参考文献本地大模型校对

原文内容截取的已经很好了

.env

@@ -19,6 +19,22 @@ client_id = 616562
 client_secret = CfMDrllyqBTFKrUkO2XaE7OmWTYqP3yd
 hmac = 8aU8WnITYhwaGTXH
 
+[promotion]
+PROMOTION_LLM_URL=http://chat.taimed.cn/v1/chat/completions
+PROMOTION_LLM_MODEL=your-model-name
+PROMOTION_LLM_TIMEOUT=30
+PROMOTION_LLM_FALLBACK="We would like to cordially invite you to consider submitting a manuscript to {{journal_name}}."
+PROMOTION_LLM_ADVISED_FALLBACK=""
+
+[unsubscribe]
+UNSUBSCRIBE_SECRET="TMR Unsubscribe Secret create on 20260427"
+UNSUBSCRIBE_BASE_URL=https://submission.tmrjournals.com/api/Unsubscribe/index
+
+[yboard]
+APPLY_URL="https://submission.tmrjournals.com/youthBoardRegister"
+
+[plagiarism]
+static_root="/home/wwwroot/api.tmrjournals.com/public"
 
 [journal]
 ;官网服务器地址

application/api/controller/Article.php

@@ -8,6 +8,8 @@ use think\Queue;
 use think\Validate;
 use PhpOffice\PhpWord\IOFactory;
 use app\common\OpenAi;
+use app\common\CrossrefService;
+use app\common\PubmedService;
 
 /**
  * @title 文章接口
@@ -17,9 +19,17 @@ class Article extends Base
 {
 
 
+    /**
+     * @var CrossrefService
+     */
+    private $crossService;
+    private $pubmedService;
+
     public function __construct(\think\Request $request = null)
     {
         parent::__construct($request);
+        $this->crossService = new CrossrefService();
+        $this->pubmedService = new PubmedService();
     }
 
 
@@ -169,16 +179,6 @@ class Article extends Base
         return jsonSuccess($re);
     }
 
-
-    public function myttt()
-    {
-        $res = $this->addProductionEx("3689");
-        echo "<pre>";
-        var_dump($res);
-        echo "</pre>";
-        die;
-    }
-
     /**获取预接收内容状态
      * @return void
      */
@@ -1053,6 +1053,36 @@ class Article extends Base
         }
     }
 
+
+    public function testCheckArticleCitation()
+    {
+        $data = $this->request->post();
+        $rule = new Validate([
+            "article_id"=>"require"
+            ]);
+        if(!$rule->check($data)){
+            return jsonError($rule->getError());
+        }
+        $mains = $this->article_main_obj->where("article_id",$data['article_id'])->where("state",0)->select();
+        $production = $this->production_article_obj->where("article_id",$data['article_id'])->find();
+        $refers = $this->production_article_refer_obj->where("p_article_id",$production['p_article_id'])->where("state",0)->order("index asc")->select();
+        $res = $this->crossService->qcArticleCitations($mains,$refers);
+        return jsonSuccess(['res'=>$res]);
+    }
+
+    public function testCheckArticlePubmed()
+    {
+        $data = $this->request->post();
+        $rule = new Validate([
+            "doi"=>"require"
+        ]);
+        if(!$rule->check($data)){
+            return jsonError($rule->getError());
+        }
+        $res = $this->pubmedService->fetchByDoi($data['doi']);
+        return jsonSuccess(['res'=>$res]);
+    }
+
     /**
      * @title 发送留言板消息
      * @description 发送留言板消息
@@ -3940,7 +3970,14 @@ class Article extends Base
 
     public function ffff(){
         $data = $this->request->post();
-        $this->ai_scor($data['article_id']);
+        $rule = new Validate([
+            "doi"=>"require"
+        ]);
+        if(!$rule->check($data)){
+            return jsonError($rule->getError());
+        }
+        $res = $this->crossService->fetchWorkSummary($data['doi']);
+        return jsonSuccess($res);
     }
 
 

application/api/controller/Base.php

@@ -271,6 +271,14 @@ class Base extends Controller
         }
         $this->production_article_refer_obj->where('p_article_id', $refer_info['p_article_id'])->where('index', ">", $refer_info['index'])->where('state', 0)->setDec('index');
         $this->production_article_refer_obj->where('p_refer_id', $p_refer_id)->update(['state' => 1]);
+
+        // 文献集合已变更，原校对结果的 reference_no 已全部错位，整篇标记为未校对
+        try {
+            (new \app\common\ReferenceCheckService())
+                ->clearArticleChecksByPArticleId(intval($refer_info['p_article_id']));
+        } catch (\Exception $e) {
+            \think\Log::error('delOneRefer clearArticleChecksByPArticleId p_refer_id=' . $p_refer_id . ' ' . $e->getMessage());
+        }
     }
 
 
@@ -1047,6 +1055,162 @@ class Base extends Controller
         return $this->production_article_obj->insertGetId($insert);
     }
 
+    public function markUnusedReferencesForArticle(int $articleId)
+    {
+        if ($articleId <= 0) return;
+
+        $production = Db::name('production_article')
+            ->where('article_id', $articleId)
+            ->where('state', 0)
+            ->field('p_article_id')
+            ->find();
+        $pArticleId = intval($production['p_article_id'] ?? 0);
+        if ($pArticleId <= 0) return;
+
+        // 1) 收集已使用的 p_refer_id（与 convertArticleMainCitationsToMycite 一致：正文 + 表格 + 图）
+        $usedIds = [];
+
+        // t_article_main：type=0 等为正文；type=2 时引用在 article_main_table，content 可能为空，仍扫描不增加成本
+        $mains = Db::name('article_main')
+            ->where('article_id', $articleId)
+            ->whereIn('state', [0, 2])
+            ->field('content')
+            ->select();
+        foreach ($mains as $row) {
+            $usedIds = array_merge($usedIds, $this->extractMyciteIds((string)($row['content'] ?? '')));
+        }
+
+        // t_article_main_table：type=2 对应表格数据（含 url）
+        $tables = Db::name('article_main_table')
+            ->where('article_id', $articleId)
+            ->where('state', 0)
+            ->field('table_data,html_data,title,note,url')
+            ->select();
+        foreach ($tables as $row) {
+            $usedIds = array_merge($usedIds, $this->extractMyciteIdsFromTableDataJson((string)($row['table_data'] ?? '')));
+            $usedIds = array_merge($usedIds, $this->extractMyciteIds((string)($row['html_data'] ?? '')));
+            $usedIds = array_merge($usedIds, $this->extractMyciteIds((string)($row['title'] ?? '')));
+            $usedIds = array_merge($usedIds, $this->extractMyciteIds((string)($row['note'] ?? '')));
+            $usedIds = array_merge($usedIds, $this->extractMyciteIds((string)($row['url'] ?? '')));
+        }
+
+        // t_article_main_image：type=1 图注等
+        $images = Db::name('article_main_image')
+            ->where('article_id', $articleId)
+            ->where('state', 0)
+            ->field('title,note,url')
+            ->select();
+        foreach ($images as $row) {
+            $usedIds = array_merge($usedIds, $this->extractMyciteIds((string)($row['title'] ?? '')));
+            $usedIds = array_merge($usedIds, $this->extractMyciteIds((string)($row['note'] ?? '')));
+            $usedIds = array_merge($usedIds, $this->extractMyciteIds((string)($row['url'] ?? '')));
+        }
+
+        $usedIds = array_values(array_unique($usedIds));
+
+        // 2) 标记：先全置 0，再把用到的置 1
+        try {
+            Db::name('production_article_refer')
+                ->where('p_article_id', $pArticleId)
+                ->where('state', 0)
+                ->update(['is_used' => 0, 'update_time' => time()]);
+
+            if (!empty($usedIds)) {
+                Db::name('production_article_refer')
+                    ->where('p_article_id', $pArticleId)
+                    ->where('state', 0)
+                    ->whereIn('p_refer_id', $usedIds)
+                    ->update(['is_used' => 1, 'update_time' => time()]);
+            }
+        } catch (\Exception $e) {
+            // 工具方法：不影响主流程，忽略异常（可按需改为记录日志）
+        }
+    }
+    public function extractMyciteIds(string $text): array
+    {
+        if ($text === '') return [];
+        $ids = [];
+        if (preg_match_all('/<\s*mycite\b[^>]*\bdata-id\s*=\s*(["\'])(.*?)\1[^>]*>/iu', $text, $m)) {
+            foreach ($m[2] as $raw) {
+                $raw = trim((string)$raw);
+                if ($raw === '') continue;
+                $parts = preg_split('/\s*,\s*/', $raw);
+                foreach ($parts as $p) {
+                    $p = trim((string)$p);
+                    if ($p === '') continue;
+                    $v = intval($p);
+                    if ($v > 0) $ids[] = $v;
+                }
+            }
+        }
+        $ids = array_values(array_unique($ids));
+        sort($ids);
+        return $ids;
+    }
+
+    /**
+     * table_data：二维数组 JSON [[{text,colspan,rowspan},...],...]；支持双重 JSON 字符串编码。
+     *
+     * @return array|null
+     */
+    protected function decodeTableDataJsonToArray(string $raw): ?array
+    {
+        $raw = trim($raw);
+        if ($raw === '') {
+            return null;
+        }
+        if (preg_match('/^\xEF\xBB\xBF/', $raw)) {
+            $raw = substr($raw, 3);
+        }
+        $decoded = json_decode($raw, true);
+        if (json_last_error() !== JSON_ERROR_NONE) {
+            return null;
+        }
+        if (is_array($decoded)) {
+            return $decoded;
+        }
+        if (is_string($decoded)) {
+            $decoded2 = json_decode($decoded, true);
+            if (json_last_error() === JSON_ERROR_NONE && is_array($decoded2)) {
+                return $decoded2;
+            }
+        }
+        return null;
+    }
+
+    /**
+     * table_data 为 JSON：递归收集所有字符串里的 <mycite>；非合法 JSON 时按整串解析。
+     *
+     * @return int[]
+     */
+    protected function extractMyciteIdsFromTableDataJson(string $tableDataJson): array
+    {
+        $tableDataJson = trim($tableDataJson);
+        if ($tableDataJson === '') {
+            return [];
+        }
+        $decoded = $this->decodeTableDataJsonToArray($tableDataJson);
+        if ($decoded === null) {
+            return $this->extractMyciteIds($tableDataJson);
+        }
+        $ids = [];
+        $walk = function ($node) use (&$walk, &$ids) {
+            if (is_string($node)) {
+                $ids = array_merge($ids, $this->extractMyciteIds($node));
+                return;
+            }
+            if (is_array($node)) {
+                foreach ($node as $v) {
+                    $walk($v);
+                }
+            }
+        };
+        $walk($decoded);
+        $ids = array_values(array_unique($ids));
+        sort($ids);
+        return $ids;
+    }
+
 
 }
 

application/api/controller/Contribute.php

@@ -6,6 +6,7 @@ use app\api\controller\Base;
 use think\Db;
 use PhpOffice\PhpWord\IOFactory;
 use think\Exception;
+use think\Validate;
 use \app\common\ArticleParserService;
 /**
  * @title 自动投稿控制器
@@ -102,6 +103,23 @@ class Contribute extends Base
         return $result;
     }
 
+
+
+    public function myTestArticle(){
+        $data = $this->request->post();
+        $rule = new Validate([
+            'article_id'=>"require"
+        ]);
+        if(!$rule->check($data)){
+            return jsonError($rule->getError());
+        }
+        $files = $this->article_file_obj->where("article_id",$data['article_id'])->where("type_name","manuscirpt")->order("file_id desc")->limit(1)->select();
+        $sFileUrl =$files[0]['file_url'];
+        $sFileUrl = rtrim(ROOT_PATH,'/').'/public/'.ltrim(ltrim($sFileUrl,'/'),'public');
+        $res = ArticleParserService::getReferencesFromWord($sFileUrl);
+        return jsonSuccess($res);
+    }
+
     /**
      * 组装数据插入相关数据表
      * @param array $aParam

application/api/controller/Country.php

@@ -0,0 +1,250 @@
+<?php
+
+namespace app\api\controller;
+
+use think\Db;
+
+class Country extends Base
+{
+    public function __construct(\think\Request $request = null)
+    {
+        parent::__construct($request);
+    }
+
+    /**
+     * 国家列表（支持筛选 + 分页）
+     *
+     * 参数:
+     *   keyword    - 搜索中文名/英文名/code
+     *   is_hot     - 0/1 筛选热门
+     *   partition  - 分区 1/2/3
+     *   page       - 页码（默认1）
+     *   per_page   - 每页条数（默认20，最大100）
+     */
+    public function getList()
+    {
+        $keyword   = trim($this->request->param('keyword', ''));
+        $isHot     = $this->request->param('is_hot', '');
+        $partition = $this->request->param('partition', '');
+        $page      = max(1, intval($this->request->param('page', 1)));
+        $perPage   = max(1, min(intval($this->request->param('per_page', 20)), 100));
+
+        $where = [];
+        if ($keyword !== '') {
+            $where[] = ['zh_name|en_name|code', 'like', '%' . $keyword . '%'];
+        }
+        if ($isHot !== '' && $isHot !== '-1') {
+            $where[] = ['is_hot', '=', intval($isHot)];
+        }
+        if ($partition !== '' && $partition !== '-1') {
+            $where[] = ['partition', '=', intval($partition)];
+        }
+
+        $countQuery = Db::name('country');
+        foreach ($where as $w) {
+            $countQuery->where($w[0], $w[1], $w[2]);
+        }
+        $total = $countQuery->count();
+
+        $listQuery = Db::name('country');
+        foreach ($where as $w) {
+            $listQuery->where($w[0], $w[1], $w[2]);
+        }
+        $list = $listQuery
+            ->order('is_hot desc, zh_name asc')
+            ->page($page, $perPage)
+            ->select();
+
+        return jsonSuccess([
+            'list'        => $list,
+            'total'       => $total,
+            'page'        => $page,
+            'per_page'    => $perPage,
+            'total_pages' => $total > 0 ? ceil($total / $perPage) : 0,
+        ]);
+    }
+
+    /**
+     * 国家下拉选项（不分页，用于筛选器）
+     *
+     * 参数:
+     *   hot_only  - 1 则只返回热门国家
+     *   partition - 按分区筛选
+     */
+    public function getOptions()
+    {
+        $hotOnly   = intval($this->request->param('hot_only', 0));
+        $partition = $this->request->param('partition', '');
+
+        $query = Db::name('country');
+
+        if ($hotOnly) {
+            $query->where('is_hot', 1);
+        }
+        if ($partition !== '' && $partition !== '-1') {
+            $query->where('partition', intval($partition));
+        }
+
+        $list = $query
+            ->field('country_id, zh_name, en_name, code, is_hot, partition')
+            ->order('is_hot desc, zh_name asc')
+            ->select();
+
+        return jsonSuccess($list);
+    }
+
+    /**
+     * 国家详情
+     */
+    public function getDetail()
+    {
+        $id = intval($this->request->param('country_id', 0));
+        if (!$id) {
+            return jsonError('country_id is required');
+        }
+
+        $row = Db::name('country')->where('country_id', $id)->find();
+        if (!$row) {
+            return jsonError('Country not found');
+        }
+
+        return jsonSuccess($row);
+    }
+
+    /**
+     * 新增国家
+     */
+    public function add()
+    {
+        $data = $this->request->post();
+
+        $zhName  = trim(isset($data['zh_name']) ? $data['zh_name'] : '');
+        $enName  = trim(isset($data['en_name']) ? $data['en_name'] : '');
+        $code    = strtoupper(trim(isset($data['code']) ? $data['code'] : ''));
+
+        if ($zhName === '' && $enName === '') {
+            return jsonError('zh_name 或 en_name 至少填一个');
+        }
+
+        if ($code !== '') {
+            $exists = Db::name('country')->where('code', $code)->find();
+            if ($exists) {
+                return jsonError('code 已存在: ' . $code);
+            }
+        }
+
+        $id = Db::name('country')->insertGetId([
+            'zh_name'   => mb_substr($zhName, 0, 32),
+            'en_name'   => mb_substr($enName, 0, 256),
+            'code'      => mb_substr($code, 0, 32),
+            'is_hot'    => intval(isset($data['is_hot']) ? $data['is_hot'] : 0),
+            'partition' => intval(isset($data['partition']) ? $data['partition'] : 2),
+        ]);
+
+        return jsonSuccess(['country_id' => $id]);
+    }
+
+    /**
+     * 编辑国家
+     */
+    public function edit()
+    {
+        $data = $this->request->post();
+        $id   = intval(isset($data['country_id']) ? $data['country_id'] : 0);
+        if (!$id) {
+            return jsonError('country_id is required');
+        }
+
+        $row = Db::name('country')->where('country_id', $id)->find();
+        if (!$row) {
+            return jsonError('Country not found');
+        }
+
+        $update = [];
+
+        if (isset($data['zh_name'])) {
+            $update['zh_name'] = mb_substr(trim($data['zh_name']), 0, 32);
+        }
+        if (isset($data['en_name'])) {
+            $update['en_name'] = mb_substr(trim($data['en_name']), 0, 256);
+        }
+        if (isset($data['code'])) {
+            $code = strtoupper(trim($data['code']));
+            $dup = Db::name('country')->where('code', $code)->where('country_id', '<>', $id)->find();
+            if ($dup) {
+                return jsonError('code 已被其他国家使用: ' . $code);
+            }
+            $update['code'] = mb_substr($code, 0, 32);
+        }
+        if (isset($data['is_hot'])) {
+            $update['is_hot'] = intval($data['is_hot']);
+        }
+        if (isset($data['partition'])) {
+            $update['partition'] = intval($data['partition']);
+        }
+
+        if (empty($update)) {
+            return jsonError('没有可更新的字段');
+        }
+
+        Db::name('country')->where('country_id', $id)->update($update);
+
+        return jsonSuccess([]);
+    }
+
+    /**
+     * 删除国家
+     */
+    public function delete()
+    {
+        $id = intval($this->request->param('country_id', 0));
+        if (!$id) {
+            return jsonError('country_id is required');
+        }
+
+        $used = Db::name('expert')->where('country_id', $id)->count();
+        if ($used > 0) {
+            return jsonError('该国家下有 ' . $used . ' 位专家关联，无法删除');
+        }
+
+        Db::name('country')->where('country_id', $id)->delete();
+
+        return jsonSuccess([]);
+    }
+
+    /**
+     * 批量设置热门
+     */
+    public function setHot()
+    {
+        $ids   = $this->request->param('country_ids', '');
+        $isHot = intval($this->request->param('is_hot', 1));
+
+        if (empty($ids)) {
+            return jsonError('country_ids is required');
+        }
+
+        $idArr = array_map('intval', explode(',', $ids));
+        $count = Db::name('country')->where('country_id', 'in', $idArr)->update(['is_hot' => $isHot]);
+
+        return jsonSuccess(['affected' => $count]);
+    }
+
+    /**
+     * 批量设置分区
+     */
+    public function setPartition()
+    {
+        $ids       = $this->request->param('country_ids', '');
+        $partition = intval($this->request->param('partition', 2));
+
+        if (empty($ids)) {
+            return jsonError('country_ids is required');
+        }
+
+        $idArr = array_map('intval', explode(',', $ids));
+        $count = Db::name('country')->where('country_id', 'in', $idArr)->update(['partition' => $partition]);
+
+        return jsonSuccess(['affected' => $count]);
+    }
+}

application/api/controller/Crossrefdoi.php

@@ -0,0 +1,326 @@
+<?php
+namespace app\api\controller;
+use app\api\controller\Base;
+use think\Db;
+use think\Validate;
+class Crossrefdoi extends Base{
+
+    public function __construct(\think\Request $request = null) {
+        parent::__construct($request);
+    }
+    // 配置项
+    private $mailto; // 邮箱（提升优先级）
+    private $timeout = 15; // 请求超时（秒）
+    private $maxRetry = 2; // 单个DOI最大重试次数
+    private $crossrefUrl = "https://api.crossref.org/works/"; //接口地址
+    /**
+     * 批量查询DOI信息（核心方法）
+     * @param array $dois DOI列表
+     * @return array 包含所有DOI的查询结果
+     */
+    public function get($aParam = []){
+
+        //获取参数
+        $aParam = empty($aParam) ? $this->request->post() : $aParam;
+
+        $iPReferId = empty($aParam['p_refer_id']) ? 0 : $aParam['p_refer_id'];
+        if(empty($iPReferId)){
+            return json_encode(['status' => 2,'msg' => 'Please select the reference to be queried']);
+        }
+        $aWhere = ['p_refer_id' => $iPReferId,'state' => ['in', [0,2]]];
+        $aRefer = Db::name('production_article_refer')->field('title,joura,author,refer_doi,doilink,cs')->where($aWhere)->find();
+        if(empty($aRefer['refer_doi'])){
+            return json_encode(['status' => 3,'msg' => 'The doi of the reference is empty']);
+        }
+
+        $finalResult = [];
+        $sDoi = empty($aRefer['refer_doi']) ? '' : $aRefer['refer_doi'];
+        $sCheckDoi = $this->filterValidDoi($sDoi); // 过滤非法DOI
+        if(empty($sCheckDoi)){
+            return json_encode(['status' => 4,'msg' => 'Doi does not comply with the rules']);
+        }
+
+        // 调用单DOI查询（带重试）
+        $aDoiInfo = $this->fetchSingleDoiWithRetry($sCheckDoi);
+        if (!$aDoiInfo) {
+            $result['status'] = 'fail';
+            $result['fail_reason'] = "请求失败（重试{$this->maxRetry}次后仍失败）";
+        }
+        // 提取核心字段
+        $sTitle = $this->getTitle($aDoiInfo);
+        //期刊信息
+        $aPublisher = $this->getPublisher($aDoiInfo);
+        $sJoura = empty($aPublisher['title']) ? $aPublisher['short_title'] : $aPublisher['title'];
+        //作者信息
+        $aAuthor = $this->getAuthors($aDoiInfo);
+        $sAuthor = empty($aAuthor) ? '' : implode(',', $aAuthor);
+        $sDateno = $this->getVolumeIssuePages($aDoiInfo);
+        
+        // 识别撤稿状态
+        $aRetractInfo = $this->checkRetracted($aDoiInfo);
+        $bIsRetracted = 2;
+        if($aRetractInfo['is_retracted'] == true){
+            $bIsRetracted = 1;
+        }
+        $sRetractReason = empty($aRetractInfo['reason']) ? '' : $aRetractInfo['reason'];
+        //获取dolink
+        $sDolink = $this->getDolink($aDoiInfo);
+        $sDolink = empty($sDolink) ? 'https://doi.org/' . $sCheckDoi : $sDolink;
+
+        //数据处理更新数据库
+        $aUpdate = [];
+        if(!empty($sTitle) && empty($aRefer['title'])){
+            $aUpdate['title'] = $sTitle;
+        }
+        if(!empty($sJoura) && empty($aRefer['joura'])){
+            $aUpdate['joura'] = $sJoura;
+        }
+        if(!empty($sAuthor) && empty($aRefer['author'])){
+            $aUpdate['author'] = $sAuthor;
+        }
+        if(!empty($sDateno)){
+            $aUpdate['dateno'] = $sDateno;
+        }
+        if($bIsRetracted == 1){
+            $aUpdate['is_retracted'] = 1;
+        }
+        if(!empty($sDolink) && empty($aRefer['doilink'])){
+            $aUpdate['doilink'] = $sDolink;
+        }
+        if(empty($aUpdate)){
+            return json_encode(['status' => 5,'msg' => 'No update information available']);
+        }
+        $aUpdate['update_time'] = time();
+        $aUpdate['cs'] = 1;
+        $aWhere = ['p_refer_id' => $iPReferId];
+        $result = Db::name('production_article_refer')->where($aWhere)->limit(1)->update($aUpdate);
+        if($result === false){
+            return json_encode(['status' => 6,'msg' => 'Update failed-Cs']);
+        }
+        return json_encode(['status' => 1,'msg' => 'Update successful']);
+    }
+
+
+    public function getOneDoi(){
+        $data = $this->request->post();
+        $rule = new Validate([
+            "doi"=>"require"
+        ]);
+        if(!$rule->check($data)){
+            return jsonError($rule->getError());
+        }
+        $sCheckDoi = $this->fetchSingleDoiWithRetry($this->filterValidDoi($data['doi'])); // 过滤非法DOI
+        return jsonSuccess($sCheckDoi);
+    }
+
+
+
+    /**
+     * 过滤非法DOI（仅保留10.xxxx/xxx格式）
+     */
+    private function filterValidDoi($doi = ''){
+        $doi = trim($doi);
+        if (empty($doi)) return '';
+        // 正则匹配：10. + 至少4位数字 + / + 任意字符
+        if (preg_match('/^10\.\d{4,}\/.+/', $doi)) {
+            return $doi;
+        }
+        return '';
+    }
+
+    /**
+     * 单DOI查询
+     */
+    private function fetchSingleDoiWithRetry($doi){
+        $retryCount = 0;
+        while ($retryCount < $this->maxRetry) {
+            $url = $this->crossrefUrl. rawurlencode($doi) . "?mailto=" . rawurlencode($this->mailto);
+            
+            $ch = curl_init();
+            curl_setopt($ch, CURLOPT_URL, $url);
+            curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
+            curl_setopt($ch, CURLOPT_TIMEOUT, $this->timeout);
+            curl_setopt($ch, CURLOPT_SSL_VERIFYPEER, false);
+            curl_setopt($ch, CURLOPT_HTTPHEADER, [
+                "User-Agent: DOI-Fetcher/1.0 (mailto:{$this->mailto})"
+            ]);
+            $response = curl_exec($ch);
+            $httpCode = curl_getinfo($ch, CURLINFO_HTTP_CODE);
+            curl_close($ch);
+
+            // 成功返回
+            if ($httpCode == 200) {
+                $data = json_decode($response, true);
+                return $data['status'] == 'ok' ? $data['message'] : null;
+            }
+
+            // 429速率限制：延长等待后重试
+            if ($httpCode == 429) {
+                sleep(5);
+                $retryCount++;
+                continue;
+            }
+
+            $retryCount++;
+            sleep(1); // 普通失败，1秒后重试
+        }
+        return null;
+    }
+
+    /**
+     * 提取标题
+     */
+    private function getTitle($aDoiInfo = []){
+        return $aDoiInfo['title'][0] ?? '';
+    }
+
+    /**
+     * 提取出版社名
+     */
+    private function getPublisher($aDoiInfo = []){
+        $aJournal = [
+            'title'      => isset($aDoiInfo['container-title'][0]) ? $aDoiInfo['container-title'][0] : '',
+            'short_title'=> isset($aDoiInfo['short-container-title'][0]) ? $aDoiInfo['short-container-title'][0] : '',
+            'ISSN'       => $aDoiInfo['ISSN'] ?? [],
+            'publisher'  => $aDoiInfo['publisher'] ?? '',
+        ];
+        return $aJournal;
+    }
+
+    /**
+     * 提取作者
+     */
+    private function getAuthors($aDoiInfo = []){
+        $authors = [];
+        if (!empty($aDoiInfo['author'])) {
+            foreach ($aDoiInfo['author'] as $author) {
+                $name = $author['family'] ?? '';
+                if (!empty($author['given'])) {
+                    $name = $author['given'] . ' ' . $name;
+                }
+                if (!empty($name)) {
+                    $authors[] = $name;
+                }
+            }
+        }
+        return $authors;
+    }
+
+    /**
+     * 提取发表年份
+     */
+    private function getPublishYear($aDoiInfo = []){
+        if (!empty($aDoiInfo['issued']['date-parts'][0][0])) {
+            return (string)$aDoiInfo['issued']['date-parts'][0][0];
+        }
+        return '';
+    }
+
+    /**
+     * 提取卷(期):起始页-终止页（格式：2024;10(2):100-120）
+     */
+    private function getVolumeIssuePages($aDoiInfo = []){
+        $parts = [];
+        // 年
+        $year = $this->getPublishYear($aDoiInfo);
+        if ($year) $parts[] = $year;
+        
+        // 卷(期)
+        $volume = $aDoiInfo['volume'] ?? '';
+        $issue = $aDoiInfo['issue'] ?? '';
+        if ($volume) {
+            $volumeIssue = $volume . ($issue ? "({$issue})" : '');
+            $parts[] = $volumeIssue;
+        }
+        
+        // 起始页-终止页
+        $pageStart = $aDoiInfo['page']['start'] ?? ($aDoiInfo['first-page'] ?? '');
+        $pageEnd = $aDoiInfo['page']['end'] ?? ($aDoiInfo['last-page'] ?? '');
+        $pages = '';
+        if ($pageStart) {
+            $pages = $pageStart . ($pageEnd ? "-{$pageEnd}" : '');
+        }else{
+            $pages = $aDoiInfo['page'] ?? '';
+        }
+        if ($pages) $parts[] = $pages;
+
+        return implode(':', $parts);
+    }
+
+    /**
+     * 识别撤稿文章
+     */
+    private function checkRetracted($aDoiInfo = []){
+        $isRetracted = false;
+        $reason = "未撤稿";
+
+        // 1. 文章类型为撤稿声明type/subtype
+        $sType = strtolower($aDoiInfo['type'] ?? '');
+        $sSubtype = strtolower($aDoiInfo['subtype'] ?? '');
+        if (isset($sType) && in_array($sType, ['retraction', 'correction'])) {
+            $isRetracted = true;
+            $reason = "文章类型为{$sType}（撤稿/更正声明）";
+        }
+
+        if (isset($sSubtype) && in_array($sSubtype, ['retraction', 'correction'])) {
+            $isRetracted = true;
+            $reason = "文章类型为{$sSubtype}（撤稿/更正声明）";
+        }
+        // 2. update-type包含撤稿
+        if (isset($aDoiInfo['update-type']) && in_array('retraction', $aDoiInfo['update-type'])) {
+            $isRetracted = true;
+            $reason = "官方标记为撤稿（update-type: retraction）";
+        }
+
+        // 3. 关联撤稿文章
+        if (isset($aDoiInfo['relation']) && !empty($aDoiInfo['relation'])) {
+            foreach ($aDoiInfo['relation'] as $relType => $relItems) {
+                if (in_array($relType, ['is-retraction-of', 'corrects'])) {
+                    $isRetracted = true;
+                    $relatedDoi = $relItems[0]['id'] ?? '未知';
+                    $reason = "关联撤稿文章{$relatedDoi}（关系：{$relType}）";
+                    break;
+                }
+            }
+        }
+
+        // 4. update-to 字段
+        if (isset($aDoiInfo['update-to']) && is_array($aDoiInfo['update-to'])) {
+            foreach ($aDoiInfo['update-to'] as $update) {
+                $updateType = strtolower($update['type'] ?? '');
+                $updateLabel = strtolower($update['label'] ?? '');
+                if (strpos($updateType, 'retract') !== false || strpos($updateLabel, 'retract') !== false) {
+                    $isRetracted = true;
+                    $retractionDetail['retraction_notice'] = [
+                        'type'  => $update['type'] ?? '',
+                        'label' => $update['label'] ?? '',
+                        'DOI'   => $update['DOI'] ?? '',
+                        'date'  => isset($update['updated']) ? $this->parseDateParts($update['updated']) : '',
+                    ];
+                    break;
+                }
+            }
+        }
+        //5.title 关键词
+        $aTitles = $aDoiInfo['title'] ?? [];
+        foreach ($aTitles as $value) {
+            $sTitleLower = strtolower($value);
+            if (strpos($sTitleLower, 'retraction') !== false || strpos($sTitleLower, 'retracted') !== false
+                || strpos($sTitleLower, 'withdrawal') !== false || strpos($sTitleLower, 'withdrawn') !== false) {
+                $isRetracted = true;
+                $retractionDetail['title_keyword'] = $value;
+                break;
+            }
+        }
+        return [
+            'is_retracted' => $isRetracted,
+            'reason' => $reason
+        ];
+    }
+    /**
+     * 识别doi链接
+     */
+    private function getDolink($aDoiInfo = []){
+        return $aDoiInfo['URL'] ?? '';
+    }
+}

application/api/controller/ExpertFinder.php

@@ -4,31 +4,22 @@ namespace app\api\controller;
 
 use think\Cache;
 use think\Db;
-use GuzzleHttp\Client;
+use think\Queue;
+use think\Validate;
+use app\common\ExpertFinderService;
 
 class ExpertFinder extends Base
 {
-    private $httpClient;
-    private $ncbiBaseUrl = 'https://eutils.ncbi.nlm.nih.gov/entrez/eutils/';
+    private $service;
 
     public function __construct(\think\Request $request = null)
     {
         parent::__construct($request);
-        $this->httpClient = new Client([
-            'timeout'         => 180,
-            'connect_timeout' => 15,
-            'verify'          => false,
-        ]);
+        $this->service = new ExpertFinderService();
     }
 
     /**
      * Main search endpoint
-     * Params:
-     *   keyword      - search term (e.g. "biomedical engineering")
-     *   page         - page number, default 1
-     *   per_page     - articles per page, default 100, max 100
-     *   min_year     - earliest publication year, default current-3
-     *   source       - "pubmed" (fast, email from affiliation) or "pmc" (slower, structured email)
      */
     public function search()
     {
@@ -49,16 +40,12 @@ class ExpertFinder extends Base
         }
 
         try {
-            if ($source === 'pmc') {
-                $result = $this->searchViaPMC($keyword, $perPage, $minYear, $page);
-            } else {
-                $result = $this->searchViaPubMed($keyword, $perPage, $minYear, $page);
-            }
+            $result = $this->service->searchExperts($keyword, $perPage, $minYear, $page, $source);
         } catch (\Exception $e) {
             return jsonError('Search failed: ' . $e->getMessage());
         }
 
-        $saveResult = $this->saveExperts($result['experts'], $keyword, $source);
+        $saveResult = $this->service->saveExperts($result['experts'], $keyword, $source);
         $result['saved_new']   = $saveResult['inserted'];
         $result['saved_exist'] = $saveResult['existing'];
 
@@ -67,23 +54,65 @@ class ExpertFinder extends Base
         return jsonSuccess($result);
     }
 
+    /**
+     * 测试单个专家的国家解析（同步执行，立刻返回结果）
+     */
+    public function cityTest(){
+        $data = $this->request->post();
+        $rule = new Validate([
+            "expert_id"=>"require",
+            "aff"=>"require"
+        ]);
+        if(!$rule->check($data)){
+            return jsonError($rule->getError());
+        }
+        $data1 = [
+            'expert_id'   => intval($data['expert_id']),
+            'affiliation' => trim((string)$data['aff']),
+        ];
+
+//        Queue::push('app\api\job\FillExpertCountry@fire', $data1, 'FetchExpertCity');
+
+
+
+        $service = new ExpertFinderService();
+        $service->fillExpertCountry($data['expert_id'], $data['aff']);
+        $expert = Db::name('expert')->where('expert_id', intval($data['expert_id']))->find();
+        return jsonSuccess([
+            'country_id' => isset($expert['country_id']) ? $expert['country_id'] : null,
+            'country'    => isset($expert['country']) ? $expert['country'] : null,
+        ]);
+    }
+
+    /**
+     * 启动国家解析：同时启动两条链，分别用不同模型并行处理。
+     * 只需调一次，两条链各自链式执行直到全部处理完。
+     */
+    public function batchFillCountry(){
+        $service = new ExpertFinderService();
+
+//        $chain1 = $service->enqueueNextCountryFill(0, 'FetchExpertCity', '');
+        $chain2 = $service->enqueueNextCountryFill(0, 'FetchExpertCityOne', 'http://125.39.141.154:10002/v1/chat/completions');
+
+        $pending = Db::name('expert')
+            ->where('affiliation', '<>', '')
+            ->where('country_id', 0)
+            ->where('state', '<>', 5)
+            ->count();
+
+        return jsonSuccess([
+//            'chain1_started' => $chain1,
+            'chain2_started' => $chain2,
+            'pending' => $pending,
+        ]);
+    }
+
     /**
      * Get experts from local database
-     * Params:
-     *   field       - filter by field keyword (searches t_expert_field)
-     *   major_id    - filter by major_id (searches t_expert_field)
-     *   state       - filter by state (0-5), -1 for all
-     *   keyword     - search name/email/affiliation
-     *   no_recent   - if 1, exclude experts promoted within last N days (default 30)
-     *   recent_days - days threshold for no_recent filter, default 30
-     *   page        - page number, default 1
-     *   per_page    - items per page, default 20
-     *   min_experts - auto-fetch threshold, default 50
      */
     public function getList()
     {
         $field      = trim($this->request->param('field', ''));
-        $majorId    = intval($this->request->param('major_id', 0));
         $state      = $this->request->param('state', '-1');
         $keyword    = trim($this->request->param('keyword', ''));
         $noRecent   = intval($this->request->param('no_recent', 0));
@@ -93,16 +122,13 @@ class ExpertFinder extends Base
         $minExperts = max(0, intval($this->request->param('min_experts', 50)));
 
         $query = Db::name('expert')->alias('e');
-        $needJoin = ($field !== '' || $majorId > 0);
+        $needJoin = ($field !== '');
 
         if ($needJoin) {
             $query->join('t_expert_field ef', 'ef.expert_id = e.expert_id AND ef.state = 0', 'inner');
             if ($field !== '') {
                 $query->where('ef.field', 'like', '%' . $field . '%');
             }
-            if ($majorId > 0) {
-                $query->where('ef.major_id', $majorId);
-            }
             $query->group('e.expert_id');
         }
 
@@ -132,6 +158,7 @@ class ExpertFinder extends Base
             $item['fields'] = Db::name('expert_field')
                 ->where('expert_id', $item['expert_id'])
                 ->where('state', 0)
+                ->group('field')
                 ->column('field');
         }
 
@@ -171,9 +198,6 @@ class ExpertFinder extends Base
 
     /**
      * Update expert state
-     * Params:
-     *   expert_id  - single id or comma-separated ids
-     *   state      - new state (0-5)
      */
     public function updateState()
     {
@@ -195,9 +219,6 @@ class ExpertFinder extends Base
 
     /**
      * Delete expert (soft: set state=5 blacklist, or hard delete)
-     * Params:
-     *   expert_id  - single id or comma-separated ids
-     *   hard       - 1 for hard delete, default 0 (blacklist)
      */
     public function deleteExpert()
     {
@@ -221,7 +242,6 @@ class ExpertFinder extends Base
 
     /**
      * Export search results to Excel
-     * Same params as search(), exports current page results
      */
     public function export()
     {
@@ -240,11 +260,7 @@ class ExpertFinder extends Base
 
         if (!$cached) {
             try {
-                if ($source === 'pmc') {
-                    $cached = $this->searchViaPMC($keyword, $perPage, $minYear, $page);
-                } else {
-                    $cached = $this->searchViaPubMed($keyword, $perPage, $minYear, $page);
-                }
+                $cached = $this->service->searchExperts($keyword, $perPage, $minYear, $page, $source);
                 Cache::set($cacheKey, $cached, 3600);
             } catch (\Exception $e) {
                 return jsonError('Search failed: ' . $e->getMessage());
@@ -276,16 +292,66 @@ class ExpertFinder extends Base
 
     // ==================== Cron / Auto Fetch ====================
 
+    /**
+     * Daily cron: auto-fetch experts for all active fields in t_expert_fetch via queue.
+     * No longer tied to journals; t_expert_fetch is the sole source of crawl targets.
+     */
+    public function dailyFetchAll()
+    {
+        $perPage = max(10, intval($this->request->param('per_page', 10)));
+        $source  = $this->request->param('source', 'pubmed');
+        $minYear = intval($this->request->param('min_year', date('Y') - 3));
+
+        $fetchList = Db::name('expert_fetch')
+            ->where('state', 0)
+            ->select();
+
+        if (empty($fetchList)) {
+            return jsonSuccess(['msg' => 'No active fetch fields found', 'queued' => 0]);
+        }
+
+        $queued     = 0;
+        $skipped    = 0;
+        $details    = [];
+        $todayStart = strtotime(date('Y-m-d'));
+
+        foreach ($fetchList as $item) {
+            $keyword    = trim($item['field']);
+            $itemSource = trim($item['source'] ?: $source);
+            if ($keyword === '') continue;
+
+            $fetchLog = $this->service->getFetchLog($keyword, $itemSource);
+            if ($fetchLog['last_time'] >= $todayStart) {
+                $skipped++;
+                continue;
+            }
+
+            $delay = $queued * 10;
+            \think\Queue::later($delay, 'app\api\job\FetchExperts@fire', [
+                'field'    => $keyword,
+                'source'   => $itemSource,
+                'per_page' => $perPage,
+                'min_year' => $minYear,
+            ], 'FetchExperts');
+
+            $queued++;
+            $details[] = [
+                'expert_fetch_id' => $item['expert_fetch_id'],
+                'field'           => $keyword,
+                'source'          => $itemSource,
+                'delay_s'         => $delay,
+            ];
+        }
+
+        return jsonSuccess([
+            'queued'  => $queued,
+            'skipped' => $skipped,
+            'details' => $details,
+        ]);
+    }
+
     /**
      * Cron job: daily fetch experts for given keywords
-     * Params:
-     *   keywords  - comma-separated keywords (e.g. "biomedical engineering,tissue engineering")
-     *   source    - pubmed or pmc, default pubmed
-     *   per_page  - articles per page, default 100
-     *   min_year  - default current-3
-     *
-     * Uses cache to remember which page was last fetched per keyword,
-     * so each cron run fetches the next page automatically.
      */
     public function cronFetch()
     {
@@ -303,39 +369,14 @@ class ExpertFinder extends Base
         $report   = [];
 
         foreach ($keywords as $kw) {
-            if (empty($kw)) {
-                continue;
-            }
-
-            $fetchLog = $this->getFetchLog($kw, $source);
-            $page     = $fetchLog['last_page'] + 1;
+            if (empty($kw)) continue;
 
             try {
-                if ($source === 'pmc') {
-                    $result = $this->searchViaPMC($kw, $perPage, $minYear, $page);
-                } else {
-                    $result = $this->searchViaPubMed($kw, $perPage, $minYear, $page);
-                }
-
-                $saveResult = $this->saveExperts($result['experts'], $kw, $source);
-
-                $nextPage   = $result['has_more'] ? $page : 0;
-                $totalPages = isset($result['total_pages']) ? $result['total_pages'] : 0;
-                $this->updateFetchLog($kw, $source, $nextPage, $totalPages);
-
-                $report[] = [
-                    'keyword'        => $kw,
-                    'page'           => $page,
-                    'experts_found'  => $result['total'],
-                    'saved_new'      => $saveResult['inserted'],
-                    'saved_exist'    => $saveResult['existing'],
-                    'field_enriched' => $saveResult['field_enriched'],
-                    'has_more'       => $result['has_more'],
-                ];
+                $result  = $this->service->doFetchForField($kw, $source, $perPage, $minYear);
+                $report[] = $result;
             } catch (\Exception $e) {
                 $report[] = [
                     'keyword' => $kw,
-                    'page'    => $page,
                     'error'   => $e->getMessage(),
                 ];
             }
@@ -357,7 +398,7 @@ class ExpertFinder extends Base
         }
         Cache::set($lockKey, 1, 300);
 
-        \think\Queue::push('app\api\job\FetchExperts', [
+        \think\Queue::push('app\api\job\FetchExperts@fire', [
             'field'    => $field,
             'source'   => 'pubmed',
             'per_page' => 100,
@@ -365,459 +406,17 @@ class ExpertFinder extends Base
         ], 'FetchExperts');
     }
 
-    /**
-     * Internal method: run a fetch for a single keyword (used by both cron and queue job)
-     */
-    public function doFetchForField($field, $source = 'pubmed', $perPage = 100, $minYear = null)
+    public function fetchOneField()
     {
-        if ($minYear === null) {
-            $minYear = date('Y') - 3;
-        }
-
-        $fetchLog = $this->getFetchLog($field, $source);
-        $page     = $fetchLog['last_page'] + 1;
-
-        if ($source === 'pmc') {
-            $result = $this->searchViaPMC($field, $perPage, $minYear, $page);
-        } else {
-            $result = $this->searchViaPubMed($field, $perPage, $minYear, $page);
-        }
-
-        $saveResult = $this->saveExperts($result['experts'], $field, $source);
-
-        $nextPage   = $result['has_more'] ? $page : 0;
-        $totalPages = isset($result['total_pages']) ? $result['total_pages'] : 0;
-        $this->updateFetchLog($field, $source, $nextPage, $totalPages);
-
-        return [
-            'keyword'        => $field,
-            'page'           => $page,
-            'experts_found'  => $result['total'],
-            'saved_new'      => $saveResult['inserted'],
-            'saved_exist'    => $saveResult['existing'],
-            'field_enriched' => $saveResult['field_enriched'],
-            'has_more'       => $result['has_more'],
-        ];
-    }
-
-    // ==================== PubMed Search ====================
-
-    private function searchViaPubMed($keyword, $perPage, $minYear, $page = 1)
-    {
-        set_time_limit(600);
-
-        $searchResult = $this->esearch('pubmed', $keyword, $perPage, $minYear, $page);
-        $ids          = $searchResult['ids'];
-        $totalArticles = $searchResult['total'];
-
-        if (empty($ids)) {
-            return $this->buildPagedResult([], 0, 0, $totalArticles, $page, $perPage, 'pubmed');
-        }
-
-        $allAuthors = [];
-        $batches = array_chunk($ids, 50);
-        foreach ($batches as $batch) {
-            $xml = $this->efetchWithRetry('pubmed', $batch);
-            if ($xml) {
-                $authors = $this->parsePubMedXml($xml);
-                $allAuthors = array_merge($allAuthors, $authors);
-            }
-            usleep(400000);
-        }
-
-        $experts = $this->aggregateExperts($allAuthors);
-
-        return $this->buildPagedResult($experts, count($experts), count($ids), $totalArticles, $page, $perPage, 'pubmed');
-    }
-
-    // ==================== PMC Search ====================
-
-    private function searchViaPMC($keyword, $perPage, $minYear, $page = 1)
-    {
-        set_time_limit(600);
-
-        $searchResult = $this->esearch('pmc', $keyword, $perPage, $minYear, $page);
-        $ids          = $searchResult['ids'];
-        $totalArticles = $searchResult['total'];
-
-        if (empty($ids)) {
-            return $this->buildPagedResult([], 0, 0, $totalArticles, $page, $perPage, 'pmc');
-        }
-
-        $allAuthors = [];
-        $batches = array_chunk($ids, 5);
-        foreach ($batches as $batch) {
-            $xml = $this->efetchWithRetry('pmc', $batch);
-            if ($xml) {
-                $authors = $this->parsePMCXml($xml);
-                $allAuthors = array_merge($allAuthors, $authors);
-            }
-            usleep(500000);
-        }
-
-        $experts = $this->aggregateExperts($allAuthors);
-
-        return $this->buildPagedResult($experts, count($experts), count($ids), $totalArticles, $page, $perPage, 'pmc');
-    }
-
-    // ==================== NCBI API Calls ====================
-
-    private function esearch($db, $keyword, $perPage, $minYear, $page = 1)
-    {
-        $term     = $keyword . ' AND ' . $minYear . ':' . date('Y') . '[pdat]';
-        $retstart = ($page - 1) * $perPage;
-
-        $response = $this->httpClient->get($this->ncbiBaseUrl . 'esearch.fcgi', [
-            'query' => [
-                'db'       => $db,
-                'term'     => $term,
-                'retstart' => $retstart,
-                'retmax'   => $perPage,
-                'retmode'  => 'json',
-                'sort'     => 'relevance',
-            ],
+        $data = $this->request->post();
+        $rule = new Validate([
+            "field" => "require"
         ]);
-
-        $data   = json_decode($response->getBody()->getContents(), true);
-        $ids    = $data['esearchresult']['idlist'] ?? [];
-        $total  = intval($data['esearchresult']['count'] ?? 0);
-
-        return ['ids' => $ids, 'total' => $total];
-    }
-
-    private function efetch($db, $ids)
-    {
-        $response = $this->httpClient->post($this->ncbiBaseUrl . 'efetch.fcgi', [
-            'form_params' => [
-                'db'      => $db,
-                'id'      => implode(',', $ids),
-                'retmode' => 'xml',
-            ],
-        ]);
-
-        return $response->getBody()->getContents();
-    }
-
-    private function efetchWithRetry($db, $ids, $maxRetries = 3)
-    {
-        for ($attempt = 1; $attempt <= $maxRetries; $attempt++) {
-            try {
-                return $this->efetch($db, $ids);
-            } catch (\Exception $e) {
-                if ($attempt === $maxRetries) {
-                    if (count($ids) > 1) {
-                        $half = ceil(count($ids) / 2);
-                        $firstHalf  = array_slice($ids, 0, $half);
-                        $secondHalf = array_slice($ids, $half);
-                        $xml1 = $this->efetchWithRetry($db, $firstHalf, 2);
-                        $xml2 = $this->efetchWithRetry($db, $secondHalf, 2);
-                        return $this->mergeXml($xml1, $xml2);
-                    }
-                    return null;
-                }
-                sleep($attempt * 2);
-            }
+        if (!$rule->check($data)) {
+            return jsonError($rule->getError());
         }
-        return null;
-    }
-
-    private function mergeXml($xml1, $xml2)
-    {
-        if (empty($xml1)) return $xml2;
-        if (empty($xml2)) return $xml1;
-        return $xml1 . "\n" . $xml2;
-    }
-
-    // ==================== PubMed XML Parsing ====================
-
-    private function parsePubMedXml($xmlString)
-    {
-        $results = [];
-
-        libxml_use_internal_errors(true);
-        $xml = simplexml_load_string($xmlString);
-        if ($xml === false) {
-            return $results;
-        }
-
-        foreach ($xml->PubmedArticle as $article) {
-            $citation   = $article->MedlineCitation;
-            $articleData = $citation->Article;
-
-            $title = $this->xmlNodeToString($articleData->ArticleTitle);
-            $pmid  = (string) $citation->PMID;
-
-            $journal = '';
-            if (isset($articleData->Journal->Title)) {
-                $journal = (string) $articleData->Journal->Title;
-            }
-
-            if (!isset($articleData->AuthorList->Author)) {
-                continue;
-            }
-
-            foreach ($articleData->AuthorList->Author as $author) {
-                $lastName   = (string) ($author->LastName ?? '');
-                $foreName   = (string) ($author->ForeName ?? '');
-                $fullName   = trim($foreName . ' ' . $lastName);
-
-                if (empty($fullName)) {
-                    continue;
-                }
-
-                $email       = '';
-                $affiliation = '';
-
-                if (isset($author->AffiliationInfo)) {
-                    foreach ($author->AffiliationInfo as $affInfo) {
-                        $affText = (string) $affInfo->Affiliation;
-                        if (empty($affiliation)) {
-                            $affiliation = $affText;
-                        }
-                        if (empty($email)) {
-                            $email = $this->extractEmailFromText($affText);
-                        }
-                    }
-                }
-
-                if (empty($email)) {
-                    continue;
-                }
-
-                $results[] = [
-                    'name'          => $fullName,
-                    'email'         => strtolower($email),
-                    'affiliation'   => $this->cleanAffiliation($affiliation),
-                    'article_title' => $title,
-                    'article_id'    => $pmid,
-                    'journal'       => $journal,
-                ];
-            }
-        }
-
-        return $results;
-    }
-
-    // ==================== PMC XML Parsing ====================
-
-    private function parsePMCXml($xmlString)
-    {
-        $results = [];
-
-        libxml_use_internal_errors(true);
-        $xml = simplexml_load_string($xmlString);
-        if ($xml === false) {
-            return $results;
-        }
-
-        $articles = $xml->article ?? $xml->children();
-
-        foreach ($articles as $article) {
-            if ($article->getName() !== 'article') {
-                continue;
-            }
-
-            $front = $article->front;
-            if (!$front) {
-                continue;
-            }
-
-            $articleMeta = $front->{'article-meta'};
-            if (!$articleMeta) {
-                continue;
-            }
-
-            $title = $this->xmlNodeToString($articleMeta->{'title-group'}->{'article-title'} ?? null);
-            $pmcId = '';
-            if (isset($articleMeta->{'article-id'})) {
-                foreach ($articleMeta->{'article-id'} as $idNode) {
-                    if ((string) $idNode['pub-id-type'] === 'pmc') {
-                        $pmcId = (string) $idNode;
-                    }
-                }
-            }
-
-            $journal = '';
-            if (isset($front->{'journal-meta'}->{'journal-title'})) {
-                $journal = (string) $front->{'journal-meta'}->{'journal-title'};
-            } elseif (isset($front->{'journal-meta'}->{'journal-title-group'}->{'journal-title'})) {
-                $journal = (string) $front->{'journal-meta'}->{'journal-title-group'}->{'journal-title'};
-            }
-
-            $correspEmails = [];
-            if (isset($articleMeta->{'author-notes'})) {
-                $this->extractEmailsFromNode($articleMeta->{'author-notes'}, $correspEmails);
-            }
-
-            $affiliationMap = [];
-            if (isset($articleMeta->{'contrib-group'})) {
-                foreach ($articleMeta->{'contrib-group'}->children() as $child) {
-                    if ($child->getName() === 'aff') {
-                        $affId   = (string) ($child['id'] ?? '');
-                        $affText = $this->xmlNodeToString($child);
-                        if ($affId) {
-                            $affiliationMap[$affId] = $affText;
-                        }
-                    }
-                }
-            }
-            if (isset($front->{'article-meta'}->{'aff'})) {
-                foreach ($front->{'article-meta'}->{'aff'} as $aff) {
-                    $affId   = (string) ($aff['id'] ?? '');
-                    $affText = $this->xmlNodeToString($aff);
-                    if ($affId) {
-                        $affiliationMap[$affId] = $affText;
-                    }
-                }
-            }
-
-            if (!isset($articleMeta->{'contrib-group'})) {
-                continue;
-            }
-
-            foreach ($articleMeta->{'contrib-group'}->contrib as $contrib) {
-                $contribType = (string) ($contrib['contrib-type'] ?? '');
-                if ($contribType !== 'author') {
-                    continue;
-                }
-
-                $nameNode  = $contrib->name;
-                if (!$nameNode) {
-                    continue;
-                }
-                $surname    = (string) ($nameNode->surname ?? '');
-                $givenNames = (string) ($nameNode->{'given-names'} ?? '');
-                $fullName   = trim($givenNames . ' ' . $surname);
-
-                if (empty($fullName)) {
-                    continue;
-                }
-
-                $email = '';
-                if (isset($contrib->email)) {
-                    $email = strtolower(trim((string) $contrib->email));
-                }
-
-                $affiliation = '';
-                if (isset($contrib->xref)) {
-                    foreach ($contrib->xref as $xref) {
-                        if ((string) $xref['ref-type'] === 'aff') {
-                            $rid = (string) $xref['rid'];
-                            if (isset($affiliationMap[$rid])) {
-                                $affiliation = $affiliationMap[$rid];
-                                break;
-                            }
-                        }
-                    }
-                }
-                if (empty($affiliation) && isset($contrib->aff)) {
-                    $affiliation = $this->xmlNodeToString($contrib->aff);
-                }
-
-                $isCorresponding = false;
-                if (isset($contrib->xref)) {
-                    foreach ($contrib->xref as $xref) {
-                        if ((string) $xref['ref-type'] === 'corresp') {
-                            $isCorresponding = true;
-                        }
-                    }
-                }
-                if ((string) ($contrib['corresp'] ?? '') === 'yes') {
-                    $isCorresponding = true;
-                }
-
-                if (empty($email) && $isCorresponding && !empty($correspEmails)) {
-                    $email = $correspEmails[0];
-                }
-
-                if (empty($email)) {
-                    $extracted = $this->extractEmailFromText($affiliation);
-                    if ($extracted) {
-                        $email = $extracted;
-                    }
-                }
-
-                if (empty($email)) {
-                    continue;
-                }
-
-                $results[] = [
-                    'name'          => $fullName,
-                    'email'         => strtolower($email),
-                    'affiliation'   => $this->cleanAffiliation($affiliation),
-                    'article_title' => $title,
-                    'article_id'    => $pmcId,
-                    'journal'       => $journal,
-                ];
-            }
-        }
-
-        return $results;
-    }
-
-    // ==================== Pagination ====================
-
-    private function buildPagedResult($experts, $expertCount, $articlesScanned, $totalArticles, $page, $perPage, $source)
-    {
-        $totalPages = $totalArticles > 0 ? ceil($totalArticles / $perPage) : 0;
-
-        return [
-            'experts'          => $experts,
-            'total'            => $expertCount,
-            'articles_scanned' => $articlesScanned,
-            'total_articles'   => $totalArticles,
-            'page'             => $page,
-            'per_page'         => $perPage,
-            'total_pages'      => $totalPages,
-            'has_more'         => $page < $totalPages,
-            'source'           => $source,
-        ];
-    }
-
-    // ==================== Aggregation ====================
-
-    private function aggregateExperts($authorRecords)
-    {
-        $map = [];
-
-        foreach ($authorRecords as $record) {
-            $key = strtolower(trim($record['email']));
-            if (empty($key)) {
-                continue;
-            }
-
-            if (!isset($map[$key])) {
-                $map[$key] = [
-                    'name'        => $record['name'],
-                    'email'       => $record['email'],
-                    'affiliation' => $record['affiliation'],
-                    'paper_count' => 0,
-                    'papers'      => [],
-                ];
-            }
-
-            $map[$key]['paper_count']++;
-
-            if (count($map[$key]['papers']) < 10) {
-                $map[$key]['papers'][] = [
-                    'title'      => $record['article_title'],
-                    'article_id' => $record['article_id'],
-                    'journal'    => $record['journal'],
-                ];
-            }
-
-            if (empty($map[$key]['affiliation']) && !empty($record['affiliation'])) {
-                $map[$key]['affiliation'] = $record['affiliation'];
-            }
-        }
-
-        $experts = array_values($map);
-
-        usort($experts, function ($a, $b) {
-            return $b['paper_count'] - $a['paper_count'];
-        });
-
-        return $experts;
+        $res = $this->service->doFetchForField($data['field'], "pubmed", 100, date('Y') - 3);
+        return jsonSuccess($res);
     }
 
     // ==================== Excel Export ====================
@@ -880,188 +479,4 @@ class ExpertFinder extends Base
             'count'     => count($experts),
         ]);
     }
-
-    // ==================== Database Storage ====================
-
-    private function saveExperts($experts, $field, $source)
-    {
-        $inserted    = 0;
-        $existing    = 0;
-        $fieldEnrich = 0;
-
-        foreach ($experts as $expert) {
-            $email = strtolower(trim($expert['email']));
-            if (empty($email)) {
-                continue;
-            }
-
-            $exists = Db::name('expert')->where('email', $email)->find();
-
-            if ($exists) {
-                $existing++;
-                $fieldEnrich += $this->enrichExpertField($exists['expert_id'], $field);
-                continue;
-            }
-
-            $insert = [
-                'name'        => mb_substr($expert['name'], 0, 255),
-                'email'       => mb_substr($email, 0, 128),
-                'affiliation' => mb_substr($expert['affiliation'], 0, 128),
-                'source'      => mb_substr($source, 0, 128),
-                'ctime'       => time(),
-                'ltime'       => 0,
-                'state'       => 0,
-            ];
-
-            try {
-                $expertId = Db::name('expert')->insertGetId($insert);
-                $this->enrichExpertField($expertId, $field);
-                $inserted++;
-            } catch (\Exception $e) {
-                $existing++;
-            }
-        }
-
-        return ['inserted' => $inserted, 'existing' => $existing, 'field_enriched' => $fieldEnrich];
-    }
-
-    // ==================== Fetch Log (t_expert_fetch) ====================
-
-    private function getFetchLog($field, $source)
-    {
-        $log = Db::name('expert_fetch')
-            ->where('field', $field)
-            ->where('source', $source)
-            ->find();
-
-        if (!$log) {
-            return ['last_page' => 0, 'total_pages' => 0, 'last_time' => 0];
-        }
-
-        return $log;
-    }
-
-    private function updateFetchLog($field, $source, $lastPage, $totalPages)
-    {
-        $exists = Db::name('expert_fetch')
-            ->where('field', $field)
-            ->where('source', $source)
-            ->find();
-
-        if ($exists) {
-            Db::name('expert_fetch')
-                ->where('expert_fetch_id', $exists['expert_fetch_id'])
-                ->update([
-                    'last_page'   => $lastPage,
-                    'total_pages' => $totalPages,
-                    'last_time'   => time(),
-                ]);
-        } else {
-            Db::name('expert_fetch')->insert([
-                'field'       => mb_substr($field, 0, 128),
-                'source'      => mb_substr($source, 0, 128),
-                'last_page'   => $lastPage,
-                'total_pages' => $totalPages,
-                'last_time'   => time(),
-            ]);
-        }
-    }
-
-    private function enrichExpertField($expertId, $field)
-    {
-        $field = trim($field);
-        if (empty($field)) {
-            return 0;
-        }
-
-        $exists = Db::name('expert_field')
-            ->where('expert_id', $expertId)
-            ->where('field', $field)
-            ->where('state', 0)
-            ->find();
-
-        if ($exists) {
-            return 0;
-        }
-
-        Db::name('expert_field')->insert([
-            'expert_id' => $expertId,
-            'major_id'  => 0,
-            'field'     => mb_substr($field, 0, 128),
-            'state'     => 0,
-        ]);
-
-        return 1;
-    }
-
-    // ==================== Helper Methods ====================
-
-    private function extractEmailFromText($text)
-    {
-        if (empty($text)) {
-            return '';
-        }
-
-        if (preg_match('/[Ee]lectronic address:\s*([^\s;,]+@[^\s;,]+)/', $text, $m)) {
-            return strtolower(trim($m[1], '.'));
-        }
-
-        if (preg_match('/[Ee]-?mail:\s*([^\s;,]+@[^\s;,]+)/', $text, $m)) {
-            return strtolower(trim($m[1], '.'));
-        }
-
-        if (preg_match('/\b([a-zA-Z0-9._%+\-]+@[a-zA-Z0-9.\-]+\.[a-zA-Z]{2,})\b/', $text, $m)) {
-            return strtolower(trim($m[1], '.'));
-        }
-
-        return '';
-    }
-
-    private function extractEmailsFromNode($node, &$emails)
-    {
-        if ($node === null) {
-            return;
-        }
-
-        foreach ($node->children() as $child) {
-            if ($child->getName() === 'email') {
-                $email = strtolower(trim((string) $child));
-                if (!empty($email) && !in_array($email, $emails)) {
-                    $emails[] = $email;
-                }
-            }
-            $this->extractEmailsFromNode($child, $emails);
-        }
-
-        $text = (string) $node;
-        if (preg_match_all('/\b([a-zA-Z0-9._%+\-]+@[a-zA-Z0-9.\-]+\.[a-zA-Z]{2,})\b/', $text, $matches)) {
-            foreach ($matches[1] as $email) {
-                $email = strtolower(trim($email, '.'));
-                if (!in_array($email, $emails)) {
-                    $emails[] = $email;
-                }
-            }
-        }
-    }
-
-    private function cleanAffiliation($text)
-    {
-        $text = preg_replace('/\s*[Ee]lectronic address:\s*[^\s;,]+@[^\s;,]+/', '', $text);
-        $text = preg_replace('/\s*[Ee]-?mail:\s*[^\s;,]+@[^\s;,]+/', '', $text);
-        $text = preg_replace('/\s*\b[a-zA-Z0-9._%+\-]+@[a-zA-Z0-9.\-]+\.[a-zA-Z]{2,}\b/', '', $text);
-        $text = trim($text, " \t\n\r\0\x0B.,;");
-        return $text;
-    }
-
-    private function xmlNodeToString($node)
-    {
-        if ($node === null) {
-            return '';
-        }
-
-        $xml = $node->asXML();
-        $text = strip_tags($xml);
-        $text = html_entity_decode($text, ENT_QUOTES | ENT_XML1, 'UTF-8');
-        return trim(preg_replace('/\s+/', ' ', $text));
-    }
 }

application/api/controller/ExpertManage.php

@@ -0,0 +1,710 @@
+<?php
+
+namespace app\api\controller;
+
+use think\Db;
+
+class ExpertManage extends Base
+{
+    private $stateMap = [
+        0 => '待联系',
+        1 => '已发邮件',
+        2 => '已回复',
+        3 => '已投稿',
+        4 => '退信/无效',
+        5 => '黑名单(退订)',
+    ];
+
+    public function __construct(\think\Request $request = null)
+    {
+        parent::__construct($request);
+    }
+
+    /**
+     * 专家列表（支持多条件筛选 + 分页）
+     *
+     * 参数:
+     *   keyword    - 搜索姓名/邮箱/单位
+     *   field      - 按领域关键词筛选
+     *   major_id   - 按学科ID筛选
+     *   state      - 状态筛选 (0-5, 传-1或不传则不过滤)
+     *   source     - 来源筛选
+     *   pageIndex  - 页码 (默认1)
+     *   pageSize   - 每页条数 (默认20)
+     */
+    public function getList()
+    {
+        $data     = $this->request->param();
+        $keyword  = trim(isset($data['keyword']) ? $data['keyword'] : '');
+        $field    = trim(isset($data['field']) ? $data['field'] : '');
+        $state    = isset($data['state']) ? $data['state'] : '-1';
+        $source   = trim(isset($data['source']) ? $data['source'] : '');
+        $page     = max(1, intval(isset($data['pageIndex']) ? $data['pageIndex'] : 1));
+        $pageSize = max(1, intval(isset($data['pageSize']) ? $data['pageSize'] : 20));
+
+        $query = Db::name('expert')->alias('e');
+        $countQuery = Db::name('expert')->alias('e');
+        $needJoin = ($field !== '');
+
+        if ($needJoin) {
+            $query->join('t_expert_field ef', 'ef.expert_id = e.expert_id AND ef.state = 0', 'inner');
+            $countQuery->join('t_expert_field ef', 'ef.expert_id = e.expert_id AND ef.state = 0', 'inner');
+            if ($field !== '') {
+                $query->where('ef.field', 'like', '%' . $field . '%');
+                $countQuery->where('ef.field', 'like', '%' . $field . '%');
+            }
+            $query->group('e.expert_id');
+            $countQuery->group('e.expert_id');
+        }
+
+        if ($state !== '-1' && $state !== '') {
+            $query->where('e.state', intval($state));
+            $countQuery->where('e.state', intval($state));
+        }
+        if ($keyword !== '') {
+            $query->where('e.name|e.email|e.affiliation', 'like', '%' . $keyword . '%');
+            $countQuery->where('e.name|e.email|e.affiliation', 'like', '%' . $keyword . '%');
+        }
+        if ($source !== '') {
+            $query->where('e.source', $source);
+            $countQuery->where('e.source', $source);
+        }
+
+//        $countQuery = clone $query;
+//        $total = $countQuery->distinct('e.expert_id')->count();
+        $total = $needJoin ? count($countQuery->group('e.expert_id')->column('e.expert_id')) : $countQuery->count();
+
+        $list = $query
+            ->field('e.*')
+            ->order('e.ctime desc')
+            ->page($page, $pageSize)
+            ->select();
+
+        foreach ($list as &$item) {
+//            $item['fields'] = Db::name('expert_field')
+//                ->where('expert_id', $item['expert_id'])
+//                ->where('state', 0)
+//                ->select();
+            $item['state_text'] = isset($this->stateMap[$item['state']]) ? $this->stateMap[$item['state']] : '未知';
+            $item['ctime_text'] = $item['ctime'] ? date('Y-m-d H:i:s', $item['ctime']) : '';
+            $item['ltime_text'] = $item['ltime'] ? date('Y-m-d H:i:s', $item['ltime']) : '';
+        }
+
+        return jsonSuccess([
+            'list'       => $list,
+            'total'      => $total,
+            'pageIndex'  => $page,
+            'pageSize'   => $pageSize,
+            'totalPages' => $total > 0 ? ceil($total / $pageSize) : 0,
+        ]);
+    }
+
+    /**
+     * 获取专家详情（含所有领域）
+     */
+    public function getDetail()
+    {
+        $expertId = intval($this->request->param('expert_id', 0));
+        if (!$expertId) {
+            return jsonError('expert_id is required');
+        }
+
+        $expert = Db::name('expert')->where('expert_id', $expertId)->find();
+        if (!$expert) {
+            return jsonError('专家不存在');
+        }
+
+        $expert['fields'] = Db::name('expert_field')
+            ->where('expert_id', $expertId)
+            ->where('state', 0)
+            ->select();
+        $expert['state_text'] = isset($this->stateMap[$expert['state']]) ? $this->stateMap[$expert['state']] : '未知';
+        $expert['ctime_text'] = $expert['ctime'] ? date('Y-m-d H:i:s', $expert['ctime']) : '';
+        $expert['ltime_text'] = $expert['ltime'] ? date('Y-m-d H:i:s', $expert['ltime']) : '';
+
+        return jsonSuccess($expert);
+    }
+
+    /**
+     * 添加专家
+     */
+    public function addExpert()
+    {
+        $data = $this->request->post();
+        $name  = trim(isset($data['name']) ? $data['name'] : '');
+        $email = trim(isset($data['email']) ? $data['email'] : '');
+
+        if ($name === '' || $email === '') {
+            return jsonError('name和email不能为空');
+        }
+
+        $exists = Db::name('expert')->where('email', $email)->find();
+        if ($exists) {
+            return jsonError('该邮箱已存在，expert_id=' . $exists['expert_id']);
+        }
+
+        $insert = [
+            'name'        => $name,
+            'email'       => $email,
+            'affiliation' => trim(isset($data['affiliation']) ? $data['affiliation'] : ''),
+            'source'      => trim(isset($data['source']) ? $data['source'] : 'manual'),
+            'ctime'       => time(),
+            'ltime'       => 0,
+            'state'       => 0,
+        ];
+
+        $expertId = Db::name('expert')->insertGetId($insert);
+
+        if (!empty($data['fields'])) {
+            $this->saveExpertFields($expertId, $data['fields']);
+        }
+
+        return jsonSuccess(['expert_id' => $expertId]);
+    }
+
+    /**
+     * 编辑专家
+     */
+    public function editExpert()
+    {
+        $data     = $this->request->post();
+        $expertId = intval(isset($data['expert_id']) ? $data['expert_id'] : 0);
+        if (!$expertId) {
+            return jsonError('expert_id is required');
+        }
+
+        $expert = Db::name('expert')->where('expert_id', $expertId)->find();
+        if (!$expert) {
+            return jsonError('专家不存在');
+        }
+
+        $update = [];
+        if (isset($data['name']))        $update['name']        = trim($data['name']);
+        if (isset($data['email']))       $update['email']       = trim($data['email']);
+        if (isset($data['affiliation'])) $update['affiliation'] = trim($data['affiliation']);
+        if (isset($data['source']))      $update['source']      = trim($data['source']);
+        if (isset($data['state']))       $update['state']       = intval($data['state']);
+
+        if (!empty($update)) {
+            Db::name('expert')->where('expert_id', $expertId)->update($update);
+        }
+
+        if (isset($data['fields'])) {
+            Db::name('expert_field')->where('expert_id', $expertId)->where('state', 0)->update(['state' => 1]);
+            if (!empty($data['fields'])) {
+                $this->saveExpertFields($expertId, $data['fields']);
+            }
+        }
+
+        return jsonSuccess(['expert_id' => $expertId]);
+    }
+
+    /**
+     * 批量修改状态
+     *
+     * 参数:
+     *   expert_ids - 逗号分隔的ID列表 "1,2,3"
+     *   state      - 目标状态 0-5
+     */
+    public function updateState()
+    {
+        $data      = $this->request->post();
+        $expertIds = isset($data['expert_ids']) ? $data['expert_ids'] : '';
+        $state     = intval(isset($data['state']) ? $data['state'] : -1);
+
+        if (empty($expertIds)) {
+            return jsonError('expert_ids is required');
+        }
+        if ($state < 0 || $state > 5) {
+            return jsonError('state取值范围0-5');
+        }
+
+        $ids   = array_map('intval', explode(',', $expertIds));
+        $count = Db::name('expert')->where('expert_id', 'in', $ids)->update(['state' => $state]);
+
+        return jsonSuccess(['updated' => $count]);
+    }
+
+    /**
+     * 删除专家（软删除，设为黑名单状态5）
+     *
+     * 参数:
+     *   expert_ids - 逗号分隔的ID列表
+     *   hard       - 传1则物理删除
+     */
+    public function deleteExpert()
+    {
+        $data      = $this->request->post();
+        $expertIds = isset($data['expert_ids']) ? $data['expert_ids'] : '';
+        $hard      = intval(isset($data['hard']) ? $data['hard'] : 0);
+
+        if (empty($expertIds)) {
+            return jsonError('expert_ids is required');
+        }
+
+        $ids = array_map('intval', explode(',', $expertIds));
+
+        if ($hard) {
+            Db::name('expert_field')->where('expert_id', 'in', $ids)->delete();
+            $count = Db::name('expert')->where('expert_id', 'in', $ids)->delete();
+        } else {
+            $count = Db::name('expert')->where('expert_id', 'in', $ids)->update(['state' => 5]);
+        }
+
+        return jsonSuccess(['affected' => $count]);
+    }
+
+    /**
+     * 给专家添加领域
+     */
+    public function addField()
+    {
+        $data     = $this->request->post();
+        $expertId = intval(isset($data['expert_id']) ? $data['expert_id'] : 0);
+        $field    = trim(isset($data['field']) ? $data['field'] : '');
+
+        if (!$expertId || $field === '') {
+            return jsonError('expert_id和field不能为空');
+        }
+
+        $exists = Db::name('expert_field')
+            ->where('expert_id', $expertId)
+            ->where('field', $field)
+            ->where('state', 0)
+            ->find();
+        if ($exists) {
+            return jsonError('该领域已存在');
+        }
+
+        $id = Db::name('expert_field')->insertGetId([
+            'expert_id' => $expertId,
+            'field'     => $field,
+            'state'     => 0,
+        ]);
+
+        return jsonSuccess(['expert_field_id' => $id]);
+    }
+
+    /**
+     * 删除领域（软删除）
+     */
+    public function removeField()
+    {
+        $efId = intval($this->request->param('expert_field_id', 0));
+        if (!$efId) {
+            return jsonError('expert_field_id is required');
+        }
+
+        // 若 t_expert_field 同时存“领域行 + 论文行”，删除领域时应同时软删该领域下所有论文行
+        $row = Db::name('expert_field')->where('expert_field_id', $efId)->find();
+        if ($row) {
+            Db::name('expert_field')
+                ->where('expert_id', intval($row['expert_id']))
+                ->where('field', (string)$row['field'])
+                ->update(['state' => 1]);
+        } else {
+            Db::name('expert_field')->where('expert_field_id', $efId)->update(['state' => 1]);
+        }
+
+        return jsonSuccess([]);
+    }
+
+    /**
+     * 获取所有不重复的领域列表（用于筛选下拉框）
+     */
+    public function getFieldOptions()
+    {
+        $list = Db::name('expert_field')
+            ->where('state', 0)
+            ->group('field')
+            ->column('field');
+
+        return jsonSuccess($list);
+    }
+
+    /**
+     * 获取所有来源列表（用于筛选下拉框）
+     */
+    public function getSourceOptions()
+    {
+        $list = Db::name('expert')
+            ->where('source', '<>', '')
+            ->group('source')
+            ->column('source');
+
+        return jsonSuccess($list);
+    }
+
+    /**
+     * 导出某个领域的专家为Excel
+     *
+     * 参数:
+     *   field      - 领域关键词（必填）
+     *   major_id   - 学科ID（可选）
+     *   state      - 状态筛选（可选，默认不过滤）
+     *   keyword    - 搜索姓名/邮箱/单位
+     *   source     - 来源筛选
+     */
+    public function exportExcel()
+    {
+        $data    = $this->request->param();
+        $field   = trim(isset($data['field']) ? $data['field'] : '');
+        $state   = isset($data['state']) ? $data['state'] : '-1';
+        $keyword = trim(isset($data['keyword']) ? $data['keyword'] : '');
+        $source  = trim(isset($data['source']) ? $data['source'] : '');
+
+        $query = Db::name('expert')->alias('e');
+
+        if ($field !== '') {
+            $query->join('t_expert_field ef', 'ef.expert_id = e.expert_id AND ef.state = 0', 'inner');
+            if ($field !== '') {
+                $query->where('ef.field', 'like', '%' . $field . '%');
+            }
+            $query->group('e.expert_id');
+        }
+
+        if ($state !== '-1' && $state !== '') {
+            $query->where('e.state', intval($state));
+        }
+        if ($keyword !== '') {
+            $query->where('e.name|e.email|e.affiliation', 'like', '%' . $keyword . '%');
+        }
+        if ($source !== '') {
+            $query->where('e.source', $source);
+        }
+
+        $list = $query->field('e.*')->order('e.ctime desc')->select();
+
+        if (empty($list)) {
+            return jsonError('没有符合条件的数据可导出');
+        }
+
+        $expertIds = array_column($list, 'expert_id');
+        $allFields = Db::name('expert_field')
+            ->where('expert_id', 'in', $expertIds)
+            ->where('state', 0)
+            ->select();
+
+        $fieldMap = [];
+        foreach ($allFields as $f) {
+            $fieldMap[$f['expert_id']][] = $f['field'];
+        }
+
+        vendor("PHPExcel.PHPExcel");
+
+        $objPHPExcel = new \PHPExcel();
+        $sheet = $objPHPExcel->getActiveSheet();
+        $sheet->setTitle('Expert List');
+
+        $headers = [
+            'A' => '#',
+            'B' => 'Name',
+            'C' => 'Email',
+            'D' => 'Affiliation',
+            'E' => 'Source',
+            'F' => 'Fields',
+            'G' => 'State',
+            'H' => 'Add Time',
+            'I' => 'Last Promotion',
+        ];
+        foreach ($headers as $col => $header) {
+            $sheet->setCellValue($col . '1', $header);
+        }
+
+        $headerStyle = [
+            'font'      => ['bold' => true, 'color' => ['rgb' => 'FFFFFF']],
+            'fill'      => ['type' => \PHPExcel_Style_Fill::FILL_SOLID, 'startcolor' => ['rgb' => '4472C4']],
+            'alignment' => ['horizontal' => \PHPExcel_Style_Alignment::HORIZONTAL_CENTER],
+        ];
+        $sheet->getStyle('A1:I1')->applyFromArray($headerStyle);
+
+        foreach ($list as $i => $item) {
+            $row = $i + 2;
+            $fields = isset($fieldMap[$item['expert_id']]) ? implode(', ', $fieldMap[$item['expert_id']]) : '';
+            $stateText = isset($this->stateMap[$item['state']]) ? $this->stateMap[$item['state']] : '未知';
+
+            $sheet->setCellValue('A' . $row, $i + 1);
+            $sheet->setCellValue('B' . $row, $item['name']);
+            $sheet->setCellValueExplicit('C' . $row, $item['email'], \PHPExcel_Cell_DataType::TYPE_STRING);
+            $sheet->setCellValue('D' . $row, $item['affiliation']);
+            $sheet->setCellValue('E' . $row, $item['source']);
+            $sheet->setCellValue('F' . $row, $fields);
+            $sheet->setCellValue('G' . $row, $stateText);
+            $sheet->setCellValue('H' . $row, $item['ctime'] ? date('Y-m-d H:i:s', $item['ctime']) : '');
+            $sheet->setCellValue('I' . $row, $item['ltime'] ? date('Y-m-d H:i:s', $item['ltime']) : '');
+        }
+
+        $sheet->getColumnDimension('A')->setWidth(6);
+        $sheet->getColumnDimension('B')->setWidth(25);
+        $sheet->getColumnDimension('C')->setWidth(35);
+        $sheet->getColumnDimension('D')->setWidth(40);
+        $sheet->getColumnDimension('E')->setWidth(15);
+        $sheet->getColumnDimension('F')->setWidth(50);
+        $sheet->getColumnDimension('G')->setWidth(15);
+        $sheet->getColumnDimension('H')->setWidth(20);
+        $sheet->getColumnDimension('I')->setWidth(20);
+
+        $label = $field !== '' ? preg_replace('/[^a-zA-Z0-9_\x{4e00}-\x{9fa5}]/u', '_', $field) : 'all';
+        $filename = 'expert_' . $label . '_' . date('Ymd_His') . '.xlsx';
+
+        $dir = ROOT_PATH . 'public' . DS . 'exports';
+        if (!is_dir($dir)) {
+            mkdir($dir, 0777, true);
+        }
+
+        $filepath = $dir . DS . $filename;
+        $writer = \PHPExcel_IOFactory::createWriter($objPHPExcel, 'Excel2007');
+        $writer->save($filepath);
+
+        return jsonSuccess([
+            'file_url'  => '/exports/' . $filename,
+            'file_name' => $filename,
+            'count'     => count($list),
+        ]);
+    }
+
+    /**
+     * 批量保存专家领域
+     * @param int $expertId
+     * @param array $fields [{"field":"xxx"}, ...]
+     */
+    private function saveExpertFields($expertId, $fields)
+    {
+        if (is_string($fields)) {
+            $fields = json_decode($fields, true);
+        }
+        if (!is_array($fields)) {
+            return;
+        }
+
+        foreach ($fields as $f) {
+            $fieldName = trim(isset($f['field']) ? $f['field'] : '');
+            if ($fieldName === '') continue;
+
+            $exists = Db::name('expert_field')
+                ->where('expert_id', $expertId)
+                ->where('field', $fieldName)
+                ->where('state', 0)
+                ->find();
+            if ($exists) continue;
+
+            Db::name('expert_field')->insert([
+                'expert_id' => $expertId,
+                'field'     => $fieldName,
+                'state'     => 0,
+            ]);
+        }
+    }
+
+    // ==================== Expert Fetch Field Management ====================
+
+    /**
+     * 获取抓取领域列表
+     * 参数: state(-1不过滤), keyword(搜索field), pageIndex, pageSize
+     */
+    public function getFetchList()
+    {
+        $data     = $this->request->param();
+        $state    = isset($data['state']) ? $data['state'] : '-1';
+        $keyword  = trim(isset($data['keyword']) ? $data['keyword'] : '');
+        $page     = max(1, intval(isset($data['pageIndex']) ? $data['pageIndex'] : 1));
+        $pageSize = max(1, intval(isset($data['pageSize']) ? $data['pageSize'] : 20));
+
+        $query = Db::name('expert_fetch');
+
+        if ($state !== '-1' && $state !== '') {
+            $query->where('state', intval($state));
+        }
+        if ($keyword !== '') {
+            $query->where('field', 'like', '%' . $keyword . '%');
+        }
+
+        $countQuery = Db::name('expert_fetch');
+        if ($state !== '-1' && $state !== '') {
+            $countQuery->where('state', intval($state));
+        }
+        if ($keyword !== '') {
+            $countQuery->where('field', 'like', '%' . $keyword . '%');
+        }
+
+        $total = $countQuery->count();
+        $list  = $query->order('expert_fetch_id desc')->page($page, $pageSize)->select();
+
+        foreach ($list as &$item) {
+            $item['last_time_text'] = $item['last_time'] ? date('Y-m-d H:i:s', $item['last_time']) : '';
+            $item['ctime_text']     = $item['ctime'] ? date('Y-m-d H:i:s', $item['ctime']) : '';
+            $fieldName = trim($item['field']);
+
+            $item['journal_count'] = Db::name('journal_promotion_field')
+                ->where('expert_fetch_id', $item['expert_fetch_id'])
+                ->where('state', 0)
+                ->count();
+
+            // 与推广选人一致：按字段关键词匹配，统计可用专家（去重）
+            $item['expert_count'] = Db::name('expert_field')->alias('ef')
+                ->join('t_expert e', 'e.expert_id = ef.expert_id', 'inner')
+                ->where('ef.state', 0)
+                ->where('e.state', 0)
+                ->where('ef.field', 'like', '%' . $fieldName . '%')
+                ->count('distinct ef.expert_id');
+
+
+        }
+
+        return jsonSuccess([
+            'list'       => $list,
+            'total'      => $total,
+            'pageIndex'  => $page,
+            'pageSize'   => $pageSize,
+            'totalPages' => $total > 0 ? ceil($total / $pageSize) : 0,
+        ]);
+    }
+
+    /**
+     * 新增抓取领域
+     * 参数: field(必填), source(选填,默认pubmed)
+     */
+    public function addFetchField()
+    {
+        $data   = $this->request->post();
+        $field  = trim(isset($data['field']) ? $data['field'] : '');
+        $source = trim(isset($data['source']) ? $data['source'] : 'pubmed');
+
+        if ($field === '') {
+            return jsonError('field不能为空');
+        }
+
+        $exists = Db::name('expert_fetch')
+            ->where('field', $field)
+            ->where('source', $source)
+            ->find();
+        if ($exists) {
+            if ($exists['state'] == 1) {
+                Db::name('expert_fetch')
+                    ->where('expert_fetch_id', $exists['expert_fetch_id'])
+                    ->update(['state' => 0]);
+                return jsonSuccess(['expert_fetch_id' => $exists['expert_fetch_id'], 'msg' => 'reactivated']);
+            }
+            return jsonError('该领域已存在 (expert_fetch_id=' . $exists['expert_fetch_id'] . ')');
+        }
+
+        $id = Db::name('expert_fetch')->insertGetId([
+            'field'       => mb_substr($field, 0, 128),
+            'source'      => mb_substr($source, 0, 128),
+            'last_page'   => 0,
+            'total_pages' => 0,
+            'last_time'   => 0,
+            'state'       => 0,
+            'ctime'       => time(),
+        ]);
+
+        return jsonSuccess(['expert_fetch_id' => $id]);
+    }
+
+    /**
+     * 编辑抓取领域
+     * 参数: expert_fetch_id(必填), field, source, state
+     */
+    public function editFetchField()
+    {
+        $data = $this->request->post();
+        $id   = intval(isset($data['expert_fetch_id']) ? $data['expert_fetch_id'] : 0);
+        if (!$id) {
+            return jsonError('expert_fetch_id is required');
+        }
+
+        $record = Db::name('expert_fetch')->where('expert_fetch_id', $id)->find();
+        if (!$record) {
+            return jsonError('记录不存在');
+        }
+
+        $update = [];
+        if (isset($data['field']))  $update['field']  = mb_substr(trim($data['field']), 0, 128);
+        if (isset($data['source'])) $update['source'] = mb_substr(trim($data['source']), 0, 128);
+        if (isset($data['state']))  $update['state']  = intval($data['state']);
+
+        if (!empty($update)) {
+            Db::name('expert_fetch')->where('expert_fetch_id', $id)->update($update);
+        }
+
+        return jsonSuccess(['expert_fetch_id' => $id]);
+    }
+
+    /**
+     * 删除/停用抓取领域（软删除 state=1）
+     * 参数: expert_fetch_id(必填), hard(传1物理删除)
+     */
+    public function deleteFetchField()
+    {
+        $data = $this->request->post();
+        $id   = intval(isset($data['expert_fetch_id']) ? $data['expert_fetch_id'] : 0);
+        $hard = intval(isset($data['hard']) ? $data['hard'] : 0);
+
+        if (!$id) {
+            return jsonError('expert_fetch_id is required');
+        }
+
+        if ($hard) {
+            Db::name('journal_promotion_field')->where('expert_fetch_id', $id)->delete();
+            Db::name('expert_fetch')->where('expert_fetch_id', $id)->delete();
+        } else {
+            Db::name('expert_fetch')->where('expert_fetch_id', $id)->update(['state' => 1]);
+        }
+
+        return jsonSuccess([]);
+    }
+
+    /**
+     * 后台手工将专家标记为退订
+     * 参数: expert_id(必填) 或 expert_ids(逗号分隔批量)
+     */
+    public function unsubscribe()
+    {
+        $data = $this->request->post();
+        $idsRaw = trim((string)(isset($data['expert_ids']) ? $data['expert_ids'] : (isset($data['expert_id']) ? $data['expert_id'] : '')));
+        if ($idsRaw === '') {
+            return jsonError('expert_id 或 expert_ids 必填');
+        }
+        $ids = array_filter(array_map('intval', explode(',', $idsRaw)));
+        if (empty($ids)) {
+            return jsonError('expert_id 无效');
+        }
+
+        $affected = Db::name('expert')
+            ->where('expert_id', 'in', $ids)
+            ->where('unsubscribed', 0)
+            ->update(['unsubscribed' => 1]);
+
+        return jsonSuccess([
+            'requested' => count($ids),
+            'affected'  => intval($affected),
+        ]);
+    }
+
+    /**
+     * 后台手工恢复专家订阅状态
+     * 参数: expert_id(必填) 或 expert_ids(逗号分隔批量)
+     */
+    public function resubscribe()
+    {
+        $data = $this->request->post();
+        $idsRaw = trim((string)(isset($data['expert_ids']) ? $data['expert_ids'] : (isset($data['expert_id']) ? $data['expert_id'] : '')));
+        if ($idsRaw === '') {
+            return jsonError('expert_id 或 expert_ids 必填');
+        }
+        $ids = array_filter(array_map('intval', explode(',', $idsRaw)));
+        if (empty($ids)) {
+            return jsonError('expert_id 无效');
+        }
+
+        $affected = Db::name('expert')
+            ->where('expert_id', 'in', $ids)
+            ->where('unsubscribed', 1)
+            ->update(['unsubscribed' => 0]);
+
+        return jsonSuccess([
+            'requested' => count($ids),
+            'affected'  => intval($affected),
+        ]);
+    }
+}

application/api/controller/Journal.php

@@ -333,6 +333,17 @@ class Journal extends Base {
             $update['wechat_app_secret'] =  $data['wechat_app_secret'];
             $aJournalUpdate['wechat_app_secret'] = $update['wechat_app_secret'];
         }
+        if (isset($data['wechat_yboard_qrcode'])){
+            $update['wechat_yboard_qrcode'] =  $data['wechat_yboard_qrcode'];
+        }
+
+        if(isset($data['editor_name'])&&$data['editor_name']!=''){
+            $update['editor_name'] = $data['editor_name'];
+        }
+
+        if(isset($data['databases'])&&$data['databases']!=''){
+            $update['databases'] = $data['databases'];
+        }
 
         if(!empty($aJournalUpdate)){
             $aJournalUpdate['issn'] = $journal_info['issn'];
@@ -440,6 +451,19 @@ class Journal extends Base {
         }
     }
 
+    public function uploadYboardQrcode()
+{
+    $file = request()->file('qrcode_url');
+    if ($file) {
+        $info = $file->move(ROOT_PATH . 'public' . DS . 'journalyboardqrcode');
+        if ($info) {
+            return json(['code' => 0, 'upurl' => str_replace("\\", "/", $info->getSaveName())]);
+        } else {
+            return json(['code' => 1, 'msg' => $file->getError()]);
+        }
+    }
+}
+
     /**
      * 获取微信公众号相关数量
      */

application/api/controller/MailTemplate.php

@@ -0,0 +1,261 @@
+<?php
+
+namespace app\api\controller;
+
+use think\Db;
+use think\Validate;
+
+class MailTemplate extends Base
+{
+    /**
+     * Create or update a mail template (V2)
+     * POST fields:
+     * - template_id (optional)
+     * - journal_id, scene, language, title, subject, body_html, body_text, variables_json, version, is_active
+     */
+    public function saveTemplate()
+    {
+        $data = $this->request->post();
+
+        $rule = new Validate([
+            'journal_id' => 'require|number',
+            'scene'      => 'require',
+            'language'   => 'require',
+            'title'      => 'require',
+            'subject'    => 'require',
+        ]);
+        if (!$rule->check($data)) {
+            return jsonError($rule->getError());
+        }
+
+        $payload = [
+            'journal_id'     => intval($data['journal_id']),
+            'scene'          => trim($data['scene']),
+            'language'       => trim($data['language']),
+            'title'          => trim($data['title']),
+            'subject'        => trim($data['subject']),
+            'body_html'      => $data['body_html'] ?? '',
+            'body_text'      => $data['body_text'] ?? '',
+            'variables_json' => $data['variables_json'] ?? '',
+            'version'        => intval($data['version'] ?? 1),
+            'is_active'      => intval($data['is_active'] ?? 1),
+            'utime'          => time(),
+        ];
+
+        $templateId = intval($data['template_id'] ?? 0);
+        if ($templateId) {
+            $exists = Db::name('mail_template')->where('template_id', $templateId)->where('state', 0)->find();
+            if (!$exists) {
+                return jsonError('Template not found');
+            }
+            Db::name('mail_template')->where('template_id', $templateId)->update($payload);
+            return jsonSuccess(['template_id' => $templateId]);
+        }
+
+        $payload['ctime'] = time();
+        $payload['state'] = 0;
+        $newId = Db::name('mail_template')->insertGetId($payload);
+        return jsonSuccess(['template_id' => $newId]);
+    }
+
+    /**
+     * Soft delete a template
+     * Params: template_id
+     */
+    public function deleteTemplate()
+    {
+        $templateId = intval($this->request->param('template_id', 0));
+        if (!$templateId) {
+            return jsonError('template_id is required');
+        }
+
+        Db::name('mail_template')->where('template_id', $templateId)->update(['state' => 1, 'utime' => time()]);
+        return jsonSuccess([]);
+    }
+
+    /**
+     * Get template detail
+     */
+    public function getTemplate()
+    {
+        $templateId = intval($this->request->param('template_id', 0));
+        if (!$templateId) {
+            return jsonError('template_id is required');
+        }
+
+        $tpl = Db::name('mail_template')->where('template_id', $templateId)->where('state', 0)->find();
+        if (!$tpl) {
+            return jsonError('Template not found');
+        }
+        return jsonSuccess(['template' => $tpl]);
+    }
+
+    /**
+     * List templates
+     * Params: journal_id, scene(optional), language(optional), is_active(optional)
+     */
+    public function listTemplates()
+    {
+        $journalId = intval($this->request->param('journal_id', 0));
+        if (!$journalId) {
+            return jsonError('journal_id is required');
+        }
+
+        $scene    = trim($this->request->param('scene', ''));
+        $language = trim($this->request->param('language', ''));
+        $isActive = $this->request->param('is_active', '');
+
+        $where = ['journal_id' => $journalId, 'state' => 0];
+        if ($scene !== '') $where['scene'] = $scene;
+        if ($language !== '') $where['language'] = $language;
+        if ($isActive !== '') $where['is_active'] = intval($isActive);
+
+        $list = Db::name('mail_template')
+            ->where($where)
+            ->order('is_active desc, utime desc, template_id desc')
+            ->select();
+
+        return jsonSuccess(['list' => $list]);
+    }
+
+    public function listTemplatesAll(){
+        $list = Db::name('mail_template')
+            ->where('state', 0)
+            ->order('is_active desc, utime desc, template_id desc')
+            ->select();
+        return jsonSuccess(['list'=>$list]);
+    }
+
+    /**
+     * Create or update a global mail style
+     * 当前 style 表字段：
+     *  - style_id, name, description, header_html, footer_html, state, ctime
+     */
+    public function saveStyle()
+    {
+        $data = $this->request->post();
+        $rule = new Validate([
+            'name' => 'require',
+        ]);
+        if (!$rule->check($data)) {
+            return jsonError($rule->getError());
+        }
+
+        $payload = [
+            'name'        => trim($data['name']),
+            'description' => trim($data['description'] ?? ''),
+            'header_html' => $data['header_html'] ?? '',
+            'footer_html' => $data['footer_html'] ?? '',
+            'state'       => 0,
+        ];
+
+        $styleId = intval($data['style_id'] ?? 0);
+        if ($styleId) {
+            $exists = Db::name('mail_style')->where('style_id', $styleId)->where('state', 0)->find();
+            if (!$exists) {
+                return jsonError('Style not found');
+            }
+            Db::name('mail_style')->where('style_id', $styleId)->update($payload);
+            return jsonSuccess(['style_id' => $styleId]);
+        }
+
+        $payload['ctime'] = time();
+        $newId = Db::name('mail_style')->insertGetId($payload);
+        return jsonSuccess(['style_id' => $newId]);
+    }
+
+    /**
+     * Delete a global mail style (soft delete)
+     */
+    public function deleteStyle()
+    {
+        $styleId = intval($this->request->param('style_id', 0));
+        if (!$styleId) {
+            return jsonError('style_id is required');
+        }
+
+        Db::name('mail_style')->where('style_id', $styleId)->update(['state' => 1]);
+        return jsonSuccess([]);
+    }
+
+    /**
+     * Get style detail
+     */
+    public function getStyle()
+    {
+        $styleId = intval($this->request->param('style_id', 0));
+        if (!$styleId) {
+            return jsonError('style_id is required');
+        }
+
+        $style = Db::name('mail_style')->where('style_id', $styleId)->where('state', 0)->find();
+        if (!$style) {
+            return jsonError('Style not found');
+        }
+        return jsonSuccess(['style' => $style]);
+    }
+
+    /**
+     * List styles
+     * 现在样式不再按 scene / language 区分，只返回全部正常状态的样式
+     */
+    public function listStyles()
+    {
+        $where = ['state' => 0];
+
+        $list = Db::name('mail_style')
+            ->where($where)
+            ->order('style_id desc')
+            ->select();
+
+        return jsonSuccess(['list' => $list]);
+    }
+
+    /**
+     * Render preview for a template预览效果，没啥用
+     * Params: template_id, vars (json string)
+     */
+    public function preview()
+    {
+        $templateId = intval($this->request->param('template_id', 0));
+        $varsJson   = $this->request->param('vars', '');
+        if (!$templateId) {
+            return jsonError('template_id is required');
+        }
+
+        $tpl = Db::name('mail_template')->where('template_id', $templateId)->where('state', 0)->find();
+        if (!$tpl) {
+            return jsonError('Template not found');
+        }
+
+        $vars = [];
+        if ($varsJson) {
+            $decoded = json_decode($varsJson, true);
+            if (is_array($decoded)) $vars = $decoded;
+        }
+
+        $subject = $this->render($tpl['subject'], $vars);
+        $body    = $this->render($tpl['body_html'], $vars);
+
+        // For preview we do not enforce a style; caller can combine with style if needed
+        return jsonSuccess(['subject' => $subject, 'body' => $body]);
+    }
+
+    private function render($tpl, $vars)
+    {
+        if (!is_string($tpl) || empty($tpl)) return '';
+        if (!is_array($vars) || empty($vars)) return $tpl;
+
+        $replace = [];
+        foreach ($vars as $k => $v) {
+            $key = trim((string)$k);
+            if ($key === '') continue;
+            $replace['{{' . $key . '}}'] = (string)$v;
+            // backward compatible placeholders
+            $replace['{' . $key . '}'] = (string)$v;
+        }
+
+        return str_replace(array_keys($replace), array_values($replace), $tpl);
+    }
+}
+

application/api/controller/Plagiarism.php

@@ -0,0 +1,226 @@
+<?php
+
+namespace app\api\controller;
+
+use think\Db;
+use think\Response;
+use app\common\PlagiarismService;
+
+/**
+ * 论文查重（Turnitin / Crossref Similarity Check）控制器。
+ *
+ * 触发方式：纯手工（编辑后台点"查重"按钮）。
+ * 报告策略：在线 viewer URL 临时签名 + PDF 永久落盘 runtime/plagiarism/。
+ *
+ * 主要接口:
+ *   POST submit              触发查重
+ *   GET  getStatus           轮询单条查重状态（前端 ajax）
+ *   GET  getList             列出某 article 的全部查重记录
+ *   GET  getReportUrl        获取/刷新在线查看 URL
+ *   GET  downloadReport      下载本地 PDF
+ *   POST retry               重新触发（创建新行）
+ *   GET  features            探活（开发调试用）
+ */
+class Plagiarism extends Base
+{
+    public function __construct(\think\Request $request = null)
+    {
+        parent::__construct($request);
+    }
+
+    /**
+     * 触发查重
+     *
+     * 入参:
+     *   article_id   必填
+     *   file_url     选填；不传则按 article_id 在 t_article_file 找 manuscirpt
+     *   editor_id    选填；触发人 user_id（前端拿不到也可以传 0）
+     */
+    public function submit()
+    {
+        $articleId = intval($this->request->param('article_id', 0));
+        $fileUrl   = trim($this->request->param('file_url', ''));
+        $editorId  = intval($this->request->param('editor_id', 0));
+
+        if ($articleId <= 0) {
+            return jsonError('article_id required');
+        }
+
+        try {
+            $svc = new PlagiarismService();
+            $localPath = $fileUrl !== ''
+                ? $svc->resolveFileUrlToLocal($fileUrl)
+                : $svc->locateArticleManuscript($articleId);
+            echo $localPath;
+            $checkId = $svc->submit($articleId, $localPath, $editorId, 'manual');
+            return jsonSuccess(['check_id' => $checkId]);
+        } catch (\Throwable $e) {
+            return jsonError($e->getMessage());
+        }
+    }
+
+
+    public function testccone(){
+        $svc = new PlagiarismService();
+        $checkId = 9;
+        $filePath = "/home/wwwroot/api.tmrjournals.com/public/manuscirpt/20260509/6832a56e8ace38fe99df390ab5221deb.docx";
+        $svc->runUploadAndTrigger($checkId,$filePath);
+
+    }
+
+    /**
+     * 重试 = 提交一次新查重（保留历史）
+     */
+    public function retry()
+    {
+        return $this->submit();
+    }
+
+    /**
+     * 取单条查重状态
+     */
+    public function getStatus()
+    {
+        $checkId = intval($this->request->param('check_id', 0));
+        if ($checkId <= 0) {
+            return jsonError('check_id required');
+        }
+        $row = Db::name('plagiarism_check')->where('check_id', $checkId)->find();
+        if (!$row) {
+            return jsonError('not found');
+        }
+        return jsonSuccess($this->formatRow($row));
+    }
+
+    /**
+     * 列出某 article 的全部查重记录（按时间倒序）
+     */
+    public function getList()
+    {
+        $articleId = intval($this->request->param('article_id', 0));
+        if ($articleId <= 0) {
+            return jsonError('article_id required');
+        }
+        $rows = Db::name('plagiarism_check')
+            ->where('article_id', $articleId)
+            ->order('check_id desc')
+            ->select();
+        $out = [];
+        foreach ($rows as $r) {
+            $out[] = $this->formatRow($r);
+        }
+        return jsonSuccess(['list' => $out]);
+    }
+
+    /**
+     * 取在线查看 URL；过期则自动刷新
+     */
+    public function getReportUrl()
+    {
+        $checkId = intval($this->request->param('check_id', 0));
+        if ($checkId <= 0) {
+            return jsonError('check_id required');
+        }
+        try {
+            $row = Db::name('plagiarism_check')->where('check_id', $checkId)->find();
+            if (!$row) {
+                return jsonError('not found');
+            }
+            if ($row['state'] != 3) {
+                return jsonError('check not completed yet, state=' . $row['state']);
+            }
+            $needRefresh = empty($row['view_only_url'])
+                || intval($row['view_only_url_expire']) < time() + 60;
+
+            if ($needRefresh) {
+                $svc = new PlagiarismService();
+                $info = $svc->refreshViewerUrlFor($checkId);
+                return jsonSuccess([
+                    'view_only_url' => $info['url'],
+                    'expire'        => $info['expire'],
+                ]);
+            }
+            return jsonSuccess([
+                'view_only_url' => $row['view_only_url'],
+                'expire'        => intval($row['view_only_url_expire']),
+            ]);
+        } catch (\Throwable $e) {
+            return jsonError($e->getMessage());
+        }
+    }
+
+    /**
+     * 直接吐 PDF 二进制流给浏览器下载
+     */
+    public function downloadReport()
+    {
+        $checkId = intval($this->request->param('check_id', 0));
+        if ($checkId <= 0) {
+            return jsonError('check_id required');
+        }
+        $row = Db::name('plagiarism_check')->where('check_id', $checkId)->find();
+        if (!$row || empty($row['pdf_local_path'])) {
+            return jsonError('report not available');
+        }
+        $rootDir = ROOT_PATH ?: dirname(dirname(dirname(__DIR__)));
+        $abs = rtrim($rootDir, '/\\') . DIRECTORY_SEPARATOR . str_replace(['/', '\\'], DIRECTORY_SEPARATOR, $row['pdf_local_path']);
+        if (!is_file($abs)) {
+            return jsonError('pdf file missing on disk: ' . $row['pdf_local_path']);
+        }
+        $filename = sprintf('plagiarism_check_%d_article_%d.pdf', $row['check_id'], $row['article_id']);
+        return Response::create(file_get_contents($abs), 'html', 200, [
+            'Content-Type'        => 'application/pdf',
+            'Content-Disposition' => 'attachment; filename="' . $filename . '"',
+            'Content-Length'      => (string)filesize($abs),
+        ]);
+    }
+
+    /**
+     * Turnitin 探活（开发调试用）
+     */
+    public function features()
+    {
+        try {
+            $tii = new \app\common\TurnitinService();
+            return jsonSuccess($tii->featuresEnabled());
+        } catch (\Throwable $e) {
+            return jsonError($e->getMessage());
+        }
+    }
+
+    // ---------- 内部 ----------
+
+    private function formatRow($r)
+    {
+        return [
+            'check_id'         => intval($r['check_id']),
+            'article_id'       => intval($r['article_id']),
+            'journal_id'       => intval($r['journal_id']),
+            'state'            => intval($r['state']),
+            'state_label'      => $this->stateLabel($r['state']),
+            'similarity_score' => floatval($r['similarity_score']),
+            'tii_report_status' => (string)$r['tii_report_status'],
+            'has_pdf'          => !empty($r['pdf_local_path']),
+            'has_viewer_url'   => !empty($r['view_only_url']) && intval($r['view_only_url_expire']) > time(),
+            'attempts'         => intval($r['attempts']),
+            'error_msg'        => (string)$r['error_msg'],
+            'source_file_name' => (string)$r['source_file_name'],
+            'trigger_source'   => (string)$r['trigger_source'],
+            'triggered_by'     => intval($r['triggered_by']),
+            'ctime'            => intval($r['ctime']),
+            'utime'            => intval($r['utime']),
+        ];
+    }
+
+    private function stateLabel($state)
+    {
+        $map = [
+            0 => '待上传',
+            1 => '上传中',
+            2 => '比对中',
+            3 => '完成',
+            4 => '失败',
+        ];
+        return isset($map[$state]) ? $map[$state] : 'unknown';
+    }
+}

application/api/controller/Preaccept.php

@@ -6,6 +6,8 @@ use think\Db;
 use think\Env;
 use think\Queue;
 use think\Validate;
+use app\common\CrossrefService;
+use app\common\ReferenceCheckService;
 
 class Preaccept extends Base
 {
@@ -14,6 +16,26 @@ class Preaccept extends Base
         parent::__construct($request);
     }
 
+    /**
+     * 新增/修改导致文献集合改变后，清空整篇校对明细，使文章状态回到"未校对"。
+     * 失败仅记日志，不阻塞主流程。
+     */
+    private function resetArticleChecksOnReferChange($pArticleId, $sourceTag = '')
+    {
+        $pArticleId = intval($pArticleId);
+        if ($pArticleId <= 0) {
+            return;
+        }
+        try {
+            (new ReferenceCheckService())->clearArticleChecksByPArticleId($pArticleId);
+        } catch (\Exception $e) {
+            \think\Log::error(
+                'resetArticleChecksOnReferChange[' . $sourceTag . '] p_article_id='
+                . $pArticleId . ' ' . $e->getMessage()
+            );
+        }
+    }
+
 
     /**获取文章参考文献列表
      * @return \think\response\Json
@@ -91,6 +113,7 @@ class Preaccept extends Base
             return jsonError($rule->getError());
         }
         $this->production_article_refer_obj->where('p_article_id',$data['p_article_id'])->update(["state"=>1]);
+        $this->resetArticleChecksOnReferChange(intval($data['p_article_id']), 'discardRefersByParticleid');
         return jsonSuccess([]);
     }
 
@@ -141,6 +164,7 @@ class Preaccept extends Base
         }
         $adId= $this->production_article_refer_obj->insertGetId($insert);
         $this->production_article_refer_obj->where('p_article_id', $p_info['p_article_id'])->where("p_refer_id", "<>", $adId)->where("index", ">", $pre_refer['index'])->where('state', 0)->setInc('index');
+        $this->resetArticleChecksOnReferChange(intval($p_info['p_article_id']), 'addRefer');
         return jsonSuccess([]);
 
 
@@ -197,6 +221,7 @@ class Preaccept extends Base
         }
         $adId= $this->production_article_refer_obj->insertGetId($insert);
         $this->production_article_refer_obj->where('p_article_id', $p_info['p_article_id'])->where("p_refer_id", "<>", $adId)->where("index", ">", $pre_refer['index'])->where('state', 0)->setInc('index');
+        $this->resetArticleChecksOnReferChange(intval($p_info['p_article_id']), 'addReferByParticleid');
         return jsonSuccess([]);
     }
 
@@ -232,6 +257,7 @@ class Preaccept extends Base
         $insert['cs'] = 1;
         $adId = $this->production_article_refer_obj->insertGetId($insert);
         $this->production_article_refer_obj->where('p_article_id', $p_info['p_article_id'])->where("p_refer_id", "<>", $adId)->where("index", ">", $pre_refer['index'])->where('state', 0)->setInc('index');
+        $this->resetArticleChecksOnReferChange(intval($p_info['p_article_id']), 'addReferNotdoi');
         return jsonSuccess([]);
     }
 
@@ -461,6 +487,17 @@ class Preaccept extends Base
 //        }
 //        $this->production_article_refer_obj->where('p_refer_id', $data['p_refer_id'])->update(['refer_doi' => $data['doi']]);
 //        my_doiToFrag2($this->production_article_refer_obj->where('p_refer_id', $data['p_refer_id'])->find());
+
+        //文献内容更新成功后异步重检该文献对应的全部校对明细（失败不阻塞主流程）
+        try {
+            (new ReferenceCheckService())->enqueueRecheckByPReferId(
+                intval($data['p_refer_id']),
+                intval($old_refer_info['p_article_id'])
+            );
+        } catch (\Exception $e) {
+            \think\Log::error('editRefer enqueueRecheckByPReferId p_refer_id=' . $data['p_refer_id'] . ' ' . $e->getMessage());
+        }
+
         return jsonSuccess([]);
     }
 
@@ -708,36 +745,66 @@ class Preaccept extends Base
     }
 
 
+    /**
+     * 通过 DOI 获取文献元数据（Crossref REST API）。
+     *
+     * POST 参数:
+     *   doi  必填，可为纯 DOI（10.xxxx/...）或 https://doi.org/10.xxxx/...
+     *
+     * 返回 data.formate 与旧版字段兼容: author, title, joura, dateno, doilink
+     * 另附 data.crossref: 原始摘要字段（不含 raw message，避免体积过大）
+     */
     public function searchDoi()
     {
         $data = $this->request->post();
         $rule = new Validate([
-            "doi" => "require"
+            'doi' => 'require',
         ]);
         if (!$rule->check($data)) {
             return jsonError($rule->getError());
         }
-        $doi = str_replace('/', '%2F', $data['doi']);
-//        $url = "https://citation.crosscite.org/format?doi=$doi&style=cancer-translational-medicine&lang=en-US";
-        $url = "https://citation.doi.org/format?doi=$doi&style=cancer-translational-medicine&lang=en-US";
-        $res = myGet($url);
-        $frag = trim(substr($res, strpos($res, '.') + 1));
-        if ($frag == "") {
-            return jsonError("not find");
+
+        $doiInput = trim((string)$data['doi']);
+        if ($doiInput === '') {
+            return jsonError('doi empty');
         }
-        if (mb_substr_count($frag, '.') != 3) {
-            return jsonError("formate fail");
+        // 去掉 URL 前缀，得到裸 DOI
+        $doiNorm = preg_replace('#^https?://(dx\.)?doi\.org/#i', '', $doiInput);
+        $doiNorm = trim($doiNorm, " \t\n\r\0\x0B/");
+
+        $svc = new CrossrefService([
+            'mailto' => trim((string)Env::get('crossref_mailto', '')),
+        ]);
+        $summary = $svc->fetchWorkSummary($doiNorm);
+        if ($summary === null || empty($summary['doi'])) {
+            return jsonError('DOI not found or invalid (Crossref)');
         }
-        $res = explode('.', $frag);
-        $f['author'] = prgeAuthor($res[0]);
-        $f['title'] = trim($res[1]);
-        $bj = bekjournal($res[2]);
-        $joura = formateJournal(trim($bj[0]));
-        $f['joura'] = $joura;
-        $f['dateno'] = str_replace(' ', '', str_replace('-', '–', trim($bj[1])));
-        $f['doilink'] = strpos($data['doi'], "http") === false ? "http://doi.org/" . $data['doi'] : $data['doi'];
-        $re['formate'] = $f;
-        return jsonSuccess($re);
+
+        $title = trim((string)($summary['title'] ?? ''));
+        $jouraRaw = trim((string)($summary['joura'] ?? ''));
+        $authorStr = trim((string)($summary['author_str'] ?? ''));
+        $dateno = trim((string)($summary['dateno'] ?? ''));
+        $doilink = trim((string)($summary['doilink'] ?? ''));
+        if ($doilink === '') {
+            $doilink = 'https://doi.org/' . $summary['doi'];
+        }
+
+        $f = [
+            'author'  => $authorStr !== '' ? prgeAuthor($authorStr) : '',
+            'title'   => $title,
+            'joura'   => $jouraRaw !== '' ? formateJournal($jouraRaw) : '',
+            'dateno'  => str_replace(' ', '', str_replace('-', '–', $dateno)),
+            'doilink' => $doilink,
+        ];
+
+        $crossrefOut = $summary;
+        unset($crossrefOut['raw']);
+
+        return jsonSuccess([
+            'formate'   => $f,
+            'crossref'  => $crossrefOut,
+            'doi'       => $summary['doi'],
+        ]);
     }
 
 
@@ -843,14 +910,14 @@ class Preaccept extends Base
         }
         $am_info = $this->article_main_obj->where("am_id",$data['am_id'])->find();
         //上一行，空行
-        $p_list = $this->article_main_obj->where("article_id",$am_info['article_id'])->where("sort","<",$am_info['sort'])->whereIn("state",[0,2])->order("sort desc")->limit(1)->select();
-        if($p_list&&($p_list[0]['type']>0||$p_list[0]['content']!="")){
-            $this->addBRow($am_info['article_id'],$p_list[0]['am_id']);
-        }
-        $n_list = $this->article_main_obj->where("article_id",$am_info['article_id'])->where("sort",">",$am_info['sort'])->whereIn("state",[0,2])->order("sort asc")->limit(1)->select();
-        if($n_list[0]['type']>0||$n_list[0]['content']!=""){
-            $this->addBRow($am_info['article_id'],$data['am_id']);
-        }
+//        $p_list = $this->article_main_obj->where("article_id",$am_info['article_id'])->where("sort","<",$am_info['sort'])->whereIn("state",[0,2])->order("sort desc")->limit(1)->select();
+//        if($p_list&&($p_list[0]['type']>0||$p_list[0]['content']!="")){
+//            $this->addBRow($am_info['article_id'],$p_list[0]['am_id']);
+//        }
+//        $n_list = $this->article_main_obj->where("article_id",$am_info['article_id'])->where("sort",">",$am_info['sort'])->whereIn("state",[0,2])->order("sort asc")->limit(1)->select();
+//        if($n_list[0]['type']>0||$n_list[0]['content']!=""){
+//            $this->addBRow($am_info['article_id'],$data['am_id']);
+//        }
         $this->article_main_obj->where("am_id",$data['am_id'])->update(["is_h1"=>1,"is_h2"=>0,"is_h3"=>0]);
         // return jsonSuccess([]);
         //返回数据 20260119 start
@@ -872,10 +939,10 @@ class Preaccept extends Base
         }
         $am_info = $this->article_main_obj->where("am_id",$data['am_id'])->find();
         //上一行，空行
-        $p_list = $this->article_main_obj->where("article_id",$am_info['article_id'])->where("sort","<",$am_info['sort'])->whereIn("state",[0,2])->order("sort desc")->limit(1)->select();
-        if($p_list&&($p_list[0]['type']>0||$p_list[0]['content']!="")){
-            $this->addBRow($am_info['article_id'],$p_list[0]['am_id']);
-        }
+//        $p_list = $this->article_main_obj->where("article_id",$am_info['article_id'])->where("sort","<",$am_info['sort'])->whereIn("state",[0,2])->order("sort desc")->limit(1)->select();
+//        if($p_list&&($p_list[0]['type']>0||$p_list[0]['content']!="")){
+//            $this->addBRow($am_info['article_id'],$p_list[0]['am_id']);
+//        }
         $this->article_main_obj->where("am_id",$data['am_id'])->update(["is_h1"=>0,"is_h2"=>1,"is_h3"=>0]);
         // return jsonSuccess([]);
         //返回数据 20260119 start
@@ -1131,6 +1198,9 @@ class Preaccept extends Base
             return jsonError($rule->getError());
         }
         $am_info = $this->article_main_obj->where("am_id",$data['am_id'])->find();
+        $old_content = $am_info['content'];
+        $new_raw_content = isset($data['content']) ? (string)$data['content'] : '';
+
         $insert['article_id'] = $am_info['article_id'];
         $insert['am_id'] = $data['am_id'];
         $insert['type'] = 0;
@@ -1140,11 +1210,21 @@ class Preaccept extends Base
         $insert['ctime'] = time();
         $this->article_main_log_obj->insert($insert);
 
+        // 判断是否存在“引用删除”（新 content 相对旧 content 缺少 <mycite>）
+        $hasCitationDeletion = $this->hasMyciteDeletion($old_content, $new_raw_content);
+
+        $update['content'] = $this->formatMain($new_raw_content);
+
+
 
 
-        $update['content'] = $this->formatMain($data['content']);
         $update['state'] = 0;
         $this->article_main_obj->where("am_id",$data['am_id'])->update($update);
+
+        // 若检测到引用删除，则进行全文扫描并标记未被引用条目的 is_used=0（含 table 内容）
+        if ($hasCitationDeletion) {
+            $this->markUnusedReferencesForArticle(intval($am_info['article_id']));
+        }
         // return jsonSuccess([]);
         //返回更新数据 20260119 start
         $update = empty($update) ? [] : array_merge($update,['am_id' => empty($data['am_id']) ? 0 : $data['am_id']]);
@@ -1155,6 +1235,46 @@ class Preaccept extends Base
         //返回更新数据 20260119 end
     }
 
+    /**
+     * 是否发生 <mycite> 删除（new 相对 old 少了任意引用 id）
+     */
+    private function hasMyciteDeletion(string $oldContent, string $newContent): bool
+    {
+        $oldIds = $this->extractMyciteIds($oldContent);
+        if (empty($oldIds)) {
+            return false;
+        }
+        $newIds = $this->extractMyciteIds($newContent);
+
+        // old 有引用，new 为空 => 删除
+        if (empty($newIds)) {
+            return true;
+        }
+        $oldSet = array_fill_keys($oldIds, 1);
+        $newSet = array_fill_keys($newIds, 1);
+        foreach ($oldSet as $id => $_) {
+            if (!isset($newSet[$id])) {
+                return true;
+            }
+        }
+        return false;
+    }
+
+    /**
+     * 从文本/HTML 中提取 <mycite data-id="1,2,3"> 的 id 列表
+     *
+     * @return int[]
+     */
+
+
+    /**
+     * 全文扫描（正文 + table），将 production_article_refer 未被引用的条目标记 is_used=0
+     * - 被引用的条目 is_used=1
+     *
+     * 注意：依赖 production_article_refer 表存在 is_used 字段（int/tinyint）
+     */
+
+
     public function getArticleMainsRecycle(){
         $data = $this->request->post();
         $rule = new Validate([
@@ -1369,6 +1489,7 @@ class Preaccept extends Base
             return jsonError($rule->getError());
         }
         $refer_info = $this->production_article_refer_obj->where('p_refer_id', $data['p_refer_id'])->find();
+        $sibling_p_refer_id = 0;
         if ($data['act'] == "up") {
             $up_info = $this->production_article_refer_obj->where('p_article_id', $refer_info['p_article_id'])->where('index', $refer_info['index'] - 1)->where('state', 0)->find();
             if (!$up_info) {
@@ -1376,6 +1497,7 @@ class Preaccept extends Base
             }
             $this->production_article_refer_obj->where('p_refer_id', $up_info['p_refer_id'])->setInc("index");
             $this->production_article_refer_obj->where('p_refer_id', $refer_info['p_refer_id'])->setDec("index");
+            $sibling_p_refer_id = intval($up_info['p_refer_id']);
         } else {
             $down_info = $this->production_article_refer_obj->where('p_article_id', $refer_info['p_article_id'])->where('index', $refer_info['index'] + 1)->where('state', 0)->find();
             if (!$down_info) {
@@ -1383,7 +1505,19 @@ class Preaccept extends Base
             }
             $this->production_article_refer_obj->where('p_refer_id', $refer_info['p_refer_id'])->setInc("index");
             $this->production_article_refer_obj->where('p_refer_id', $down_info['p_refer_id'])->setDec("index");
+            $sibling_p_refer_id = intval($down_info['p_refer_id']);
         }
+
+        // 仅同步本次交换的两条 p_refer_id 对应的校对明细 reference_no / refer_index
+        try {
+            (new ReferenceCheckService())->syncReferenceNoByPReferIds(
+                [intval($refer_info['p_refer_id']), $sibling_p_refer_id],
+                intval($refer_info['p_article_id'])
+            );
+        } catch (\Exception $e) {
+            \think\Log::error('sortRefer syncReferenceNoByPReferIds: ' . $e->getMessage());
+        }
+
         return jsonSuccess([]);
     }
 

application/api/controller/Production.php

@@ -10,6 +10,7 @@ use think\Db;
 use think\Queue;
 use think\Validate;
 use think\log;
+use app\common\ArticleSymbolNormalizer;
 
 /**
  * @title 公共管理相关
@@ -1380,6 +1381,10 @@ class Production extends Base
         return $html;
     }
 
+    public function testsym(){
+        ArticleSymbolNormalizer::normalize("");
+    }
+
 
     public function doTypeSettingNew()
     {
@@ -1399,7 +1404,7 @@ class Production extends Base
         $editor_info = $this->user_obj->where('user_id', $journal_info['editor_id'])->find();
 
         $typesetInfo = [];
-        $typesetInfo['info_title'] = $p_info['title'];
+        $typesetInfo['info_title'] = ArticleSymbolNormalizer::normalize($p_info['title']);
         $typesetInfo['info_type'] = $p_info['type'];
         $typesetInfo['doi'] = $p_info['doi'];
         $typesetInfo['topic'] = '';
@@ -1920,6 +1925,7 @@ class Production extends Base
             $cache_insert['p_article_id'] = $p_info['p_article_id'];
             $cache_insert['refer_content'] = $v;
             $cache_insert['index'] = $k;
+            $cache_insert['old_index'] = $k;
             $cache_insert['ctime'] = time();
             $this->production_article_refer_obj->insert($cache_insert);
         }

application/api/controller/PromotionFactory.php

@@ -0,0 +1,296 @@
+<?php
+
+namespace app\api\controller;
+
+use think\Db;
+use think\Validate;
+
+class PromotionFactory extends Base
+{
+    public function __construct(\think\Request $request = null)
+    {
+        parent::__construct($request);
+    }
+
+    /**
+     * 工厂列表
+     * 参数: journal_id, state(-1=全部), page, per_page
+     */
+    public function getList()
+    {
+        $journalId = intval($this->request->param('journal_id', 0));
+        $editor_id = $this->request->param("user_id",0);
+        $state     = $this->request->param('state', '-1');
+        $page      = max(1, intval($this->request->param('page', 1)));
+        $perPage   = max(1, min(intval($this->request->param('per_page', 20)), 100));
+
+//        if (!$journalId) {
+//            return jsonError('journal_id is required');
+//        }
+        if($journalId!==0){
+            $where[] = ['journal_id', '=', $journalId];
+        }else{
+            $journalIds = $this->journal_obj->where("editor_id",$editor_id)->where("state",0)->column("journal_id");
+            $where[] = ['journal_id',"in",$journalIds];
+        }
+
+        if ($state !== '-1' && $state !== '') {
+            $where[] = ['state', '=', intval($state)];
+        }
+
+        $countQuery = Db::name('promotion_factory');
+        foreach ($where as $w) {
+            $countQuery->where($w[0], $w[1], $w[2]);
+        }
+        $total = $countQuery->count();
+
+        $listQuery = Db::name('promotion_factory');
+        foreach ($where as $w) {
+            $listQuery->where($w[0], $w[1], $w[2]);
+        }
+        $list = $listQuery
+            ->order('promotion_factory_id desc')
+            ->page($page, $perPage)
+            ->select();
+
+
+
+        foreach ($list as &$item) {
+            $item['ctime_text'] = $item['ctime'] ? date('Y-m-d H:i:s', $item['ctime']) : '';
+            $item['fetch_fields'] = $this->resolveFetchFields($item['fetch_ids']);
+            $item['country_scope_label'] = $this->buildCountryScopeLabel(
+                isset($item['target_partitions']) ? (string)$item['target_partitions'] : '',
+                isset($item['target_country_ids']) ? (string)$item['target_country_ids'] : ''
+            );
+        }
+
+        return jsonSuccess([
+            'list'        => $list,
+            'total'       => $total,
+            'page'        => $page,
+            'per_page'    => $perPage,
+            'total_pages' => $total > 0 ? ceil($total / $perPage) : 0,
+        ]);
+    }
+
+    /**
+     * 工厂详情
+     * 参数: promotion_factory_id
+     */
+    public function getDetail()
+    {
+        $id = intval($this->request->param('promotion_factory_id', 0));
+        if (!$id) {
+            return jsonError('promotion_factory_id is required');
+        }
+
+        $row = Db::name('promotion_factory')->where('promotion_factory_id', $id)->find();
+        if (!$row) {
+            return jsonError('Factory not found');
+        }
+
+        $row['ctime_text'] = $row['ctime'] ? date('Y-m-d H:i:s', $row['ctime']) : '';
+        $row['fetch_fields'] = $this->resolveFetchFields($row['fetch_ids']);
+        $row['country_scope_label'] = $this->buildCountryScopeLabel(
+            isset($row['target_partitions']) ? (string)$row['target_partitions'] : '',
+            isset($row['target_country_ids']) ? (string)$row['target_country_ids'] : ''
+        );
+
+        return jsonSuccess($row);
+    }
+
+    /**
+     * 新增工厂
+     */
+    public function add()
+    {
+        $data = $this->request->post();
+
+        $journalId = intval(isset($data['journal_id']) ? $data['journal_id'] : 0);
+        if (!$journalId) {
+            return jsonError('journal_id is required');
+        }
+
+        $journal = Db::name('journal')->where('journal_id', $journalId)->where('state', 0)->find();
+        if (!$journal) {
+            return jsonError('Journal not found');
+        }
+
+        $now = time();
+        $id = Db::name('promotion_factory')->insertGetId([
+            'journal_id'         => $journalId,
+            'type'               => intval(isset($data['type']) ? $data['type'] : 0),
+            'expert_type'        => intval(isset($data['expert_type']) ? $data['expert_type'] : 5),
+            'email_ids'          => trim(isset($data['email_ids']) ? (string)$data['email_ids'] : ''),
+            'send_count'         => max(1, intval(isset($data['send_count']) ? $data['send_count'] : 100)),
+            'template_id'        => intval(isset($data['template_id']) ? $data['template_id'] : 0),
+            'style_id'           => intval(isset($data['style_id']) ? $data['style_id'] : 0),
+            'fetch_ids'          => trim(isset($data['fetch_ids']) ? (string)$data['fetch_ids'] : ''),
+            'target_partitions'  => trim(isset($data['target_partitions']) ? (string)$data['target_partitions'] : ''),
+            'target_country_ids' => trim(isset($data['target_country_ids']) ? (string)$data['target_country_ids'] : ''),
+            'state'              => 0,
+            'ctime'              => $now,
+        ]);
+
+        return jsonSuccess(['promotion_factory_id' => $id]);
+    }
+
+    public function changePromotionAct(){
+        $data = $this->request->post();
+        $rule = new Validate([
+            "promotion_factory_id"=>"require",
+            "start_promotion"=>"require"
+        ]);
+        if(!$rule->check($data)){
+            return jsonError($rule->getError());
+        }
+        Db::name("promotion_factory")->where("promotion_factory_id",$data['promotion_factory_id'])->update(['start_promotion'=>$data['start_promotion']]);
+        return jsonSuccess();
+    }
+
+    public function getCountForPromotionEmailIds(){
+        $data = $this->request->post();
+        $rule = new Validate([
+            "ids"=>"require"
+        ]);
+        if(!$rule->check($data)){
+            return jsonError($rule->getError());
+        }
+        $ids = explode(",",$data["ids"]);
+        $emails = Db::name("journal_email")->whereIn("j_email_id",$ids)->select();
+        $count = 0;
+        foreach ($emails as $item){
+            $count += $item["daily_limit"];
+        }
+        return jsonSuccess(['limit_count'=>$count]);
+    }
+
+    /**
+     * 编辑工厂
+     */
+    public function edit()
+    {
+        $data = $this->request->post();
+        $id   = intval(isset($data['promotion_factory_id']) ? $data['promotion_factory_id'] : 0);
+        if (!$id) {
+            return jsonError('promotion_factory_id is required');
+        }
+
+        $row = Db::name('promotion_factory')->where('promotion_factory_id', $id)->find();
+        if (!$row) {
+            return jsonError('Factory not found');
+        }
+
+        $update = [];
+        $allowedFields = [
+            'type', 'expert_type', 'email_ids', 'send_count',
+            'template_id', 'style_id', 'fetch_ids',
+            'target_partitions', 'target_country_ids',
+        ];
+        $intFields = ['type', 'expert_type', 'send_count', 'template_id', 'style_id'];
+
+        foreach ($allowedFields as $f) {
+            if (isset($data[$f])) {
+                $update[$f] = in_array($f, $intFields) ? intval($data[$f]) : trim((string)$data[$f]);
+            }
+        }
+
+        if (isset($update['send_count'])) {
+            $update['send_count'] = max(1, $update['send_count']);
+        }
+
+        if (empty($update)) {
+            return jsonError('没有可更新的字段');
+        }
+
+        Db::name('promotion_factory')->where('promotion_factory_id', $id)->update($update);
+
+        return jsonSuccess([]);
+    }
+
+    /**
+     * 删除工厂（软删除 state=1）
+     */
+    public function delete()
+    {
+        $id = intval($this->request->param('promotion_factory_id', 0));
+        if (!$id) {
+            return jsonError('promotion_factory_id is required');
+        }
+
+        $row = Db::name('promotion_factory')->where('promotion_factory_id', $id)->find();
+        if (!$row) {
+            return jsonError('Factory not found');
+        }
+
+        Db::name('promotion_factory')->where('promotion_factory_id', $id)->delete();
+
+        return jsonSuccess([]);
+    }
+
+    /**
+     * 启用/禁用切换
+     */
+    public function toggle()
+    {
+        $id = intval($this->request->param('promotion_factory_id', 0));
+        if (!$id) {
+            return jsonError('promotion_factory_id is required');
+        }
+
+        $row = Db::name('promotion_factory')->where('promotion_factory_id', $id)->find();
+        if (!$row) {
+            return jsonError('Factory not found');
+        }
+
+        $newState = ($row['state'] == 0) ? 1 : 0;
+        Db::name('promotion_factory')->where('promotion_factory_id', $id)->update(['state' => $newState]);
+
+        return jsonSuccess(['state' => $newState]);
+    }
+
+    /**
+     * 根据 fetch_ids 解析出领域名称列表
+     */
+    private function resolveFetchFields($fetchIds)
+    {
+        $fetchIds = trim((string)$fetchIds);
+        if ($fetchIds === '') {
+            return [];
+        }
+        $ids = array_map('intval', explode(',', $fetchIds));
+        $ids = array_filter($ids);
+        if (empty($ids)) {
+            return [];
+        }
+        return Db::name('expert_fetch')
+            ->where('expert_fetch_id', 'in', $ids)
+            ->column('field', 'expert_fetch_id');
+    }
+
+    /**
+     * 构建国家范围描述标签
+     */
+    private function buildCountryScopeLabel($targetPartitions, $targetCountryIds)
+    {
+        $tp = trim($targetPartitions);
+        $tc = trim($targetCountryIds);
+
+        if ($tp === '' && $tc === '') {
+            return '全部国家';
+        }
+
+        $parts = [];
+        if ($tp !== '') {
+            $parts[] = '分区' . $tp;
+        }
+        if ($tc !== '') {
+            $cids = array_map('intval', explode(',', $tc));
+            $names = Db::name('country')->where('country_id', 'in', $cids)->column('zh_name');
+            if (!empty($names)) {
+                $parts[] = implode(',', $names);
+            }
+        }
+        return implode(' + ', $parts);
+    }
+}

application/api/controller/Ucenter.php

@@ -5,6 +5,7 @@ namespace app\api\controller;
 use app\api\controller\Base;
 use think\Validate;
 use think\Queue;
+use think\Db;
 
 
 class Ucenter extends Base{
@@ -376,6 +377,14 @@ class Ucenter extends Base{
         if(isset($data['phone'])&&$data['phone']!=''){
             $update['phone'] = $data['phone'];
         }
+        $old = $this->user_obj->where(['user_id'=>$data['user_id']])->find();
+        if(isset($data['email'])&&$data['email']!=$old['email']){
+            $c = $this->user_obj->where("email",trim($data['email']))->find();
+            if($c){
+                return jsonError("Email already exists");
+            }
+            $update['email'] = trim($data['email']);
+        }
         $update['localname'] = isset($data['localname'])?trim($data['localname']):'';
         $this->user_obj->where(['user_id'=>$data['user_id']])->update($update);
         $updata1=[
@@ -742,6 +751,73 @@ class Ucenter extends Base{
         return $data;
     }
 
+    public function getApplyYboardForExpertBaseInfo(){
+        $data = $this->request->post();
+        $rule = new Validate([
+            "expert_id"=>"require",
+            "journal_id"=>"require"
+        ]);
+        if(!$rule->check($data)){
+            return jsonError($rule->getError());
+        }
+        $expert_info = Db::name("expert")->where("expert_id",$data['expert_id'])->find();
+        $journal_info = Db::name("journal")->field("journal_id,title,issn,abbr,journal_icon,wechat_yboard_qrcode")->where("journal_id",$data['journal_id'])->find();
+        $re['expert_info'] = $expert_info;
+        $re['journal_info'] = $journal_info;
+        return jsonSuccess($re);
+    }
+
+    public function submitApplyYboardForExpert(){
+        $data = $this->request->post();
+        $rule = new Validate([
+            "expert_id"=>"require",
+            "journal_id"=>"require",
+            "password"=>"require",
+            "cv"=>"require"
+        ]);
+        if(!$rule->check($data)){
+            return jsonError($rule->getError());
+        }
+        $expert_info = Db::name("expert")->where("expert_id",$data['expert_id'])->find();
+        $journal_info = Db::name("journal")->field("journal_id,title,issn,abbr,wechat_yboard_qrcode")->where("journal_id",$data['journal_id'])->find();
+        if(isset($data['email'])&&$data['email']!=""){
+            $email = $data['email'];
+        }else{
+            $email = $expert_info['email'];
+        }
+        if(isset($data['name'])&&$data['name']!=""){
+            $name = $data['name'];
+        }else{
+            $name = $expert_info['realname'];
+        }
+        $has = $this->user_obj->where("email",$email)->find();
+        if(!$has){
+            $check = $this->user_obj->where("account",$name)->find();
+            $add_user['account'] = $check?$email:$name;
+            $add_user['email'] = $email;
+            $add_user['password'] = md5($data['password']);
+            $add_user['realname'] = $name;
+            $add_user['ctime'] = time();
+            $id = $this->user_obj->insertGetId($add_user);
+            $insert_reviewer['reviewer_id'] = $id;
+            $insert_reviewer['test_from'] = "submitApplyYboardForExpert";
+            $this->user_reviewer_info_obj->insert($insert_reviewer);
+            $has = $this->user_obj->where("user_id",$id)->find();
+        }
+        $insert['user_id'] = $has['user_id'];
+        $insert['cv'] = $data['cv'];
+        $insert['ctime'] = time();
+        $this->user_cv_obj->insert($insert);
+        $inserta['user_id'] = $has['user_id'];
+        $inserta['journal_id'] = $data['journal_id'];
+        $inserta['ctime'] = time();
+        $id = $this->apply_yboard_obj->insertGetId($inserta);
+
+        $re['journal_info'] = $journal_info;
+        return jsonSuccess($re);
+
+    }
+
     
     public function up_cv_file()
     {

application/api/controller/Unsubscribe.php

@@ -0,0 +1,190 @@
+<?php
+
+namespace app\api\controller;
+
+use think\Controller;
+use think\Db;
+use think\Response;
+use app\common\UnsubscribeService;
+
+/**
+ * 推广 / 约稿邮件退订入口（公开访问，无需登录）
+ *
+ * 路由：
+ *   GET  /api/Unsubscribe/index?k=&id=&t=    展示确认页
+ *   POST /api/Unsubscribe/confirm?k=&id=&t=  执行退订
+ *
+ * k = expert  → 操作 t_expert.unsubscribed
+ * k = user    → 操作 t_user.unsubscribed
+ *
+ * 安全：
+ *   - URL 内带 sha256 签名 (kind, id, email, secret)，防伪造
+ *   - 必须确认页二次点击才执行（防爬虫预取链接误退订）
+ */
+class Unsubscribe extends Controller
+{
+    /**
+     * 退订确认页（GET）
+     */
+    public function index()
+    {
+        $kind  = UnsubscribeService::normalizeKind($this->request->param('k', UnsubscribeService::KIND_EXPERT));
+        $id    = intval($this->request->param('id', 0));
+        $token = trim((string)$this->request->param('t', ''));
+
+        $audience = $this->loadAudience($kind, $id);
+        if (!$audience) {
+            return $this->html($this->pageInvalid(), 404);
+        }
+        if (!UnsubscribeService::verifyToken($kind, $id, (string)$audience['email'], $token)) {
+            return $this->html($this->pageInvalid(), 403);
+        }
+        if (!empty($audience['unsubscribed'])) {
+            return $this->html($this->pageAlreadyDone((string)$audience['email']));
+        }
+
+        return $this->html($this->pageConfirm(
+            $kind,
+            $id,
+            $token,
+            (string)$audience['email'],
+            (string)$audience['name']
+        ));
+    }
+
+    /**
+     * 真正执行退订（POST 推荐；GET 也允许）
+     */
+    public function confirm()
+    {
+        $kind  = UnsubscribeService::normalizeKind($this->request->param('k', UnsubscribeService::KIND_EXPERT));
+        $id    = intval($this->request->param('id', 0));
+        $token = trim((string)$this->request->param('t', ''));
+
+        $audience = $this->loadAudience($kind, $id);
+        if (!$audience) {
+            return $this->html($this->pageInvalid(), 404);
+        }
+        if (!UnsubscribeService::verifyToken($kind, $id, (string)$audience['email'], $token)) {
+            return $this->html($this->pageInvalid(), 403);
+        }
+
+        if (empty($audience['unsubscribed'])) {
+            $table = $kind === UnsubscribeService::KIND_USER ? 'user' : 'expert';
+            $pk    = $kind === UnsubscribeService::KIND_USER ? 'user_id' : 'expert_id';
+            Db::name($table)->where($pk, $id)->update([
+                'unsubscribed' => 1,
+            ]);
+        }
+
+        return $this->html($this->pageSuccess((string)$audience['email']));
+    }
+
+    /**
+     * 按 kind 加载受众的最少必要信息（id, email, name, unsubscribed）
+     */
+    private function loadAudience($kind, $id)
+    {
+        if ($id <= 0) return null;
+
+        if ($kind === UnsubscribeService::KIND_USER) {
+            $row = Db::name('user')
+                ->where('user_id', $id)
+                ->field('user_id, email, realname AS name, unsubscribed')
+                ->find();
+        } else {
+            $row = Db::name('expert')
+                ->where('expert_id', $id)
+                ->field('expert_id, email, name, unsubscribed')
+                ->find();
+        }
+        return $row ?: null;
+    }
+
+    // ==================== HTML 页面 ====================
+
+    private function html($html, $status = 200)
+    {
+        $resp = Response::create($html, 'html', $status);
+        $resp->header('Content-Type', 'text/html; charset=utf-8');
+        return $resp;
+    }
+
+    private function pageShell($title, $bodyHtml)
+    {
+        $titleSafe = htmlspecialchars($title, ENT_QUOTES, 'UTF-8');
+        $css = "
+            body{margin:0;padding:0;font-family:-apple-system,Segoe UI,Roboto,Helvetica,Arial,sans-serif;background:#f5f7fa;color:#1f2937;}
+            .wrap{max-width:520px;margin:64px auto;padding:32px;background:#fff;border-radius:12px;box-shadow:0 2px 16px rgba(0,0,0,.06);}
+            h1{font-size:22px;margin:0 0 16px;color:#111827;}
+            p{font-size:15px;line-height:1.6;margin:0 0 16px;color:#374151;}
+            .email{font-weight:600;color:#111827;word-break:break-all;}
+            .btn{display:inline-block;padding:10px 24px;border-radius:6px;border:0;cursor:pointer;font-size:15px;text-decoration:none;}
+            .btn-primary{background:#dc2626;color:#fff;}
+            .btn-primary:hover{background:#b91c1c;}
+            .btn-secondary{background:#e5e7eb;color:#374151;margin-left:8px;}
+            .muted{color:#6b7280;font-size:13px;margin-top:24px;}
+            .ok{color:#16a34a;font-weight:600;}
+            .warn{color:#d97706;font-weight:600;}
+            .err{color:#dc2626;font-weight:600;}
+        ";
+        return '<!doctype html><html lang="en"><head><meta charset="utf-8">'
+            . '<meta name="viewport" content="width=device-width,initial-scale=1">'
+            . '<meta name="robots" content="noindex,nofollow">'
+            . '<title>' . $titleSafe . '</title>'
+            . '<style>' . $css . '</style></head>'
+            . '<body><div class="wrap">' . $bodyHtml . '</div></body></html>';
+    }
+
+    private function pageConfirm($kind, $id, $token, $email, $name)
+    {
+        $kindSafe  = htmlspecialchars($kind, ENT_QUOTES, 'UTF-8');
+        $idSafe    = intval($id);
+        $tokenSafe = htmlspecialchars($token, ENT_QUOTES, 'UTF-8');
+        $emailSafe = htmlspecialchars($email, ENT_QUOTES, 'UTF-8');
+        $nameSafe  = htmlspecialchars($name !== '' ? $name : $email, ENT_QUOTES, 'UTF-8');
+
+        $body = '<h1>Confirm unsubscribe</h1>'
+            . '<p>Hi ' . $nameSafe . ',</p>'
+            . '<p>You are about to unsubscribe <span class="email">' . $emailSafe
+            . '</span> from all promotion and invitation emails sent by TMR Journals. '
+            . 'After unsubscribing you will no longer receive marketing emails from us.</p>'
+            . '<form method="post" action="confirm" style="margin-top:24px;">'
+            . '<input type="hidden" name="k"  value="' . $kindSafe . '">'
+            . '<input type="hidden" name="id" value="' . $idSafe . '">'
+            . '<input type="hidden" name="t"  value="' . $tokenSafe . '">'
+            . '<button type="submit" class="btn btn-primary">Confirm unsubscribe</button>'
+            . '<a class="btn btn-secondary" href="javascript:window.close();">Cancel</a>'
+            . '</form>'
+            . '<p class="muted">If you didn\'t expect this email, you can safely close this page.</p>';
+
+        return $this->pageShell('Confirm unsubscribe - TMR Journals', $body);
+    }
+
+    private function pageSuccess($email)
+    {
+        $emailSafe = htmlspecialchars($email, ENT_QUOTES, 'UTF-8');
+        $body = '<h1>You have been unsubscribed</h1>'
+            . '<p class="ok">' . $emailSafe . ' will no longer receive promotion or invitation emails from TMR Journals.</p>'
+            . '<p>If this was a mistake, please contact us and we will be happy to resubscribe you.</p>'
+            . '<p class="muted">Thank you for your past interest in our journals.</p>';
+        return $this->pageShell('Unsubscribed - TMR Journals', $body);
+    }
+
+    private function pageAlreadyDone($email)
+    {
+        $emailSafe = htmlspecialchars($email, ENT_QUOTES, 'UTF-8');
+        $body = '<h1>Already unsubscribed</h1>'
+            . '<p class="warn">' . $emailSafe . ' is already unsubscribed from our promotion emails.</p>'
+            . '<p class="muted">No further action is needed.</p>';
+        return $this->pageShell('Already unsubscribed - TMR Journals', $body);
+    }
+
+    private function pageInvalid()
+    {
+        $body = '<h1>Invalid or expired link</h1>'
+            . '<p class="err">This unsubscribe link is invalid or has expired.</p>'
+            . '<p>Please use the most recent unsubscribe link in our emails, or contact us for help.</p>';
+        return $this->pageShell('Invalid link - TMR Journals', $body);
+    }
+}

application/api/job/CitationRelevanceQueue.php

@@ -0,0 +1,90 @@
+<?php
+namespace app\api\job;
+
+use think\queue\Job;
+use app\common\QueueJob;
+use app\common\QueueRedis;
+use app\api\controller\References;
+
+class CitationRelevanceQueue
+{
+    private $oQueueJob;
+    private $QueueRedis;
+    private $lockExpire = 900;        // 单条参考文献鉴别锁 15 分钟
+    private $completedExprie = 86400; // 结果保留 1 天
+
+    public function __construct()
+    {
+        $this->oQueueJob = new QueueJob;
+        $this->QueueRedis = QueueRedis::getInstance();
+    }
+
+    public function fire(Job $job, $data)
+    {
+        // 通用初始化（进程超时 / Redis / DB 自检）
+        $this->oQueueJob->init($job);
+
+        // 校验参数
+        $pReferId = empty($data['p_refer_id']) ? 0 : intval($data['p_refer_id']);
+        if (empty($pReferId)) {
+            $this->oQueueJob->log("无效的 p_refer_id，删除任务");
+            $job->delete();
+            return;
+        }
+
+        // 生成 Redis 键：一个参考文献只允许一个队列在跑
+        $sClassName = get_class($this);
+        $sRedisKey = "queue_job:{$sClassName}:{$pReferId}";
+        $sRedisValue = uniqid() . '_' . getmypid();
+
+        // 加锁并标记为 processing（幂等控制）
+        $isLocked = $this->QueueRedis->startJob($sRedisKey, $sRedisValue, $this->lockExpire);
+        if (!$isLocked) {
+            $jobStatus = $this->QueueRedis->getJobStatus($sRedisKey);
+            if (in_array($jobStatus, ['completed', 'failed'])) {
+                $this->oQueueJob->log("任务已完成或失败，删除任务 | 状态: {$jobStatus}");
+                $job->delete();
+            } else {
+                $attempts = $job->attempts();
+                if ($attempts >= 3) {
+                    $this->oQueueJob->log("超过最大重试次数，停止重试");
+                    $job->delete();
+                } else {
+                    $lockTtl = $this->QueueRedis->getLockTtl($sRedisKey);
+                    $delay = $lockTtl > 0 ? $lockTtl + 5 : 30;
+                    $this->oQueueJob->log("锁竞争，{$delay}秒后重试({$attempts}/3)");
+                    $job->release($delay);
+                }
+            }
+            return;
+        }
+
+        try {
+            // 实际调用 References 控制器里的同步方法，拿到完整 JSON 结果
+            $ctrl = new References();
+            $resp = $ctrl->checkCitationRelevance(['p_refer_id' => $pReferId]);
+
+            $decoded = is_string($resp) ? json_decode($resp, true) : $resp;
+            if (!is_array($decoded)) {
+                $decoded = ['status' => 2, 'msg' => 'Unexpected response from checkCitationRelevance', 'raw' => $resp];
+            }
+
+            // 结果写入 Redis，供 HTTP 轮询读取
+            $this->QueueRedis->setRedisValue($sRedisKey . ':result', json_encode($decoded, JSON_UNESCAPED_UNICODE));
+
+            // 标记完成并释放锁
+            $this->QueueRedis->finishJob($sRedisKey, 'completed', $this->completedExprie, $sRedisValue);
+            $job->delete();
+            $this->oQueueJob->log("CitationRelevanceQueue 任务执行成功 | p_refer_id: {$pReferId}");
+        } catch (\RuntimeException $e) {
+            $this->oQueueJob->handleRetryableException($e, $sRedisKey, $sRedisValue, $job);
+        } catch (\LogicException $e) {
+            $this->oQueueJob->handleNonRetryableException($e, $sRedisKey, $sRedisValue, $job);
+        } catch (\Exception $e) {
+            $this->oQueueJob->handleRetryableException($e, $sRedisKey, $sRedisValue, $job);
+        } finally {
+            $this->oQueueJob->finnal();
+        }
+    }
+}
+

application/api/job/FetchExperts.php

@@ -3,25 +3,35 @@
 namespace app\api\job;
 
 use think\queue\Job;
-use think\Log;
+use app\common\ExpertFinderService;
 
+/**
+ * 专家抓取队列任务。
+ * 注意：此任务推送到队列名 "FetchExperts"，必须单独启动 worker 才会执行：
+ *   php think queue:work --queue FetchExperts --sleep=3 --tries=3 --daemon
+ *
+ * 单条任务可能耗时较长（NCBI 接口翻页 + 写库），常驻 worker 受 wait_timeout 影响，
+ * 由 QueueJob 在进程超过 6h 或致命 DB 错误时主动 exit(1) 让 supervisor 拉起新进程。
+ */
 class FetchExperts
 {
+
     public function fire(Job $job, $data)
     {
-        try {
-            $finder = new \app\api\controller\ExpertFinder();
-            $result = $finder->doFetchForField(
-                $data['field'],
-                $data['source'] ?? 'pubmed',
-                $data['per_page'] ?? 100,
-                $data['min_year'] ?? null
-            );
-            Log::info('FetchExperts completed: ' . json_encode($result));
-        } catch (\Exception $e) {
-            Log::error('FetchExperts failed: ' . $e->getMessage());
+
+        $field = isset($data['field']) ? (string)$data['field'] : '';
+        if ($field === '') {
+            $job->delete();
+            return;
         }
 
-        $job->delete();
+            $service = new ExpertFinderService();
+            $service->doFetchForField(
+                $field,
+                isset($data['source'])   ? $data['source']        : 'pubmed',
+                isset($data['per_page']) ? intval($data['per_page']) : 100,
+                isset($data['min_year']) ? $data['min_year']      : null
+            );
+            $job->delete();
     }
 }

application/api/job/FillExpertCountry.php

@@ -0,0 +1,38 @@
+<?php
+
+namespace app\api\job;
+
+use think\queue\Job;
+use app\common\ExpertFinderService;
+
+/**
+ * 队列任务：用本地大模型从 affiliation 推断国家，写入 expert.country_id / country。
+ * 处理完当前专家后，自动找下一个推入同一队列（链式执行），直到全部处理完。
+ *
+ * 支持多队列并行：通过 $data['queue'] 和 $data['chat_url'] 区分不同的链/模型。
+ *
+ * 单条任务受本地 LLM 响应时间影响（一般 2-10s），常驻 worker 由 QueueJob
+ * 在进程超 6h 或遇致命 DB 错误时主动 exit(1) 让 supervisor 拉起新进程。
+ */
+class FillExpertCountry
+{
+
+    public function fire(Job $job, $data)
+    {
+
+        $expertId    = intval(isset($data['expert_id']) ? $data['expert_id'] : 0);
+        $affiliation = isset($data['affiliation']) ? trim((string)$data['affiliation']) : '';
+        $queue       = isset($data['queue'])    ? (string)$data['queue']    : 'FetchExperts';
+        $chatUrl     = isset($data['chat_url']) ? (string)$data['chat_url'] : '';
+
+        $service = new ExpertFinderService();
+
+            if ($expertId && $affiliation !== '') {
+                $service->fillExpertCountry($expertId, $affiliation, $chatUrl);
+            }
+            $job->delete();
+
+            // 链式：处理完当前专家立刻拉下一个进来
+            $service->enqueueNextCountryFill(1, $queue, $chatUrl);
+    }
+}

application/api/job/PlagiarismPoll.php

@@ -0,0 +1,37 @@
+<?php
+
+namespace app\api\job;
+
+use think\queue\Job;
+use app\common\PlagiarismService;
+
+/**
+ * 队列任务：轮询 Turnitin similarity 状态。
+ *
+ * 未完成会再次入队（链式延迟），完成后下载 PDF 报告并写本地永久保留。
+ *
+ * data:
+ *   - check_id  t_plagiarism_check.check_id
+ *   - attempt   当前轮询次数（首次为 1）
+ *
+ * 注意：单条 job 通常很短（1 个 HTTP 请求），但会反复入队，常驻 worker 长时间运行
+ * 由 QueueJob 在进程超 6h 或致命 DB 错误时主动 exit(1) 让 supervisor 拉起新进程。
+ */
+class PlagiarismPoll
+{
+
+    public function fire(Job $job, $data)
+    {
+
+        $checkId = isset($data['check_id']) ? intval($data['check_id']) : 0;
+        $attempt = isset($data['attempt'])  ? intval($data['attempt'])  : 1;
+
+        if ($checkId <= 0) {
+            $job->delete();
+            return;
+        }
+        $svc = new PlagiarismService();
+        $svc->runPollStatus($checkId, $attempt);
+        $job->delete();
+    }
+}

application/api/job/PlagiarismRun.php

@@ -0,0 +1,36 @@
+<?php
+
+namespace app\api\job;
+
+use think\queue\Job;
+use app\common\PlagiarismService;
+
+/**
+ * 队列任务：上传论文到 Turnitin + 触发 similarity 检测。
+ *
+ * 完成后会自动入队 PlagiarismPoll 进行后续轮询。
+ *
+ * data:
+ *   - check_id   t_plagiarism_check.check_id
+ *   - file_path  本地可读的 PDF/DOCX 绝对路径
+ *
+ * 注意：上传单个 PDF 可能耗时数十秒，常驻 worker 由 QueueJob 在进程超 6h 或致命 DB
+ * 错误时主动 exit(1) 让 supervisor 拉起新进程。
+ */
+class PlagiarismRun
+{
+
+    public function fire(Job $job, $data)
+    {
+        $checkId  = isset($data['check_id'])  ? intval($data['check_id'])    : 0;
+        $filePath = isset($data['file_path']) ? (string)$data['file_path']   : '';
+        if ($checkId <= 0 || $filePath === '') {
+            $job->delete();
+            return;
+        }
+        $svc = new PlagiarismService();
+        $svc->log("PlagiarismRun job act!!");
+        $svc->runUploadAndTrigger($checkId, $filePath);
+        $job->delete();
+    }
+}

application/api/job/PromotionPrepare.php

@@ -0,0 +1,31 @@
+<?php
+
+namespace app\api\job;
+
+use think\queue\Job;
+use app\common\PromotionService;
+
+/**
+ * 【已废弃 / 兼容保留】
+ *
+ * 旧版 task 级 prepare Job；新逻辑请使用 PromotionPrepareTask + PromotionPrepareEmail。
+ *
+ * 本类仅为兼容队列中可能遗留的旧 payload：接收到旧 job 时转发到新的调度器，
+ * 保证旧消息不会丢失。新代码不应再 push 此 Job。
+ */
+class PromotionPrepare
+{
+
+    public function fire(Job $job, $data)
+    {
+        $this->oQueueJob->init($job);
+        $taskId = isset($data['task_id']) ? intval($data['task_id']) : 0;
+        if ($taskId <= 0) {
+            $job->delete();
+            return;
+        }
+            $service = new PromotionService();
+            $service->enqueuePrepareTask($taskId);
+            $job->delete();
+    }
+}

application/api/job/PromotionPrepareEmail.php

@@ -0,0 +1,29 @@
+<?php
+
+namespace app\api\job;
+
+use think\queue\Job;
+use app\common\PromotionService;
+
+/**
+ * 队列任务：单封邮件 prepare（调用 LLM 生成个性化描述 + 渲染模板 + 写入 log）。
+ *
+ * 单条 job 可能耗时 30s+（LLM 调用）。常驻 worker 必须配合 supervisor 守护，
+ * 当进程超过 6h 或遇到致命 DB 错误时，QueueJob 会主动 exit(1)，由 supervisor 拉起新进程，
+ * 避开 MySQL wait_timeout 把连接关掉后所有任务持续 fail 的问题。
+ *
+ * 启动建议：
+ *   php think queue:work --queue PromotionPrepareEmail --sleep=3 --tries=3 --daemon
+ */
+class PromotionPrepareEmail
+{
+
+    public function fire(Job $job, $data)
+    {
+        $logId = isset($data['log_id']) ? intval($data['log_id']) : 0;
+            $service = new PromotionService();
+            $service->log("id:".$logId);
+            $result  = $service->prepareSingleEmail($logId);
+            $job->delete();
+    }
+}

application/api/job/PromotionPrepareTask.php

@@ -0,0 +1,33 @@
+<?php
+
+namespace app\api\job;
+
+use think\queue\Job;
+use app\common\PromotionService;
+use app\common\QueueJob;
+
+/**
+ * 队列任务：task 级别的 prepare 调度器。
+ *
+ * 职责：遍历 task 下待准备的 promotion_email_log，
+ *       将每封邮件拆成一条 PromotionPrepareEmail 推到 promotion_email 队列，
+ *       以便并行调用 LLM 生成个性化描述。
+ *
+ * 队列名：PromotionPrepareTask
+ */
+class PromotionPrepareTask
+{
+
+    public function fire(Job $job, $data)
+    {
+        $taskId = isset($data['task_id']) ? intval($data['task_id']) : 0;
+        if ($taskId <= 0) {
+            $job->delete();
+            return;
+        }
+            $service = new PromotionService();
+            $result  = $service->dispatchPrepareEmails($taskId);
+
+            $job->delete();
+    }
+}

application/api/job/PromotionSend.php

@@ -0,0 +1,32 @@
+<?php
+
+namespace app\api\job;
+
+use think\queue\Job;
+use app\common\PromotionService;
+
+/**
+ * 队列任务：发送 task 下"已 prepare"的邮件，按 min/max_interval 控制节奏。
+ *
+ * processNextEmail 内部会自动把下一封邮件 later() 入队（链式延迟），
+ * 因此本 job 处理完一封即可 delete，无需在这里再 dispatch 下一封。
+ *
+ * 队列名：PromotionSend
+ */
+class PromotionSend
+{
+
+    public function fire(Job $job, $data)
+    {
+
+        $taskId = isset($data['task_id']) ? intval($data['task_id']) : 0;
+        if ($taskId <= 0) {
+            $job->delete();
+            return;
+        }
+            $service = new PromotionService();
+            $result  = $service->processNextEmail($taskId);
+
+            $job->delete();
+    }
+}

application/api/job/ReferenceCheck.php

@@ -0,0 +1,114 @@
+<?php
+namespace app\api\job;
+
+use think\Db;
+use think\queue\Job;
+use app\common\QueueJob;
+use app\common\QueueRedis;
+use app\common\ReferenceCheckService;
+
+class ReferenceCheck
+{
+    private $oQueueJob;
+    private $QueueRedis;
+    private $completedExprie = 3600;
+
+    public function __construct()
+    {
+        $this->oQueueJob = new QueueJob();
+        $this->QueueRedis = QueueRedis::getInstance();
+    }
+
+    public function fire(Job $job, $data)
+    {
+        $this->oQueueJob->init($job);
+
+        $rawBody = empty($job->getRawBody()) ? '' : $job->getRawBody();
+        $jobData = empty($rawBody) ? [] : json_decode($rawBody, true);
+        $jobId = empty($jobData['id']) ? 'unknown' : $jobData['id'];
+
+        $sRedisKey = '';
+        $sRedisValue = '';
+
+        $this->oQueueJob->log("-----------队列任务开始-----------");
+        $this->oQueueJob->log("当前任务ID: {$jobId}, 尝试次数: {$job->attempts()}");
+
+        try {
+            $checkId = intval(isset($data['check_id']) ? $data['check_id'] : 0);
+            if ($checkId <= 0 && !empty($jobData['data']['check_id'])) {
+                $checkId = intval($jobData['data']['check_id']);
+            }
+            if ($checkId <= 0) {
+                $job->delete();
+                return;
+            }
+
+            $row = Db::name('article_reference_check_result')->where('id', $checkId)->find();
+            if (empty($row)) {
+                $job->delete();
+                return;
+            }
+
+            if (intval($row['status']) === 1) {
+                $job->delete();
+                return;
+            }
+
+            $sClassName = get_class($this);
+            $sRedisKey = "queue_job:{$sClassName}:{$checkId}";
+            $sRedisValue = uniqid() . '_' . getmypid();
+
+            $svc = new ReferenceCheckService();
+            $svc->clearReferenceCheckQueueLock($checkId);
+
+            if (!$this->oQueueJob->acquireLock($sRedisKey, $sRedisValue, $job)) {
+                return;
+            }
+
+            try {
+                $svc->runReferenceCheckOnce($checkId);
+
+                $amId = intval(isset($row['am_id']) ? $row['am_id'] : 0);
+                if ($amId > 0) {
+                    $svc->syncAmRefCheckStatus($amId);
+                }
+                $this->QueueRedis->finishJob($sRedisKey, 'completed', $this->completedExprie, $sRedisValue);
+                $job->delete();
+                $this->oQueueJob->log("任务执行成功 | 日志ID: {$sRedisKey}");
+            } catch (\Exception $e) {
+                $this->oQueueJob->log('ReferenceCheck error: ' . $e->getMessage());
+                if ($job->attempts() >= 3) {
+                    $this->markFailed($checkId, $e->getMessage());
+                    $job->delete();
+                    return;
+                }
+                $job->release(30);
+            }
+        } catch (\RuntimeException $e) {
+            $this->oQueueJob->handleRetryableException($e, $sRedisKey, $sRedisValue, $job);
+        } catch (\LogicException $e) {
+            $this->oQueueJob->handleNonRetryableException($e, $sRedisKey, $sRedisValue, $job);
+        } catch (\Exception $e) {
+            $this->oQueueJob->handleRetryableException($e, $sRedisKey, $sRedisValue, $job);
+        } finally {
+            $this->oQueueJob->finnal();
+        }
+    }
+
+    private function markFailed($checkId, $msg)
+    {
+        $row = Db::name('article_reference_check_result')->where('id', $checkId)->find();
+        try {
+            (new ReferenceCheckService())->updateCheckResult($checkId, [
+                'status'    => 2,
+                'error_msg' => $msg,
+            ]);
+        } catch (\Exception $e) {
+            \think\Log::error('ReferenceCheck markFailed: ' . $e->getMessage());
+        }
+        $amId = empty($row) ? 0 : intval(isset($row['am_id']) ? $row['am_id'] : 0);
+        if ($amId > 0) {
+            (new ReferenceCheckService())->syncAmRefCheckStatus($amId);
+        }
+    }
+}

application/api/job/ReferenceCheckTwo.php

@@ -0,0 +1,162 @@
+<?php
+namespace app\api\job;
+
+use think\Db;
+use think\queue\Job;
+use app\common\QueueJob;
+use app\common\QueueRedis;
+use app\common\ReferenceCheckService;
+use app\common\service\LLMService;
+
+class ReferenceCheckTwo
+{
+    private $oQueueJob;
+    private $QueueRedis;
+    private $completedExprie = 3600;
+
+    public function __construct()
+    {
+        $this->oQueueJob = new QueueJob();
+        $this->QueueRedis = QueueRedis::getInstance();
+    }
+
+    public function fire(Job $job, $data)
+    {
+        $this->oQueueJob->init($job);
+
+        $rawBody = empty($job->getRawBody()) ? '' : $job->getRawBody();
+        $jobData = empty($rawBody) ? [] : json_decode($rawBody, true);
+        $jobId = empty($jobData['id']) ? 'unknown' : $jobData['id'];
+
+        $sRedisKey = '';
+        $sRedisValue = '';
+
+        $this->oQueueJob->log("-----------队列任务开始-----------");
+        $this->oQueueJob->log("当前任务ID: {$jobId}, 尝试次数: {$job->attempts()}");
+
+        try {
+            $checkId = intval(isset($data['check_id']) ? $data['check_id'] : 0);
+            if ($checkId <= 0 && !empty($jobData['data']['check_id'])) {
+                $checkId = intval($jobData['data']['check_id']);
+            }
+            $sClassName = get_class($this);
+            $sRedisKey = "queue_job_two:{$sClassName}:{$checkId}";
+            $sRedisValue = uniqid() . '_' . getmypid();
+
+            if (!$this->oQueueJob->acquireLock($sRedisKey, $sRedisValue, $job)) {
+                return;
+            }
+
+            if ($checkId <= 0) {
+                $job->delete();
+                return;
+            }
+
+            $row = Db::name('article_reference_check_result')->where('id', $checkId)->find();
+            if (empty($row)) {
+                $job->delete();
+                return;
+            }
+
+//            if (intval($row['status']) === 1) {
+//                $job->delete();
+//                return;
+//            }
+
+            try {
+                $svc = new ReferenceCheckService();
+
+                $contentA = $svc->resolveMainContentForJob($row);
+                $referText = trim((string)(isset($row['refer_text']) ? $row['refer_text'] : ''));
+                $refer = null;
+
+                if (intval($row['p_refer_id']) > 0) {
+                    $refer = Db::name('production_article_refer')
+                        ->where('p_refer_id', intval($row['p_refer_id']))
+                        ->where('state', 0)
+                        ->find();
+                }
+
+                $payload = $svc->prepareRecheckPayload(is_array($refer) ? $refer : [], $referText);
+                $doiBlock = $payload['doi_block'];
+
+                if ($contentA === '' || $referText === '') {
+                    $this->markFailed($checkId, 'Missing article_main.content or refer_text');
+                    $job->delete();
+                    return;
+                }
+                $llm = new LLMService();
+                $llmResult = $llm->checkReference($contentA, $referText, true, $doiBlock);
+
+                $requestFailed = !empty($llmResult['request_failed']);
+                $canSupport = $svc->parseLlmCanSupport($llmResult);
+                $tag = $payload['has_abstract']
+                    ? ('[Crossref复核' . ($payload['doi_used'] !== '' ? ' ' . $payload['doi_used'] : '') . ']')
+                    : '[Crossref复核-无摘要]';
+                $reason = $tag . ' ' . (isset($llmResult['reason']) ? $llmResult['reason'] : '');
+
+                // LLM 通讯失败：写 status=2 并抛异常触发队列重试
+                if ($requestFailed) {
+                    $svc->updateCheckResult($checkId, [
+                        'confidence' => floatval($llmResult['confidence']),
+                        'reason'     => $reason,
+                        'status'     => 2,
+                        'error_msg'  => isset($llmResult['reason']) ? $llmResult['reason'] : 'LLM request failed',
+                    ]);
+                    throw new \RuntimeException(isset($llmResult['reason']) ? $llmResult['reason'] : 'LLM request failed');
+                }
+
+                $affected = $svc->updateCheckResult($checkId, [
+                    'can_support' => $canSupport ? 1 : 0,
+                    'is_match'    => $canSupport ? 1 : 0,
+                    'confidence'  => floatval($llmResult['confidence']),
+                    'reason'      => $reason,
+                    'status'      => 1,
+                    'error_msg'   => '',
+                ]);
+                $this->oQueueJob->log("Crossref复核写入 id={$checkId} affected={$affected} can_support=" . ($canSupport ? 1 : 0) . " confidence=" . floatval($llmResult['confidence']));
+
+                $amId = intval(isset($row['am_id']) ? $row['am_id'] : 0);
+                if ($amId > 0) {
+                    $svc->syncAmRefCheckStatus($amId);
+                }
+                $this->QueueRedis->finishJob($sRedisKey, 'completed', $this->completedExprie, $sRedisValue);
+                $job->delete();
+                $this->oQueueJob->log("任务执行成功 | 日志ID: {$sRedisKey}");
+            } catch (\Exception $e) {
+                $this->oQueueJob->log('ReferenceCheckTwo error: ' . $e->getMessage());
+                if ($job->attempts() >= 3) {
+                    $this->markFailed($checkId, $e->getMessage());
+                    $job->delete();
+                    return;
+                }
+                $job->release(30);
+            }
+        } catch (\RuntimeException $e) {
+            $this->oQueueJob->handleRetryableException($e, $sRedisKey, $sRedisValue, $job);
+        } catch (\LogicException $e) {
+            $this->oQueueJob->handleNonRetryableException($e, $sRedisKey, $sRedisValue, $job);
+        } catch (\Exception $e) {
+            $this->oQueueJob->handleRetryableException($e, $sRedisKey, $sRedisValue, $job);
+        } finally {
+            $this->oQueueJob->finnal();
+        }
+    }
+
+    private function markFailed($checkId, $msg)
+    {
+        $row = Db::name('article_reference_check_result')->where('id', $checkId)->find();
+        try {
+            (new ReferenceCheckService())->updateCheckResult($checkId, [
+                'status'    => 2,
+                'error_msg' => $msg,
+            ]);
+        } catch (\Exception $e) {
+            \think\Log::error('ReferenceCheckTwo markFailed: ' . $e->getMessage());
+        }
+        $amId = empty($row) ? 0 : intval(isset($row['am_id']) ? $row['am_id'] : 0);
+        if ($amId > 0) {
+            (new ReferenceCheckService())->syncAmRefCheckStatus($amId);
+        }
+    }
+}

application/api/job/myQueue.php

@@ -0,0 +1,31 @@
+<?php
+
+namespace app\api\job;
+
+use think\queue\Job;
+class myQueue
+{
+
+
+    public function fire(Job $job, $data)
+    {
+        try {
+            // 你的业务逻辑（哪怕是空的）
+            // 这里写任何代码
+
+            // 执行成功，删除任务
+            $job->delete();
+
+        } catch (\Throwable $e) {
+            // 如果你不想重试，直接删除
+            $job->delete();
+        }
+    }
+
+    // 可选：任务失败时执行
+    public function failed($data)
+    {
+        // 失败后不做处理，直接跳过
+    }
+
+}

application/common/ArticleParserService.php

@@ -1151,6 +1151,123 @@ class ArticleParserService
             ]
         ];
     }
+
+    /**
+     * 提取 Word 文档中的参考文献列表（仅返回数组，不做入库）
+     * @return array 每条为一个参考文献的纯文本字符串
+     */
+    public static function getReferencesFromWord($filePath): array
+    {
+        $othis = new self($filePath) ;
+        if (empty($othis->sections)) {
+            return [];
+        }
+
+        $lines = [];
+        foreach ($othis->sections as $section) {
+            foreach ($section->getElements() as $element) {
+                $text = $othis->getTextFromElement($element);
+                $text = trim((string)$text);
+                if ($text === '') continue;
+                $lines[] = $text;
+            }
+        }
+
+        if (empty($lines)) {
+            return [];
+        }
+
+        // 识别参考文献段落起点（允许同一行包含域代码或第一条内容）
+        $startIdx = -1;
+        $startRemainder = ''; // 标题行后可能跟着第一条参考文献内容
+        foreach ($lines as $i => $line) {
+            $t = trim($line);
+            if ($t === '') continue;
+
+            // 行首命中即可（避免 “References { ADDIN... }” / “References 1. ...” 漏判）
+            if (preg_match('/^\s*(references|reference|bibliography|参考文献|文献)\b\s*[:：]?\s*/iu', $t, $m)) {
+                $startIdx = $i;
+                $remainder = preg_replace('/^\s*(references|reference|bibliography|参考文献|文献)\b\s*[:：]?\s*/iu', '', $t);
+                $remainder = trim($remainder);
+                // 过滤 EndNote 域代码（允许其出现在标题行后）
+                if ($remainder !== '' && !preg_match('/^\{\s*ADDIN\s+EN\.REFLIST\s*\}$/i', $remainder)) {
+                    $startRemainder = $remainder;
+                }
+                break;
+            }
+        }
+
+        if ($startIdx < 0) {
+            return [];
+        }
+
+        // 收集参考文献区域内容，遇到常见结尾段落标题则停止
+        $stopKeywords = [
+            'acknowledgements', 'acknowledgments', 'funding', 'appendix', 'supplementary',
+            'conflict of interest', 'competing interests', 'author contributions',
+            '致谢', '基金', '附录', '补充材料', '利益冲突', '作者贡献',
+        ];
+
+        // startRemainder 已在起点识别时处理
+
+        $raw = [];
+        if ($startRemainder !== '') {
+            $raw[] = $startRemainder;
+        }
+
+        for ($i = $startIdx + 1; $i < count($lines); $i++) {
+            $line = trim($lines[$i]);
+            if ($line === '') continue;
+            // 跳过 EndNote / Word 域代码
+            if (preg_match('/^\{\s*ADDIN\s+EN\.REFLIST\s*\}$/i', $line)) {
+                continue;
+            }
+
+            $lineLower = strtolower($line);
+            foreach ($stopKeywords as $sk) {
+                $skLower = strtolower($sk);
+                if ($lineLower === $skLower || $lineLower === $skLower . ':' || $lineLower === $skLower . '：') {
+                    $i = count($lines); // break outer
+                    continue 2;
+                }
+            }
+
+            $raw[] = $line;
+        }
+
+        if (empty($raw)) {
+            return [];
+        }
+
+        // 合并多行：以 “数字.” / “[数字]” / “数字]” 等作为新条目起始
+        $refs = [];
+        $current = '';
+        foreach ($raw as $line) {
+            $isNew = false;
+            if (preg_match('/^\s*(\[\d+\]|\d+\s*[\.\)]|\d+\s*\])\s*/u', $line)) {
+                $isNew = true;
+            }
+
+            if ($isNew) {
+                if (trim($current) !== '') {
+                    $refs[] = trim(preg_replace('/\s+/u', ' ', $current));
+                }
+                $current = $line;
+            } else {
+                // 续行拼接
+                if ($current === '') {
+                    $current = $line;
+                } else {
+                    $current .= ' ' . $line;
+                }
+            }
+        }
+        if (trim($current) !== '') {
+            $refs[] = trim(preg_replace('/\s+/u', ' ', $current));
+        }
+
+        return $refs;
+    }
     /**
      * 核心解码方法
      * @param string $str 待解码字符串

application/common/ArticleSymbolNormalizer.php

@@ -0,0 +1,194 @@
+<?php
+
+namespace app\common;
+
+/**
+ * 期刊文章内容「符号层」校对：只调整标点、空白、全角半角等，不增删语义文字。
+ *
+ * 设计原则：
+ * - 默认规则保守，可通过 $options 逐项关闭；
+ * - 纯文本用 normalize()；含 HTML 时用 normalizeHtml()（仅处理标签之间的文本段，避免破坏属性里的 URL）。
+ * - Abstract 常用：存储时被转义为 &gt; &lt; &amp; 等，可用 normalizeAbstract() 先解码再符号校对。
+ * - 英文期刊正文/摘要通常不含中文：设 english_journal=true（或 normalizeEnglishAbstract）可关闭仅针对汉字的规则。
+ */
+class ArticleSymbolNormalizer
+{
+    /** @var string 常用汉字 BMP 段（含扩展 A 前部，足够覆盖正文） */
+    private static $han = '\x{4E00}-\x{9FFF}\x{3400}-\x{4DBF}';
+
+    /**
+     * 纯文本符号校对。
+     *
+     * @param string $text
+     * @param array  $options 可选键（均为 bool，默认 true）：
+     *   - line_endings        CRLF / CR → LF
+     *   - fullwidth_space     U+3000 全角空格 → 普通空格
+     *   - collapse_spaces     连续半角空格（不含换行）压成单个空格
+     *   - remove_zwsp         删除零宽空格等不可见格式字符（不改变可见字）
+     *   - comma_cjk           两个汉字之间的英文逗号「,」→「，」
+     *   - comma_latin         两个 ASCII 字母/数字之间的全角逗号「，」→「,」
+     *   - period_cjk          汉字后的全角句点「．」(U+FF0E) →「。」
+     *   - bracket_latin       仅由 ASCII 标识包裹时「（）」→「()」（如 (a) 类简单情形，保守：仅当括号内全为 ASCII）
+     *   - decode_html_entities 将 &gt; &lt; &amp; &quot; &#39; 及数字实体等转为真实字符（默认 false；abstract 见 normalizeAbstract）
+     *   - english_journal      英文期刊：关闭「两汉字间英文逗号→，」「汉字后 FF0E→。」等中文专用规则（默认 false；见 normalizeEnglishAbstract）
+     *
+     * @return string
+     */
+    public static function normalize($text, array $options = [])
+    {
+        $text = (string)$text;
+        if ($text === '') {
+            return '';
+        }
+
+        $o = array_merge([
+            'line_endings'         => true,
+            'fullwidth_space'      => true,
+            'collapse_spaces'      => true,
+            'remove_zwsp'          => true,
+            'comma_cjk'            => true,
+            'comma_latin'          => true,
+            'period_cjk'           => true,
+            'bracket_latin'        => false,
+            'decode_html_entities' => false,
+            'english_journal'      => false,
+        ], $options);
+
+        if (!empty($o['english_journal'])) {
+            if (!array_key_exists('comma_cjk', $options)) {
+                $o['comma_cjk'] = false;
+            }
+            if (!array_key_exists('period_cjk', $options)) {
+                $o['period_cjk'] = false;
+            }
+        }
+
+        if (!empty($o['decode_html_entities'])) {
+            $text = html_entity_decode($text, ENT_QUOTES | ENT_HTML5, 'UTF-8');
+        }
+
+        if (!empty($o['line_endings'])) {
+            $text = str_replace(["\r\n", "\r"], "\n", $text);
+        }
+        if (!empty($o['fullwidth_space'])) {
+            $text = str_replace("\u{3000}", ' ', $text);
+        }
+        if (!empty($o['remove_zwsp'])) {
+            // 零宽空格、零宽非断空格、BOM、软连字符等（不改变可见字符）
+            $text = preg_replace('/[\x{200B}-\x{200D}\x{FEFF}\x{00AD}]/u', '', $text);
+        }
+        if (!empty($o['collapse_spaces'])) {
+            $text = preg_replace('/[ \t]{2,}/u', ' ', $text);
+        }
+
+        $han = self::$han;
+
+        if (!empty($o['comma_cjk'])) {
+            // 汉字 , 汉字 → 汉字 ， 汉字
+            $text = preg_replace('/(?<=[' . $han . ']),(?=[' . $han . '])/u', '，', $text);
+        }
+        if (!empty($o['comma_latin'])) {
+            // 字母/数字 ， 字母/数字 → ,
+            $text = preg_replace('/(?<=[0-9A-Za-z])，(?=[0-9A-Za-z])/u', ',', $text);
+        }
+        if (!empty($o['period_cjk'])) {
+            // 汉字后的全角英文句点 FF0E → 中文句号 。
+            $text = preg_replace('/(?<=[' . $han . '])．/u', '。', $text);
+        }
+        if (!empty($o['bracket_latin'])) {
+            // （ 仅 ASCII + 常见标点 + 空格 ）
+            $text = preg_replace_callback(
+                '/（([0-9A-Za-z\s\.,;:\-\+/=]+)）/u',
+                static function ($m) {
+                    return '(' . $m[1] . ')';
+                },
+                $text
+            );
+        }
+
+        return $text;
+    }
+
+    /**
+     * 对 HTML 片段做符号校对：只替换「标签外」的文本，不修改标签名与属性值。
+     *
+     * 实现：按 `<...>` 切分，对偶数段（文本）调用 normalize()，奇数段（标签）原样保留。
+     * 注意：畸形 HTML、属性值中含未转义 `<` 时可能误判，复杂场景请先抽纯文本再校对。
+     *
+     * @param string $html
+     * @param array  $options 同 normalize()
+     * @return string
+     */
+    public static function normalizeHtml($html, array $options = [])
+    {
+        $html = (string)$html;
+        if ($html === '') {
+            return '';
+        }
+
+        $parts = preg_split('/(<[^>]*>)/u', $html, -1, PREG_SPLIT_DELIM_CAPTURE);
+        if ($parts === false) {
+            return self::normalize($html, $options);
+        }
+
+        $out = '';
+        foreach ($parts as $i => $chunk) {
+            if ($chunk === '') {
+                continue;
+            }
+            // 偶数索引为文本，奇数索引且以 < 开头为标签
+            if ($i % 2 === 1 && isset($chunk[0]) && $chunk[0] === '<') {
+                $out .= $chunk;
+            } else {
+                $out .= self::normalize($chunk, $options);
+            }
+        }
+
+        return $out;
+    }
+
+    /**
+     * Abstract 专用：先 HTML 实体解码（&gt; → > 等），再执行与普通正文相同的符号校对。
+     *
+     * 适用于摘要字段在库中/接口中以 htmlspecialchars 形式存储的场景。
+     * 若摘要内本身含真实 HTML 标签且需保留标签结构，请改用 normalizeHtml() 并自行传入 decode_html_entities。
+     *
+     * @param string $abstract
+     * @param array  $options 同 normalize()，默认会合并 decode_html_entities=true（可被显式 false 覆盖）
+     * @return string
+     */
+    public static function normalizeAbstract($abstract, array $options = [])
+    {
+        $opts = array_merge(['decode_html_entities' => true], $options);
+        return self::normalize($abstract, $opts);
+    }
+
+    /**
+     * 带 HTML 标签的摘要：仅在「标签外文本」中做实体解码 + 符号校对，不改动标签与属性。
+     *
+     * @param string $html
+     * @param array  $options 同 normalize()，默认 decode_html_entities=true
+     * @return string
+     */
+    public static function normalizeAbstractHtml($html, array $options = [])
+    {
+        $opts = array_merge(['decode_html_entities' => true], $options);
+        return self::normalizeHtml($html, $opts);
+    }
+
+    /**
+     * 英文期刊 Abstract：实体解码 + 符号校对，且默认关闭中文专用标点规则。
+     */
+    public static function normalizeEnglishAbstract($abstract, array $options = [])
+    {
+        return self::normalizeAbstract($abstract, array_merge(['english_journal' => true], $options));
+    }
+
+    /**
+     * 英文期刊、带 HTML 的摘要（标签外文本）：实体解码 + 符号校对，且默认关闭中文专用规则。
+     */
+    public static function normalizeEnglishAbstractHtml($html, array $options = [])
+    {
+        return self::normalizeAbstractHtml($html, array_merge(['english_journal' => true], $options));
+    }
+}

application/common/CitationRelevanceService.php

@@ -0,0 +1,331 @@
+<?php
+
+namespace app\common;
+
+/**
+ * 引用相关性检测服务（PubMed + embedding）
+ *
+ * 依赖：
+ * - PubmedService：用 DOI 抓取 title/abstract/mesh/publication_types
+ * - CrossrefService：撤稿/更正识别（补充）
+ *
+ * embedding：
+ * - 使用你们内部大模型的 embedding 接口（无需 token 付费，但速度慢）
+ * - 通过构造参数传入 embedding_url / headers / timeout
+ * - 内置文件缓存，减少重复 embedding 成本
+ */
+class CitationRelevanceService
+{
+    private $pubmed;
+    private $crossref;
+
+    private $embeddingUrl = '';
+    private $embeddingHeaders = [];
+    private $timeout = 120;
+    private $chatUrl = '';
+    private $chatModel = '';
+    private $embeddingDim = 256;
+    private $chatMaxTokens = 1200;
+
+    public function __construct(PubmedService $pubmed = null, CrossrefService $crossref = null, array $config = [])
+    {
+        $this->pubmed = $pubmed ?: new PubmedService();
+        $this->crossref = $crossref ?: new CrossrefService();
+
+        if (isset($config['embedding_url'])) $this->embeddingUrl = (string)$config['embedding_url'];
+        if (isset($config['embedding_headers']) && is_array($config['embedding_headers'])) $this->embeddingHeaders = $config['embedding_headers'];
+        if (isset($config['timeout'])) $this->timeout = max(10, intval($config['timeout']));
+        if (isset($config['chat_url'])) $this->chatUrl = (string)$config['chat_url'];
+        if (isset($config['chat_model'])) $this->chatModel = (string)$config['chat_model'];
+        if (isset($config['embedding_dim'])) $this->embeddingDim = max(32, intval($config['embedding_dim']));
+        if (isset($config['chat_max_tokens'])) $this->chatMaxTokens = max(256, intval($config['chat_max_tokens']));
+    }
+
+    /**
+     * 单条引用相关性检测
+     *
+     * @param string $contextText 引用处上下文（英文）
+     * @param array $referRow production_article_refer 行（至少含 refer_doi/doilink/title/author/joura/dateno）
+     * @param array $options
+     *  - sentence_is_background(bool) 若外部已判断为背景堆引用，可直接降级
+     *  - sim_related(float) related 阈值，默认 0.75
+     *  - sim_unsure(float) unsure 阈值，默认 0.60
+     *  - check_retraction(bool) 是否检查撤稿/更正，默认 true
+     */
+    public function checkOne(string $contextText, array $referRow, array $options = []): array
+    {
+        $contextText = trim($contextText);
+        $simRelated = isset($options['sim_related']) ? (float)$options['sim_related'] : 0.75;
+        $simUnsure = isset($options['sim_unsure']) ? (float)$options['sim_unsure'] : 0.60;
+        $checkRetraction = isset($options['check_retraction']) ? (bool)$options['check_retraction'] : true;
+        $isBackground = !empty($options['sentence_is_background']);
+
+        // 1) 问题条目（退稿/更正）：先 Crossref（有 DOI 才能判断）
+        $problemFlag = 'unknown';
+        $problemReason = '';
+        if ($checkRetraction) {
+            $qc = $this->crossref->qcCitation($contextText, $referRow, ['check_retraction' => true]);
+            $problemFlag = $qc['problem_flag'] ?? 'unknown';
+            $problemReason = $qc['problem_reason'] ?? '';
+        }
+
+        // 2) PubMed 抓取 abstract/mesh（提升语义）
+        $doi = $this->extractDoiFromRefer($referRow);
+        $pub = $doi ? $this->pubmed->fetchByDoi($doi) : null;
+
+        $pubText = '';
+        $pubTypes = [];
+        if ($pub) {
+            $pubTypes = $pub['publication_types'] ?? [];
+            $mesh = $pub['mesh_terms'] ?? [];
+            $pubText = trim(
+                ($pub['title'] ?? '') . "\n" .
+                ($pub['abstract'] ?? '') . "\n" .
+                (!empty($mesh) ? ('MeSH: ' . implode('; ', $mesh)) : '')
+            );
+        }
+
+        // 3) embedding 相似度（context vs pubmed_text），无 pubmed_text 则退化为 crossref 的证据法
+        if ($pubText !== '') {
+            $v1 = $this->embedCached($contextText);
+            $v2 = $this->embedCached($pubText);
+            $sim = ($v1 && $v2) ? $this->cosine($v1, $v2) : 0.0;
+
+            $relevanceFlag = 'unsure';
+            if ($sim >= $simRelated) {
+                $relevanceFlag = 'related';
+            } elseif ($sim >= $simUnsure) {
+                $relevanceFlag = 'unsure';
+            } else {
+                $relevanceFlag = $isBackground ? 'unsure_background' : 'suspicious_unrelated';
+            }
+
+            // PubMed 自身也能提示撤稿/更正（作为补充）
+            if ($checkRetraction && $problemFlag !== 'retracted_or_corrected' && !empty($pubTypes)) {
+                $ptLower = strtolower(implode(' | ', $pubTypes));
+                if (strpos($ptLower, 'retracted publication') !== false
+                    || strpos($ptLower, 'retraction of publication') !== false
+                    || strpos($ptLower, 'published erratum') !== false
+                ) {
+                    $problemFlag = 'retracted_or_corrected';
+                    $problemReason = 'PubMed publication type indicates retraction/correction';
+                }
+            }
+
+            return [
+                'problem_flag' => $problemFlag,
+                'problem_reason' => $problemReason,
+                'relevance_flag' => $relevanceFlag,
+                'relevance_score' => round($sim, 4),
+                'reason' => $pubText !== '' ? 'embedding(context,pubmed_text)' : 'embedding_unavailable',
+                'pubmed' => [
+                    'pmid' => $pub['pmid'] ?? '',
+                    'year' => $pub['year'] ?? '',
+                    'journal' => $pub['journal'] ?? '',
+                    'publication_types' => $pubTypes,
+                ],
+            ];
+        }
+
+        // 退化：没有 PubMed 信息时，用 CrossrefService 证据法（会偏保守）
+        $fallback = $this->crossref->qcCitation($contextText, $referRow, ['check_retraction' => false]);
+        $fallback['problem_flag'] = $problemFlag;
+        $fallback['problem_reason'] = $problemReason;
+        $fallback['reason'] = 'fallback_crossref_evidence; ' . ($fallback['reason'] ?? '');
+        return $fallback;
+    }
+
+    // ---------------- embedding ----------------
+
+    private function embedCached(string $text): ?array
+    {
+        $text = trim($text);
+        if ($text === '') return null;
+
+        $key = 'emb_' . sha1($text);
+        $cached = $this->cacheGet($key, 90 * 86400);
+        if (is_array($cached) && !empty($cached)) return $cached;
+
+        $vec = $this->embed($text);
+        if (is_array($vec) && !empty($vec)) {
+            $this->cacheSet($key, $vec);
+            return $vec;
+        }
+        return null;
+    }
+
+    /**
+     * 调用内部 embedding 接口
+     * 兼容返回格式：
+     * - OpenAI embeddings: {data:[{embedding:[...] }]}
+     * - {embedding:[...]}
+     * - 直接返回数组 [...]
+     */
+    private function embed(string $text): ?array
+    {
+        // 1) 优先使用独立 embeddings 接口
+        if ($this->embeddingUrl !== '') {
+            $payload = json_encode(['text' => $text], JSON_UNESCAPED_UNICODE);
+            $ch = curl_init();
+            curl_setopt($ch, CURLOPT_URL, $this->embeddingUrl);
+            curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
+            curl_setopt($ch, CURLOPT_TIMEOUT, $this->timeout);
+            curl_setopt($ch, CURLOPT_SSL_VERIFYPEER, false);
+            curl_setopt($ch, CURLOPT_POST, true);
+            curl_setopt($ch, CURLOPT_POSTFIELDS, $payload);
+
+            $headers = array_merge(['Content-Type: application/json'], $this->embeddingHeaders);
+            curl_setopt($ch, CURLOPT_HTTPHEADER, $headers);
+
+            $res = curl_exec($ch);
+            curl_close($ch);
+            if (!is_string($res) || trim($res) === '') return null;
+
+            $decoded = json_decode($res, true);
+            if (is_array($decoded)) {
+                if (isset($decoded['data'][0]['embedding']) && is_array($decoded['data'][0]['embedding'])) {
+                    return $this->normalizeVector($decoded['data'][0]['embedding']);
+                }
+                if (isset($decoded['embedding']) && is_array($decoded['embedding'])) {
+                    return $this->normalizeVector($decoded['embedding']);
+                }
+                $isVec = isset($decoded[0]) && (is_float($decoded[0]) || is_int($decoded[0]));
+                if ($isVec) return $this->normalizeVector($decoded);
+            }
+            return null;
+        }
+
+        // 2) 没有 embeddings 接口时，使用 chat/completions 生成固定维度向量
+        if ($this->chatUrl === '' || $this->chatModel === '') {
+            return null;
+        }
+
+        $sys = "You are an embedding generator. Output ONLY valid JSON in this exact shape: {\"embedding\":[...]}.\n"
+            . "Rules:\n"
+            . "- embedding must be an array of exactly {$this->embeddingDim} floats\n"
+            . "- each float must be between -1 and 1\n"
+            . "- do not include any other keys or any extra text\n";
+
+        $payload = json_encode([
+            'model' => $this->chatModel,
+            'temperature' => 0,
+            'max_tokens' => $this->chatMaxTokens,
+            'messages' => [
+                ['role' => 'system', 'content' => $sys],
+                ['role' => 'user', 'content' => $text],
+            ],
+        ], JSON_UNESCAPED_UNICODE);
+
+        $ch = curl_init();
+        curl_setopt($ch, CURLOPT_URL, $this->chatUrl);
+        curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
+        curl_setopt($ch, CURLOPT_TIMEOUT, $this->timeout);
+        curl_setopt($ch, CURLOPT_SSL_VERIFYPEER, false);
+        curl_setopt($ch, CURLOPT_POST, true);
+        curl_setopt($ch, CURLOPT_POSTFIELDS, $payload);
+
+        $headers = array_merge(['Content-Type: application/json'], $this->embeddingHeaders);
+        curl_setopt($ch, CURLOPT_HTTPHEADER, $headers);
+
+        $res = curl_exec($ch);
+        curl_close($ch);
+        if (!is_string($res) || trim($res) === '') return null;
+
+        $decoded = json_decode($res, true);
+        $content = '';
+        if (is_array($decoded) && isset($decoded['choices'][0]['message']['content'])) {
+            $content = (string)$decoded['choices'][0]['message']['content'];
+        }
+        $content = trim($content);
+        if ($content === '') return null;
+
+        // content 可能被包裹在 ```json ... ```
+        if (preg_match('/```(?:json)?\\s*([\\s\\S]*?)\\s*```/i', $content, $m)) {
+            $content = trim($m[1]);
+        }
+        $j = json_decode($content, true);
+        if (!is_array($j) || !isset($j['embedding']) || !is_array($j['embedding'])) {
+            return null;
+        }
+
+        $vec = $j['embedding'];
+        if (count($vec) !== $this->embeddingDim) {
+            return null;
+        }
+        return $this->normalizeVector($vec);
+    }
+
+    private function cosine(array $a, array $b): float
+    {
+        $n = min(count($a), count($b));
+        if ($n <= 0) return 0.0;
+        $dot = 0.0; $na = 0.0; $nb = 0.0;
+        for ($i = 0; $i < $n; $i++) {
+            $x = (float)$a[$i];
+            $y = (float)$b[$i];
+            $dot += $x * $y;
+            $na += $x * $x;
+            $nb += $y * $y;
+        }
+        if ($na <= 0.0 || $nb <= 0.0) return 0.0;
+        return $dot / (sqrt($na) * sqrt($nb));
+    }
+
+    private function normalizeVector(array $v): array
+    {
+        $sum = 0.0;
+        $out = [];
+        foreach ($v as $x) {
+            $fx = (float)$x;
+            $out[] = $fx;
+            $sum += $fx * $fx;
+        }
+        if ($sum <= 0.0) return $out;
+        $norm = sqrt($sum);
+        for ($i = 0; $i < count($out); $i++) {
+            $out[$i] = $out[$i] / $norm;
+        }
+        return $out;
+    }
+
+    private function extractDoiFromRefer(array $referRow): string
+    {
+        // 复用 CrossrefService 内部逻辑（通过 qcCitation 的抽取函数不可直接访问，所以这里简单实现）
+        $doi = trim((string)($referRow['refer_doi'] ?? ''));
+        if ($doi !== '') return $doi;
+
+        $doilink = trim((string)($referRow['doilink'] ?? ''));
+        if ($doilink === '') return '';
+        if (preg_match('#doi\\.org/([^?#]+)#i', $doilink, $m)) {
+            return trim((string)$m[1]);
+        }
+        return $doilink;
+    }
+
+    // ---------------- cache ----------------
+
+    private function cacheDir(): string
+    {
+        return rtrim(ROOT_PATH, '/') . '/runtime/embed_cache';
+    }
+
+    private function cacheGet(string $key, int $ttlSeconds)
+    {
+        $file = $this->cacheDir() . '/' . $key . '.json';
+        if (!is_file($file)) return null;
+        $mtime = filemtime($file);
+        if (!$mtime || (time() - $mtime) > $ttlSeconds) return null;
+        $raw = @file_get_contents($file);
+        $decoded = json_decode((string)$raw, true);
+        return $decoded;
+    }
+
+    private function cacheSet(string $key, $value): void
+    {
+        $dir = $this->cacheDir();
+        if (!is_dir($dir)) @mkdir($dir, 0777, true);
+        $file = $dir . '/' . $key . '.json';
+        @file_put_contents($file, json_encode($value, JSON_UNESCAPED_UNICODE));
+    }
+}
+

application/common/CountryResolverService.php

@@ -0,0 +1,108 @@
+<?php
+
+namespace app\common;
+
+/**
+ * 使用本地 chat/completions 接口，从 affiliation 文本推断国家。
+ *
+ * 输出约定：返回数组 ['code' => 'US', 'en_name' => 'United States']，任一可为空。
+ */
+class CountryResolverService
+{
+    private $chatUrl = '';
+    private $chatModel = '';
+    private $apiKey = '';
+    private $timeout = 60;
+
+    public function __construct(array $config = [])
+    {
+        if (isset($config['chat_url'])) $this->chatUrl = (string)$config['chat_url'];
+        if (isset($config['chat_model'])) $this->chatModel = (string)$config['chat_model'];
+        if (isset($config['api_key'])) $this->apiKey = (string)$config['api_key'];
+        if (isset($config['timeout'])) $this->timeout = max(5, intval($config['timeout']));
+    }
+
+    public function resolve(string $affiliation): array
+    {
+        $affiliation = trim($affiliation);
+        if ($affiliation === '' || $this->chatUrl === '' || $this->chatModel === '') {
+            return [];
+        }
+
+        $messages = [
+            [
+                'role' => 'system',
+                'content' => 'You extract the country from an academic affiliation string. Reply ONLY with minified JSON.',
+            ],
+            [
+                'role' => 'user',
+                'content' => "Affiliation:\n" . $affiliation . "\n\nReturn JSON with keys:\n- code: ISO 3166-1 alpha-3 (preferred)\n- en_name: English country name (optional)\nIf uncertain, return {\"code\":\"\",\"en_name\":\"\"}.",
+            ],
+        ];
+
+        $payload = [
+            'model' => $this->chatModel,
+            'temperature' => 0,
+            'messages' => $messages,
+        ];
+
+        $ch = curl_init();
+        curl_setopt($ch, CURLOPT_URL, $this->chatUrl);
+        curl_setopt($ch, CURLOPT_POST, true);
+        curl_setopt($ch, CURLOPT_POSTFIELDS, json_encode($payload, JSON_UNESCAPED_UNICODE));
+        curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
+        curl_setopt($ch, CURLOPT_CONNECTTIMEOUT, min(10, $this->timeout));
+        curl_setopt($ch, CURLOPT_TIMEOUT, $this->timeout);
+
+        $headers = [
+            'Content-Type: application/json',
+        ];
+        if ($this->apiKey !== '') {
+            $headers[] = 'Authorization: Bearer ' . $this->apiKey;
+        }
+        curl_setopt($ch, CURLOPT_HTTPHEADER, $headers);
+
+        $raw = curl_exec($ch);
+        if ($raw === false) {
+            curl_close($ch);
+            return [];
+        }
+        $httpCode = intval(curl_getinfo($ch, CURLINFO_HTTP_CODE));
+        curl_close($ch);
+        if ($httpCode < 200 || $httpCode >= 300) {
+            return [];
+        }
+
+        $data = json_decode($raw, true);
+        if (!is_array($data)) return [];
+
+        // 兼容 OpenAI chat/completions 结构：choices[0].message.content
+        $content = '';
+        if (isset($data['choices'][0]['message']['content'])) {
+            $content = (string)$data['choices'][0]['message']['content'];
+        } elseif (isset($data['content'])) {
+            $content = (string)$data['content'];
+        }
+
+        $content = trim($content);
+        if ($content === '') return [];
+
+        // 尝试提取 JSON（允许模型包裹 ```json）
+        if (preg_match('/\{.*\}/s', $content, $m)) {
+            $content = $m[0];
+        }
+
+        $obj = json_decode($content, true);
+        if (!is_array($obj)) return [];
+
+        $code = isset($obj['code']) ? strtoupper(trim((string)$obj['code'])) : '';
+        $enName = isset($obj['en_name']) ? trim((string)$obj['en_name']) : '';
+
+        // 防止模型乱回长段文本
+        if (strlen($code) > 8) $code = '';
+        if (strlen($enName) > 128) $enName = mb_substr($enName, 0, 128);
+
+        return ['code' => $code, 'en_name' => $enName];
+    }
+}
+

application/common/CrossrefService.php

@@ -0,0 +1,765 @@
+<?php
+
+namespace app\common;
+
+/**
+ * Crossref API 工具类
+ *
+ * 说明：
+ * - 仿照 application/api/controller/Crossrefdoi.php 的实现风格抽成 Service
+ * - 仅做「请求 + 解析」；不包含任何数据库读写
+ */
+class CrossrefService
+{
+    // 配置项
+    private $mailto = ''; // 邮箱（提升优先级）
+    private $timeout = 15; // 请求超时（秒）
+    private $maxRetry = 2; // 单个DOI最大重试次数
+    private $crossrefUrl = "https://api.crossref.org/works/"; // 接口地址
+
+    public function __construct($config = [])
+    {
+        if (is_array($config)) {
+            if (isset($config['mailto'])) $this->mailto = (string)$config['mailto'];
+            if (isset($config['timeout'])) $this->timeout = intval($config['timeout']);
+            if (isset($config['maxRetry'])) $this->maxRetry = intval($config['maxRetry']);
+            if (isset($config['crossrefUrl'])) $this->crossrefUrl = (string)$config['crossrefUrl'];
+        }
+    }
+
+    public function setMailto($mailto)
+    {
+        $this->mailto = (string)$mailto;
+        return $this;
+    }
+
+    /**
+     * 引用标号转换工具：正文里的 [n] 对应 production_article_refer.index = n-1（index 从 0 开始）。
+     *
+     * @param int $citationMark 正文引用编号，如 13（来自 [13]）
+     * @return int production_article_refer.index，如 12
+     */
+    public function referIndexFromCitationMark(int $citationMark): int
+    {
+        $citationMark = intval($citationMark);
+        return max(0, $citationMark - 1);
+    }
+
+    /**
+     * 反向转换工具：production_article_refer.index（从 0 开始）→ 正文引用编号 [n]。
+     *
+     * @param int $referIndex production_article_refer.index，如 12
+     * @return int 正文引用编号 n，如 13
+     */
+    public function citationMarkFromReferIndex(int $referIndex): int
+    {
+        $referIndex = intval($referIndex);
+        return max(0, $referIndex + 1);
+    }
+
+    /**
+     * 批量引用质检（不查库版）：\n
+     * - 输入文章分节内容（t_article_main 的 content 列表）\n
+     * - 输入引用条目（production_article_refer 的行列表）\n
+     * - 自动抽取每个 [n] 的英文句子上下文，并映射到 refer.index=n-1 后调用 qcCitation()\n
+     *
+     * 说明：本方法不做任何数据库查询，方便你在 controller/service 中自由组合数据来源。
+     *
+     * @param array $articleMainContents 文章内容片段数组（按 sort 顺序），元素为 string 或含 content 的数组
+     * @param array $referRows production_article_refer 行数组（至少含 index/title/author/joura/dateno/refer_doi/doilink）
+     * @param array $options 透传给 qcCitation 的 options，并支持：
+     *  - sentence_window(int) 上下文句子窗口，默认 1（即前1句+本句+后1句）
+     * @return array 结果列表，每条包含 citation_mark/refer_index/context/ref_meta/qc
+     */
+    public function qcArticleCitations(array $articleMainContents, array $referRows, array $options = []): array
+    {
+        $window = isset($options['sentence_window']) ? max(0, intval($options['sentence_window'])) : 1;
+
+        // 1) 组装全文纯文本（保留 [n]）
+        $chunks = [];
+        foreach ($articleMainContents as $row) {
+            if (is_array($row)) {
+                $text = (string)($row['content'] ?? '');
+            } else {
+                $text = (string)$row;
+            }
+            if ($text === '') continue;
+            // 去掉常见标签，保留 [n]
+            $text = preg_replace('/<\s*\/?\s*blue[^>]*>/i', '', $text);
+            $text = strip_tags($text);
+            $text = html_entity_decode($text, ENT_QUOTES | ENT_HTML5, 'UTF-8');
+            $text = preg_replace('/\s+/u', ' ', trim($text));
+            if ($text !== '') $chunks[] = $text;
+        }
+        $fullText = implode("\n", $chunks);
+
+        if ($fullText === '') return [];
+
+        // 2) 构建引用条目映射：refer_index => row
+        $referMap = [];
+        foreach ($referRows as $r) {
+            if (!is_array($r)) continue;
+            if (!isset($r['index'])) continue;
+            $idx = intval($r['index']);
+            $referMap[$idx] = $r;
+        }
+
+        // 3) 英文切句（简单稳健版）
+        $sentences = $this->splitEnglishSentences($fullText);
+        if (empty($sentences)) return [];
+
+        // 4) 遍历句子，抓取其中的 [n]
+        $results = [];
+        foreach ($sentences as $si => $sent) {
+            if (!preg_match_all('/\[(\d+)\]/', $sent, $m)) {
+                continue;
+            }
+            $marks = array_unique(array_map('intval', $m[1]));
+            foreach ($marks as $citationMark) {
+                if ($citationMark <= 0) continue;
+                $referIndex = $this->referIndexFromCitationMark($citationMark);
+                if (!isset($referMap[$referIndex])) {
+                    continue;
+                }
+
+                $start = max(0, $si - $window);
+                $end = min(count($sentences) - 1, $si + $window);
+                $ctx = implode(' ', array_slice($sentences, $start, $end - $start + 1));
+                $ctx = trim(preg_replace('/\s+/u', ' ', $ctx));
+
+                $refMeta = $referMap[$referIndex];
+                $qc = $this->qcCitation($ctx, $refMeta, $options);
+
+                $results[] = [
+                    'citation_mark' => $citationMark,     // 正文编号 n（来自 [n]）
+                    'refer_index'   => $referIndex,       // production_article_refer.index
+                    'context'       => $ctx,
+                    'ref_meta'      => [
+                        'p_refer_id' => $refMeta['p_refer_id'] ?? 0,
+                        'title'      => $refMeta['title'] ?? '',
+                        'author'     => $refMeta['author'] ?? '',
+                        'joura'      => $refMeta['joura'] ?? '',
+                        'dateno'     => $refMeta['dateno'] ?? '',
+                        'refer_doi'  => $refMeta['refer_doi'] ?? '',
+                        'doilink'    => $refMeta['doilink'] ?? '',
+                        'index'      => $refMeta['index'] ?? $referIndex,
+                    ],
+                    'qc'            => $qc,
+                ];
+            }
+        }
+
+        return $results;
+    }
+
+    /**
+     * 过滤非法DOI（仅保留10.xxxx/xxx格式）
+     * @param string $doi
+     * @return string
+     */
+    public function filterValidDoi($doi = '')
+    {
+        $doi = trim((string)$doi);
+        if ($doi === '') return '';
+        if (preg_match('/^10\.\d{4,}\/.+/', $doi)) {
+            return $doi;
+        }
+        return '';
+    }
+
+    /**
+     * 获取 Crossref message（带重试）
+     * @param string $doi
+     * @return array|null
+     */
+    public function fetchWork($doi)
+    {
+        $doi = $this->filterValidDoi($doi);
+        if ($doi === '') return null;
+        return $this->fetchSingleDoiWithRetry($doi);
+    }
+
+    /**
+     * 返回常用字段集合（标题/期刊/作者/卷期页/撤稿/URL）
+     * @param string $doi
+     * @return array|null
+     */
+    public function fetchWorkSummary($doi)
+    {
+        $msg = $this->fetchWork($doi);
+        if (!$msg) return null;
+
+        $title = $this->getTitle($msg);
+        $publisher = $this->getPublisher($msg);
+        $joura = !empty($publisher['title']) ? $publisher['title'] : ($publisher['short_title'] ?? '');
+        $authors = $this->getAuthors($msg);
+        $dateno = $this->getVolumeIssuePages($msg);
+        $retractInfo = $this->checkRetracted($msg);
+        $dolink = $this->getDolink($msg);
+        if (empty($dolink)) {
+            $dolink = 'https://doi.org/' . $this->filterValidDoi($doi);
+        }
+
+        return [
+            'doi' => $this->filterValidDoi($doi),
+            'title' => $title,
+            'joura' => $joura,
+            'publisher' => $publisher,
+            'authors' => $authors,
+            'author_str' => empty($authors) ? '' : implode(',', $authors),
+            'dateno' => $dateno,
+            'is_retracted' => !empty($retractInfo['is_retracted']) ? 1 : 0,
+            'retract_reason' => $retractInfo['reason'] ?? '',
+            'doilink' => $dolink,
+            'raw' => $msg,
+        ];
+    }
+
+    /**
+     * 单DOI查询（带重试）
+     * @param string $doi
+     * @return array|null
+     */
+    private function fetchSingleDoiWithRetry($doi)
+    {
+        $retryCount = 0;
+        while ($retryCount < $this->maxRetry) {
+            $url = $this->crossrefUrl . rawurlencode($doi);
+            if (!empty($this->mailto)) {
+                $url .= "?mailto=" . rawurlencode($this->mailto);
+            }
+
+            $ch = curl_init();
+            curl_setopt($ch, CURLOPT_URL, $url);
+            curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
+            curl_setopt($ch, CURLOPT_TIMEOUT, $this->timeout);
+            curl_setopt($ch, CURLOPT_SSL_VERIFYPEER, false);
+            curl_setopt($ch, CURLOPT_HTTPHEADER, [
+                "User-Agent: DOI-Fetcher/1.0 (mailto:{$this->mailto})"
+            ]);
+            $response = curl_exec($ch);
+            $httpCode = curl_getinfo($ch, CURLINFO_HTTP_CODE);
+            curl_close($ch);
+
+            if ($httpCode == 200) {
+                $data = json_decode($response, true);
+                return (isset($data['status']) && $data['status'] == 'ok') ? ($data['message'] ?? null) : null;
+            }
+
+            if ($httpCode == 429) {
+                sleep(5);
+                $retryCount++;
+                continue;
+            }
+
+            $retryCount++;
+            sleep(1);
+        }
+
+        return null;
+    }
+
+    /**
+     * 提取标题
+     */
+    public function getTitle($aDoiInfo = [])
+    {
+        return $aDoiInfo['title'][0] ?? '';
+    }
+
+    /**
+     * 提取期刊/出版社相关信息
+     */
+    public function getPublisher($aDoiInfo = [])
+    {
+        return [
+            'title'       => isset($aDoiInfo['container-title'][0]) ? $aDoiInfo['container-title'][0] : '',
+            'short_title' => isset($aDoiInfo['short-container-title'][0]) ? $aDoiInfo['short-container-title'][0] : '',
+            'ISSN'        => $aDoiInfo['ISSN'] ?? [],
+            'publisher'   => $aDoiInfo['publisher'] ?? '',
+        ];
+    }
+
+    /**
+     * 提取作者列表
+     */
+    public function getAuthors($aDoiInfo = [])
+    {
+        $authors = [];
+        if (!empty($aDoiInfo['author'])) {
+            foreach ($aDoiInfo['author'] as $author) {
+                $name = $author['family'] ?? '';
+                if (!empty($author['given'])) {
+                    $name = $author['given'] . ' ' . $name;
+                }
+                if (!empty($name)) {
+                    $authors[] = $name;
+                }
+            }
+        }
+        return $authors;
+    }
+
+    /**
+     * 提取发表年份
+     */
+    public function getPublishYear($aDoiInfo = [])
+    {
+        if (!empty($aDoiInfo['issued']['date-parts'][0][0])) {
+            return (string)$aDoiInfo['issued']['date-parts'][0][0];
+        }
+        return '';
+    }
+
+    /**
+     * 提取卷(期):起始页-终止页（格式：2024:10(2):100-120）
+     */
+    public function getVolumeIssuePages($aDoiInfo = [])
+    {
+        $parts = [];
+
+        $year = $this->getPublishYear($aDoiInfo);
+        if ($year) $parts[] = $year;
+
+        $volume = $aDoiInfo['volume'] ?? '';
+        $issue  = $aDoiInfo['issue'] ?? '';
+        if ($volume) {
+            $parts[] = $volume . ($issue ? "({$issue})" : '');
+        }
+
+        $pageStart = $aDoiInfo['page']['start'] ?? ($aDoiInfo['first-page'] ?? '');
+        $pageEnd   = $aDoiInfo['page']['end'] ?? ($aDoiInfo['last-page'] ?? '');
+        $pages = '';
+        if ($pageStart) {
+            $pages = $pageStart . ($pageEnd ? "-{$pageEnd}" : '');
+        } else {
+            $pages = $aDoiInfo['page'] ?? '';
+        }
+        if ($pages) $parts[] = $pages;
+
+        return implode(':', $parts);
+    }
+
+    /**
+     * 识别撤稿文章（与 Crossrefdoi.php 同逻辑）
+     */
+    public function checkRetracted($aDoiInfo = [])
+    {
+        $isRetracted = false;
+        $reason = "未撤稿";
+
+        $sType = strtolower($aDoiInfo['type'] ?? '');
+        $sSubtype = strtolower($aDoiInfo['subtype'] ?? '');
+        if ($sType && in_array($sType, ['retraction', 'correction'])) {
+            $isRetracted = true;
+            $reason = "文章类型为{$sType}（撤稿/更正声明）";
+        }
+        if ($sSubtype && in_array($sSubtype, ['retraction', 'correction'])) {
+            $isRetracted = true;
+            $reason = "文章类型为{$sSubtype}（撤稿/更正声明）";
+        }
+
+        if (isset($aDoiInfo['update-type']) && is_array($aDoiInfo['update-type']) && in_array('retraction', $aDoiInfo['update-type'])) {
+            $isRetracted = true;
+            $reason = "官方标记为撤稿（update-type: retraction）";
+        }
+
+        if (isset($aDoiInfo['relation']) && !empty($aDoiInfo['relation'])) {
+            foreach ($aDoiInfo['relation'] as $relType => $relItems) {
+                if (in_array($relType, ['is-retraction-of', 'corrects'])) {
+                    $isRetracted = true;
+                    $relatedDoi = $relItems[0]['id'] ?? '未知';
+                    $reason = "关联撤稿文章{$relatedDoi}（关系：{$relType}）";
+                    break;
+                }
+            }
+        }
+
+        if (isset($aDoiInfo['update-to']) && is_array($aDoiInfo['update-to'])) {
+            foreach ($aDoiInfo['update-to'] as $update) {
+                $updateType  = strtolower($update['type'] ?? '');
+                $updateLabel = strtolower($update['label'] ?? '');
+                if (strpos($updateType, 'retract') !== false || strpos($updateLabel, 'retract') !== false) {
+                    $isRetracted = true;
+                    $reason = "update-to 标记撤稿（{$updateType}/{$updateLabel}）";
+                    break;
+                }
+            }
+        }
+
+        $aTitles = $aDoiInfo['title'] ?? [];
+        foreach ($aTitles as $value) {
+            $sTitleLower = strtolower($value);
+            if (strpos($sTitleLower, 'retraction') !== false || strpos($sTitleLower, 'retracted') !== false
+                || strpos($sTitleLower, 'withdrawal') !== false || strpos($sTitleLower, 'withdrawn') !== false) {
+                $isRetracted = true;
+                $reason = "标题包含撤稿关键词";
+                break;
+            }
+        }
+
+        return [
+            'is_retracted' => $isRetracted,
+            'reason' => $reason,
+        ];
+    }
+
+    /**
+     * 识别 doi 链接
+     */
+    public function getDolink($aDoiInfo = [])
+    {
+        return $aDoiInfo['URL'] ?? '';
+    }
+
+    /**
+     * 解析 Crossref date-parts
+     */
+    public function parseDateParts($dateObj)
+    {
+        $parts = $dateObj['date-parts'][0] ?? [];
+        if (empty($parts)) return '';
+        $y = $parts[0] ?? '';
+        $m = $parts[1] ?? '';
+        $d = $parts[2] ?? '';
+        $out = (string)$y;
+        if ($m !== '') $out .= '-' . str_pad((string)$m, 2, '0', STR_PAD_LEFT);
+        if ($d !== '') $out .= '-' . str_pad((string)$d, 2, '0', STR_PAD_LEFT);
+        return $out;
+    }
+
+    /**
+     * 引用质检：判断(1) 被引条目是否疑似退稿/更正(基于 Crossref)；(2) 引用上下文是否与被引条目相关(基于证据命中)。
+     *
+     * 说明：
+     * - 适用于没有 abstract/keywords 的场景（仅用 title/author/journal/year + 引用上下文句子）。
+     * - 如果 refer_doi/doilink 为空，则 problem_flag 只能返回 unknown。
+     *
+     * @param string $contextText 引用处的上下文句子（英文，最好只包含引用所在句 + 少量相邻句）
+     * @param array $refMeta 被引条目元信息（建议来自 production_article_refer）
+     *  - refer_doi / doilink / title / author / joura / dateno
+     * @param array $options 可选参数
+     *  - check_retraction(bool) 是否调用 Crossref 判断退稿/更正；默认 true
+     *  - background_phrases(array) 背景堆引用触发短语；默认使用内置
+     *
+     * @return array
+     *  [
+     *    'problem_flag' => 'ok'|'retracted_or_corrected'|'unknown',
+     *    'problem_reason' => string,
+     *    'relevance_flag' => 'related'|'unsure'|'unsure_background'|'suspicious_unrelated',
+     *    'relevance_score' => float,
+     *    'reason' => string
+     *  ]
+     */
+    public function qcCitation(string $contextText, array $refMeta, array $options = []): array
+    {
+        $contextText = trim($contextText);
+        $checkRetraction = isset($options['check_retraction']) ? (bool)$options['check_retraction'] : true;
+
+        $refTitle = (string)($refMeta['title'] ?? '');
+        $refAuthor = (string)($refMeta['author'] ?? '');
+        $refJoura = (string)($refMeta['joura'] ?? '');
+        $refDateno = (string)($refMeta['dateno'] ?? '');
+        $referDoi = (string)($refMeta['refer_doi'] ?? '');
+        $doilink = (string)($refMeta['doilink'] ?? '');
+
+        $doi = $this->extractDoiFromMeta($referDoi, $doilink);
+
+        // 1) 退稿/更正判断（强规则，影响 problem_flag）
+        $problemFlag = 'unknown';
+        $problemReason = '';
+        if ($checkRetraction) {
+            if (!empty($doi)) {
+                $summary = $this->fetchWorkSummary($doi);
+                if ($summary && isset($summary['is_retracted'])) {
+                    if ((int)$summary['is_retracted'] === 1) {
+                        $problemFlag = 'retracted_or_corrected';
+                        $problemReason = !empty($summary['retract_reason']) ? $summary['retract_reason'] : 'Crossref indicates retraction/correction';
+                    } else {
+                        $problemFlag = 'ok';
+                        $problemReason = 'Crossref indicates not retracted/corrected';
+                    }
+                } else {
+                    $problemFlag = 'unknown';
+                    $problemReason = 'Crossref fetch failed or returned unexpected data';
+                }
+            } else {
+                $problemFlag = 'unknown';
+                $problemReason = 'DOI is empty';
+            }
+        } else {
+            $problemFlag = 'unknown';
+            $problemReason = 'Skip retraction check';
+        }
+
+        // 2) 相关性判断（弱规则+证据命中）
+        $backgroundPhrases = isset($options['background_phrases']) ? (array)$options['background_phrases'] : [
+            'several studies',
+            'many studies',
+            'the literature',
+            'the existing literature',
+            'has been reported',
+            'have been reported',
+            'it has been shown',
+            'previous studies',
+            'the study suggests',
+            'the literature suggests',
+            'in the literature',
+        ];
+
+        $ctxLower = strtolower($contextText);
+        $isBackground = false;
+        foreach ($backgroundPhrases as $ph) {
+            $ph = strtolower(trim((string)$ph));
+            if ($ph !== '' && $ph !== '0' && strpos($ctxLower, $ph) !== false) {
+                $isBackground = true;
+                break;
+            }
+        }
+
+        $refTokens = $this->buildEvidenceTokens([
+            'title' => $refTitle,
+            'author' => $refAuthor,
+            'journal' => $refJoura,
+            'year' => $refDateno,
+        ]);
+
+        $ctxTokens = $this->tokenize($contextText);
+
+        $titleOverlap = 0.0;
+        $authorHit = 0.0;
+        $journalOverlap = 0.0;
+        $yearHit = 0.0;
+
+        $titleTokens = $refTokens['titleTokens'] ?? [];
+        $authorTokens = $refTokens['authorTokens'] ?? [];
+        $journalTokens = $refTokens['journalTokens'] ?? [];
+        $yearToken = $refTokens['yearToken'] ?? '';
+
+        if (!empty($titleTokens)) {
+            $inter = array_intersect($titleTokens, $ctxTokens);
+            $titleOverlap = count($inter) / max(1, count($titleTokens));
+        }
+
+        if (!empty($authorTokens)) {
+            foreach ($authorTokens as $at) {
+                if ($at !== '' && in_array($at, $ctxTokens, true)) {
+                    $authorHit = 1.0;
+                    break;
+                }
+            }
+        }
+
+        if (!empty($journalTokens)) {
+            $interJ = array_intersect($journalTokens, $ctxTokens);
+            $journalOverlap = count($interJ) / max(1, count($journalTokens));
+        }
+
+        if (!empty($yearToken) && strpos($ctxLower, (string)$yearToken) !== false) {
+            $yearHit = 1.0;
+        }
+
+        // 综合得分（保持解释性：越高越相关）
+        $score = round((
+            0.60 * $titleOverlap +
+            0.20 * $authorHit +
+            0.15 * $yearHit +
+            0.05 * $journalOverlap
+        ), 4);
+
+        $relevanceFlag = 'unsure';
+        $reasonParts = [];
+
+        if ($score >= 0.35 && ($authorHit > 0.0 || $yearHit > 0.0)) {
+            $relevanceFlag = 'related';
+            $reasonParts[] = 'title_keyword_overlap_high=' . $titleOverlap;
+        } elseif ($score >= 0.25) {
+            $relevanceFlag = 'unsure';
+            $reasonParts[] = 'evidence_score_mid=' . $score;
+        } else {
+            if ($isBackground) {
+                $relevanceFlag = 'unsure_background';
+                $reasonParts[] = 'background_phrases_detected';
+            } else {
+                $relevanceFlag = 'suspicious_unrelated';
+                $reasonParts[] = 'evidence_score_low=' . $score;
+            }
+        }
+
+        $reasonParts[] = 'titleOverlap=' . $titleOverlap;
+        $reasonParts[] = 'authorHit=' . $authorHit;
+        $reasonParts[] = 'yearHit=' . $yearHit;
+        $reasonParts[] = 'journalOverlap=' . $journalOverlap;
+
+        $reason = implode('; ', $reasonParts);
+
+        return [
+            'problem_flag' => $problemFlag,
+            'problem_reason' => $problemReason,
+            'relevance_flag' => $relevanceFlag,
+            'relevance_score' => (float)$score,
+            'reason' => $reason,
+        ];
+    }
+
+    /**
+     * 从 refer_doi / doilink 中抽取 DOI 字符串。
+     * @param string $referDoi
+     * @param string $doilink
+     * @return string
+     */
+    private function extractDoiFromMeta(string $referDoi, string $doilink): string
+    {
+        $doi = trim($referDoi);
+        if (!empty($doi)) {
+            return $this->filterValidDoi($doi);
+        }
+
+        $link = trim($doilink);
+        if ($link === '') return '';
+
+        // 常见：https://doi.org/10.xxxx/xxxx 或 http://doi.org/...
+        if (preg_match('#doi\.org/([^?#]+)#i', $link, $m)) {
+            $candidate = trim((string)$m[1]);
+            return $this->filterValidDoi($candidate);
+        }
+
+        // 兜底：如果doilink本身就是doi格式
+        return $this->filterValidDoi($link);
+    }
+
+    /**
+     * 构建证据 token（用于证据命中/相似度粗判）
+     * @param array $src
+     * @return array
+     */
+    private function buildEvidenceTokens(array $src): array
+    {
+        $stop = [
+            'the','a','an','and','or','of','in','on','for','with','to','from','by','at','as','is','are',
+            'was','were','be','been','being','that','this','these','those','which','who','whom','it','its',
+            'we','our','us','they','their','them','i','you','your','he','she','his','her',
+            'study','studies','report','reported','reports','model','models','analysis','analyses','method','methods',
+            'results','result','using','used','show','shown','demonstrated','demonstrate',
+        ];
+
+        $titleTokens = $this->tokenize((string)($src['title'] ?? ''));
+        $titleTokens = array_values(array_filter(array_unique($titleTokens), function ($t) use ($stop) {
+            return !in_array($t, $stop, true) && mb_strlen($t) >= 4;
+        }));
+
+        $authorTokens = $this->extractAuthorTokens((string)($src['author'] ?? ''));
+        $authorTokens = array_values(array_unique(array_filter(array_map(function ($t) use ($stop) {
+            $t = trim($t);
+            if ($t === '') return '';
+            if (in_array($t, $stop, true)) return '';
+            return $t;
+        }, $authorTokens))));
+
+        $journalTokens = $this->tokenize((string)($src['journal'] ?? ''));
+        $journalTokens = array_values(array_filter(array_unique($journalTokens), function ($t) use ($stop) {
+            return !in_array($t, $stop, true) && mb_strlen($t) >= 4;
+        }));
+
+        $yearToken = '';
+        $yearRaw = (string)($src['year'] ?? '');
+        if (preg_match('/(19\d{2}|20\d{2})/', $yearRaw, $m)) {
+            $yearToken = (string)$m[1];
+        }
+
+        return [
+            'titleTokens' => $titleTokens,
+            'authorTokens' => $authorTokens,
+            'journalTokens' => $journalTokens,
+            'yearToken' => $yearToken,
+        ];
+    }
+
+    /**
+     * 提取作者姓/缩写 token（简化版）
+     * @param string $authorStr
+     * @return array
+     */
+    private function extractAuthorTokens(string $authorStr): array
+    {
+        $authorStr = trim($authorStr);
+        if ($authorStr === '') return [];
+
+        // 把常见分隔符拆开
+        $parts = preg_split('/[,;]| and /i', $authorStr);
+        $tokens = [];
+        foreach ($parts as $p) {
+            $p = trim((string)$p);
+            if ($p === '') continue;
+
+            // 取最后一个词当作姓（例如 "Smith J" -> "Smith"），或取首段词
+            $words = preg_split('/\s+/', $p);
+            if (empty($words)) continue;
+
+            $cand = trim((string)end($words));
+            if ($cand === '') $cand = trim((string)($words[0] ?? ''));
+
+            // 只保留字母/点号（去掉异常符号）
+            $cand = preg_replace('/[^A-Za-z\.\-]/', '', $cand);
+            $cand = strtolower($cand);
+            if ($cand !== '') {
+                $tokens[] = $cand;
+            }
+        }
+
+        // 去掉过短的 token
+        $tokens = array_values(array_filter(array_unique($tokens), function ($t) {
+            return mb_strlen($t) >= 4;
+        }));
+
+        return $tokens;
+    }
+
+    /**
+     * 文本 tokenize（英文下的轻量分词）
+     * @param string $text
+     * @return array
+     */
+    private function tokenize(string $text): array
+    {
+        $text = strtolower(trim($text));
+        if ($text === '') return [];
+
+        $parts = preg_split('/[^a-z0-9]+/i', $text);
+        $tokens = [];
+        foreach ($parts as $p) {
+            $p = trim((string)$p);
+            if ($p === '') continue;
+            // 保留较有信息量的 token
+            if (mb_strlen($p) < 3) continue;
+            $tokens[] = $p;
+        }
+
+        return array_values(array_unique($tokens));
+    }
+
+    /**
+     * 英文切句（轻量实现）：按 .?! 分割，同时保留句内的 [n]。
+     * @param string $text
+     * @return array
+     */
+    private function splitEnglishSentences(string $text): array
+    {
+        $text = trim($text);
+        if ($text === '') return [];
+
+        // 先把换行统一为空格，避免断句被打断
+        $text = preg_replace('/\s+/u', ' ', $text);
+
+        // 按句末标点断句：. ? ! 后面跟空格/结尾
+        $parts = preg_split('/(?<=[\.\?\!])\s+/', $text);
+        $sentences = [];
+        foreach ($parts as $p) {
+            $p = trim((string)$p);
+            if ($p === '') continue;
+            $sentences[] = $p;
+        }
+        return $sentences;
+    }
+}
+

application/common/ExpertFinderService.php

@@ -0,0 +1,750 @@
+<?php
+
+namespace app\common;
+
+use think\Db;
+use think\Queue;
+use GuzzleHttp\Client;
+use think\Env;
+
+class ExpertFinderService
+{
+    private $httpClient;
+    private $ncbiBaseUrl = 'https://eutils.ncbi.nlm.nih.gov/entrez/eutils/';
+    private $logFile;
+
+    public function __construct()
+    {
+        $this->httpClient = new Client([
+            'timeout'         => 180,
+            'connect_timeout' => 15,
+            'verify'          => false,
+        ]);
+        $this->logFile = ROOT_PATH . 'runtime' . DS . 'expert_finder.log';
+    }
+
+    public function doFetchForField($field, $source = 'pubmed', $perPage = 100, $minYear = null)
+    {
+        if ($minYear === null) {
+            $minYear = date('Y') - 3;
+        }
+
+        $fetchLog = $this->getFetchLog($field, $source);
+        $page     = $fetchLog['last_page'] + 1;
+
+        if ($source === 'pmc') {
+            $result = $this->searchViaPMC($field, $perPage, $minYear, $page);
+        } else {
+            $result = $this->searchViaPubMed($field, $perPage, $minYear, $page);
+        }
+
+        if(!isset($result['total'])){
+            return [
+                "has_more"=>"no"
+            ];
+        }
+        $saveResult = $this->saveExperts($result['experts'], $field, $source);
+
+        $nextPage   = $result['has_more'] ? $page : $fetchLog['last_page'];
+        $totalPages = $result['total_pages'] ?? $fetchLog['total_pages'];
+        $this->updateFetchLog($field, $source, $nextPage, $totalPages);
+
+        return [
+            'keyword'        => $field,
+            'page'           => $page,
+            'experts_found'  => $result['total'],
+            'saved_new'      => $saveResult['inserted'],
+            'saved_exist'    => $saveResult['existing'],
+            'list'           => $result['experts'],
+            'field_enriched' => $saveResult['field_enriched'],
+            'has_more'       => $result['has_more'],
+        ];
+    }
+
+    public function searchExperts($keyword, $perPage, $minYear, $page, $source)
+    {
+        if ($source === 'pmc') {
+            return $this->searchViaPMC($keyword, $perPage, $minYear, $page);
+        }
+        return $this->searchViaPubMed($keyword, $perPage, $minYear, $page);
+    }
+
+    public function saveExperts($experts, $field, $source)
+    {
+        $inserted    = 0;
+        $existing    = 0;
+        $fieldEnrich = 0;
+
+        foreach ($experts as $expert) {
+            $email = strtolower(trim($expert['email']));
+            if (empty($email)) {
+                continue;
+            }
+
+            $exists   = Db::name('expert')->where('email', $email)->find();
+            $expertId = null;
+
+            if ($exists) {
+                $existing++;
+                $expertId = intval($exists['expert_id']);
+            } else {
+                try {
+                    $expertId = Db::name('expert')->insertGetId([
+                        'name'        => mb_substr($expert['name'], 0, 255),
+                        'email'       => mb_substr($email, 0, 128),
+                        'affiliation' => mb_substr($expert['affiliation'], 0, 128),
+                        'source'      => mb_substr($source, 0, 128),
+                        'ctime'       => time(),
+                        'ltime'       => 0,
+                        'state'       => 0,
+                    ]);
+                    $inserted++;
+                } catch (\Exception $e) {
+                    $existing++;
+                    continue;
+                }
+            }
+
+            $papers = (isset($expert['papers']) && is_array($expert['papers'])) ? $expert['papers'] : [];
+            $fieldEnrich += $this->saveFieldWithPapers($expertId, $field, $source, $papers);
+        }
+
+        return ['inserted' => $inserted, 'existing' => $existing, 'field_enriched' => $fieldEnrich];
+    }
+
+    /**
+     * 保存领域与论文的关联。
+     * 有论文时：每篇论文一行（expert_id + field + source + paper_article_id 去重）。
+     * 无论文时：只存一条领域行（expert_id + field 去重）。
+     */
+    private function saveFieldWithPapers($expertId, $field, $source, $papers)
+    {
+        $field = trim($field);
+        if (empty($field)) return 0;
+
+        $added = 0;
+
+        if (empty($papers)) {
+            $exists = Db::name('expert_field')
+                ->where('expert_id', $expertId)
+                ->where('field', $field)
+                ->where('state', 0)
+                ->find();
+            if (!$exists) {
+                Db::name('expert_field')->insert([
+                    'expert_id'         => $expertId,
+                    'source'            => mb_substr((string)$source, 0, 64),
+                    'field'             => mb_substr($field, 0, 128),
+                    'paper_title'       => '',
+                    'paper_article_id'  => '',
+                    'paper_journal'     => '',
+                    'state'             => 0,
+                ]);
+                $added = 1;
+            }
+        } else {
+            foreach ($papers as $paper) {
+                $articleId = isset($paper['article_id']) ? (string)$paper['article_id'] : '';
+                if ($articleId === '' || $articleId === '0') {
+                    continue;
+                }
+
+                $check = Db::name('expert_field')
+                    ->where('expert_id', $expertId)
+                    ->where('field', $field)
+                    ->where('source', $source)
+                    ->where('paper_article_id', $articleId)
+                    ->where('state', 0)
+                    ->find();
+                if ($check) {
+                    continue;
+                }
+
+                Db::name('expert_field')->insert([
+                    'expert_id'         => $expertId,
+                    'source'            => mb_substr((string)$source, 0, 64),
+                    'paper_title'       => isset($paper['title']) ? mb_substr((string)$paper['title'], 0, 255) : '',
+                    'paper_article_id'  => mb_substr($articleId, 0, 64),
+                    'paper_journal'     => isset($paper['journal']) ? mb_substr((string)$paper['journal'], 0, 255) : '',
+                    'field'             => mb_substr($field, 0, 128),
+                    'state'             => 0,
+                ]);
+                $added++;
+            }
+        }
+
+        return $added;
+    }
+
+    public function getFetchLog($field, $source)
+    {
+        $log = Db::name('expert_fetch')
+            ->where('field', $field)
+            ->where('source', $source)
+            ->find();
+
+        if (!$log) {
+            return ['last_page' => 0, 'total_pages' => 0, 'last_time' => 0];
+        }
+
+        return $log;
+    }
+
+    public function updateFetchLog($field, $source, $lastPage, $totalPages)
+    {
+        $exists = Db::name('expert_fetch')
+            ->where('field', $field)
+            ->where('source', $source)
+            ->find();
+
+        if ($exists) {
+            Db::name('expert_fetch')
+                ->where('expert_fetch_id', $exists['expert_fetch_id'])
+                ->update([
+                    'last_page'   => $lastPage,
+                    'total_pages' => $totalPages,
+                    'last_time'   => time(),
+                ]);
+        } else {
+            Db::name('expert_fetch')->insert([
+                'field'       => mb_substr($field, 0, 128),
+                'source'      => mb_substr($source, 0, 128),
+                'last_page'   => $lastPage,
+                'total_pages' => $totalPages,
+                'last_time'   => time(),
+            ]);
+        }
+    }
+
+    // ==================== PubMed Search ====================
+
+    private function searchViaPubMed($keyword, $perPage, $minYear, $page = 1)
+    {
+        set_time_limit(600);
+
+        $searchResult  = $this->esearch('pubmed', $keyword, $perPage, $minYear, $page);
+        $ids           = $searchResult['ids'];
+        $totalArticles = $searchResult['total'];
+
+        if (empty($ids)) {
+            return $this->buildPagedResult([], 0, 0, $totalArticles, $page, $perPage, 'pubmed');
+        }
+
+        $allAuthors = [];
+        $batches = array_chunk($ids, 50);
+        foreach ($batches as $batch) {
+            $xml = $this->efetchWithRetry('pubmed', $batch);
+            if ($xml) {
+                $authors = $this->parsePubMedXml($xml);
+                $allAuthors = array_merge($allAuthors, $authors);
+            }
+            usleep(400000);
+        }
+
+        $experts = $this->aggregateExperts($allAuthors);
+
+        return $this->buildPagedResult($experts, count($experts), count($ids), $totalArticles, $page, $perPage, 'pubmed');
+    }
+
+    // ==================== PMC Search ====================
+
+    private function searchViaPMC($keyword, $perPage, $minYear, $page = 1)
+    {
+        set_time_limit(600);
+
+        $searchResult  = $this->esearch('pmc', $keyword, $perPage, $minYear, $page);
+        $ids           = $searchResult['ids'];
+        $totalArticles = $searchResult['total'];
+
+        if (empty($ids)) {
+            return $this->buildPagedResult([], 0, 0, $totalArticles, $page, $perPage, 'pmc');
+        }
+
+        $allAuthors = [];
+        $batches = array_chunk($ids, 5);
+        foreach ($batches as $batch) {
+            $xml = $this->efetchWithRetry('pmc', $batch);
+            if ($xml) {
+                $authors = $this->parsePMCXml($xml);
+                $allAuthors = array_merge($allAuthors, $authors);
+            }
+            usleep(500000);
+        }
+
+        $experts = $this->aggregateExperts($allAuthors);
+
+        return $this->buildPagedResult($experts, count($experts), count($ids), $totalArticles, $page, $perPage, 'pmc');
+    }
+
+    // ==================== NCBI API ====================
+
+    private function esearch($db, $keyword, $perPage, $minYear, $page = 1)
+    {
+        $term     = $keyword . ' AND ' . $minYear . ':' . date('Y') . '[pdat]';
+        $retstart = ($page - 1) * $perPage;
+
+        $response = $this->httpClient->get($this->ncbiBaseUrl . 'esearch.fcgi', [
+            'query' => [
+                'db'       => $db,
+                'term'     => $term,
+                'retstart' => $retstart,
+                'retmax'   => $perPage,
+                'retmode'  => 'json',
+                'sort'     => 'relevance',
+            ],
+        ]);
+
+        $data  = json_decode($response->getBody()->getContents(), true);
+        $ids   = $data['esearchresult']['idlist'] ?? [];
+        $total = intval($data['esearchresult']['count'] ?? 0);
+
+        return ['ids' => $ids, 'total' => $total];
+    }
+
+    private function efetch($db, $ids)
+    {
+        $response = $this->httpClient->post($this->ncbiBaseUrl . 'efetch.fcgi', [
+            'form_params' => [
+                'db'      => $db,
+                'id'      => implode(',', $ids),
+                'retmode' => 'xml',
+            ],
+        ]);
+
+        return $response->getBody()->getContents();
+    }
+
+    private function efetchWithRetry($db, $ids, $maxRetries = 3)
+    {
+        for ($attempt = 1; $attempt <= $maxRetries; $attempt++) {
+            try {
+                return $this->efetch($db, $ids);
+            } catch (\Exception $e) {
+                if ($attempt === $maxRetries) {
+                    if (count($ids) > 1) {
+                        $half = ceil(count($ids) / 2);
+                        $firstHalf  = array_slice($ids, 0, $half);
+                        $secondHalf = array_slice($ids, $half);
+                        $xml1 = $this->efetchWithRetry($db, $firstHalf, 2);
+                        $xml2 = $this->efetchWithRetry($db, $secondHalf, 2);
+                        return $this->mergeXml($xml1, $xml2);
+                    }
+                    return null;
+                }
+                sleep($attempt * 2);
+            }
+        }
+        return null;
+    }
+
+    private function mergeXml($xml1, $xml2)
+    {
+        if (empty($xml1)) return $xml2;
+        if (empty($xml2)) return $xml1;
+        return $xml1 . "\n" . $xml2;
+    }
+
+    // ==================== PubMed XML Parsing ====================
+
+    private function parsePubMedXml($xmlString)
+    {
+        $results = [];
+        libxml_use_internal_errors(true);
+        $xml = simplexml_load_string($xmlString);
+        if ($xml === false) {
+            return $results;
+        }
+
+        foreach ($xml->PubmedArticle as $article) {
+            $citation   = $article->MedlineCitation;
+            $articleData = $citation->Article;
+            $title = $this->xmlNodeToString($articleData->ArticleTitle);
+            $pmid  = (string) $citation->PMID;
+
+            $journal = '';
+            if (isset($articleData->Journal->Title)) {
+                $journal = (string) $articleData->Journal->Title;
+            }
+            if (!isset($articleData->AuthorList->Author)) {
+                continue;
+            }
+
+            foreach ($articleData->AuthorList->Author as $author) {
+                $lastName  = (string) ($author->LastName ?? '');
+                $foreName  = (string) ($author->ForeName ?? '');
+                $fullName  = trim($foreName . ' ' . $lastName);
+                if (empty($fullName)) continue;
+
+                $email       = '';
+                $affiliation = '';
+                if (isset($author->AffiliationInfo)) {
+                    foreach ($author->AffiliationInfo as $affInfo) {
+                        $affText = (string) $affInfo->Affiliation;
+                        if (empty($affiliation)) $affiliation = $affText;
+                        if (empty($email)) $email = $this->extractEmailFromText($affText);
+                    }
+                }
+                if (empty($email)) continue;
+
+                $results[] = [
+                    'name'          => $fullName,
+                    'email'         => strtolower($email),
+                    'affiliation'   => $this->cleanAffiliation($affiliation),
+                    'article_title' => $title,
+                    'article_id'    => $pmid,
+                    'journal'       => $journal,
+                ];
+            }
+        }
+
+        return $results;
+    }
+
+    // ==================== PMC XML Parsing ====================
+
+    private function parsePMCXml($xmlString)
+    {
+        $results = [];
+        libxml_use_internal_errors(true);
+        $xml = simplexml_load_string($xmlString);
+        if ($xml === false) {
+            return $results;
+        }
+
+        $articles = $xml->article ?? $xml->children();
+
+        foreach ($articles as $article) {
+            if ($article->getName() !== 'article') continue;
+
+            $front = $article->front;
+            if (!$front) continue;
+            $articleMeta = $front->{'article-meta'};
+            if (!$articleMeta) continue;
+
+            $title = $this->xmlNodeToString($articleMeta->{'title-group'}->{'article-title'} ?? null);
+            $pmcId = '';
+            if (isset($articleMeta->{'article-id'})) {
+                foreach ($articleMeta->{'article-id'} as $idNode) {
+                    if ((string) $idNode['pub-id-type'] === 'pmc') {
+                        $pmcId = (string) $idNode;
+                    }
+                }
+            }
+
+            $journal = '';
+            if (isset($front->{'journal-meta'}->{'journal-title'})) {
+                $journal = (string) $front->{'journal-meta'}->{'journal-title'};
+            } elseif (isset($front->{'journal-meta'}->{'journal-title-group'}->{'journal-title'})) {
+                $journal = (string) $front->{'journal-meta'}->{'journal-title-group'}->{'journal-title'};
+            }
+
+            $correspEmails = [];
+            if (isset($articleMeta->{'author-notes'})) {
+                $this->extractEmailsFromNode($articleMeta->{'author-notes'}, $correspEmails);
+            }
+
+            $affiliationMap = [];
+            if (isset($articleMeta->{'contrib-group'})) {
+                foreach ($articleMeta->{'contrib-group'}->children() as $child) {
+                    if ($child->getName() === 'aff') {
+                        $affId   = (string) ($child['id'] ?? '');
+                        $affText = $this->xmlNodeToString($child);
+                        if ($affId) $affiliationMap[$affId] = $affText;
+                    }
+                }
+            }
+            if (isset($front->{'article-meta'}->{'aff'})) {
+                foreach ($front->{'article-meta'}->{'aff'} as $aff) {
+                    $affId   = (string) ($aff['id'] ?? '');
+                    $affText = $this->xmlNodeToString($aff);
+                    if ($affId) $affiliationMap[$affId] = $affText;
+                }
+            }
+
+            if (!isset($articleMeta->{'contrib-group'})) continue;
+
+            foreach ($articleMeta->{'contrib-group'}->contrib as $contrib) {
+                if ((string) ($contrib['contrib-type'] ?? '') !== 'author') continue;
+                $nameNode = $contrib->name;
+                if (!$nameNode) continue;
+
+                $surname    = (string) ($nameNode->surname ?? '');
+                $givenNames = (string) ($nameNode->{'given-names'} ?? '');
+                $fullName   = trim($givenNames . ' ' . $surname);
+                if (empty($fullName)) continue;
+
+                $email = '';
+                if (isset($contrib->email)) {
+                    $email = strtolower(trim((string) $contrib->email));
+                }
+
+                $affiliation = '';
+                if (isset($contrib->xref)) {
+                    foreach ($contrib->xref as $xref) {
+                        if ((string) $xref['ref-type'] === 'aff') {
+                            $rid = (string) $xref['rid'];
+                            if (isset($affiliationMap[$rid])) {
+                                $affiliation = $affiliationMap[$rid];
+                                break;
+                            }
+                        }
+                    }
+                }
+                if (empty($affiliation) && isset($contrib->aff)) {
+                    $affiliation = $this->xmlNodeToString($contrib->aff);
+                }
+
+                $isCorresponding = false;
+                if (isset($contrib->xref)) {
+                    foreach ($contrib->xref as $xref) {
+                        if ((string) $xref['ref-type'] === 'corresp') $isCorresponding = true;
+                    }
+                }
+                if ((string) ($contrib['corresp'] ?? '') === 'yes') $isCorresponding = true;
+
+                if (empty($email) && $isCorresponding && !empty($correspEmails)) {
+                    $email = $correspEmails[0];
+                }
+                if (empty($email)) {
+                    $extracted = $this->extractEmailFromText($affiliation);
+                    if ($extracted) $email = $extracted;
+                }
+                if (empty($email)) continue;
+
+                $results[] = [
+                    'name'          => $fullName,
+                    'email'         => strtolower($email),
+                    'affiliation'   => $this->cleanAffiliation($affiliation),
+                    'article_title' => $title,
+                    'article_id'    => $pmcId,
+                    'journal'       => $journal,
+                ];
+            }
+        }
+
+        return $results;
+    }
+
+    // ==================== Aggregation / Pagination ====================
+
+    private function aggregateExperts($authorRecords)
+    {
+        $map = [];
+        foreach ($authorRecords as $record) {
+            $key = strtolower(trim($record['email']));
+            if (empty($key)) continue;
+
+            if (!isset($map[$key])) {
+                $map[$key] = [
+                    'name'        => $record['name'],
+                    'email'       => $record['email'],
+                    'affiliation' => $record['affiliation'],
+                    'paper_count' => 0,
+                    'papers'      => [],
+                ];
+            }
+            $map[$key]['paper_count']++;
+            if (count($map[$key]['papers']) < 10) {
+                $map[$key]['papers'][] = [
+                    'title'      => $record['article_title'],
+                    'article_id' => $record['article_id'],
+                    'journal'    => $record['journal'],
+                ];
+            }
+            if (empty($map[$key]['affiliation']) && !empty($record['affiliation'])) {
+                $map[$key]['affiliation'] = $record['affiliation'];
+            }
+        }
+
+        $experts = array_values($map);
+        usort($experts, function ($a, $b) {
+            return $b['paper_count'] - $a['paper_count'];
+        });
+        return $experts;
+    }
+
+    private function buildPagedResult($experts, $expertCount, $articlesScanned, $totalArticles, $page, $perPage, $source)
+    {
+        $totalPages = $totalArticles > 0 ? ceil($totalArticles / $perPage) : 0;
+        return [
+            'experts'          => $experts,
+            'total'            => $expertCount,
+            'articles_scanned' => $articlesScanned,
+            'total_articles'   => $totalArticles,
+            'page'             => $page,
+            'per_page'         => $perPage,
+            'total_pages'      => $totalPages,
+            'has_more'         => $page < $totalPages,
+            'source'           => $source,
+        ];
+    }
+
+    // ==================== Country Resolution ====================
+
+    /**
+     * 启动国家解析链：找到下一个缺国家的专家推入队列。
+     * 队列 Job 处理完一个后会再调此方法，自动找下一个，直到全部处理完。
+     * 控制器只需调一次即可。
+     *
+     * @param int $delay 延迟秒数（防止打满模型，默认1秒）
+     * @return bool 是否成功推入了一条
+     */
+    /**
+     * 启动国家解析链：找到下一个缺国家的专家推入指定队列。
+     *
+     * @param int    $delay     延迟秒数
+     * @param string $queue     队列名（不同队列跑不同 worker，互不阻塞）
+     * @param string $chatUrl   该链使用的模型地址（为空则用默认）
+     * @return bool
+     */
+    public function enqueueNextCountryFill($delay = 1, $queue = 'FetchExperts', $chatUrl = '')
+    {
+        $row = Db::name('expert')
+            ->where('affiliation', '<>', '')
+            ->where('country_id', 0)
+            ->where('state', '<>', 5)
+            ->field('expert_id, affiliation')
+            ->order('expert_id asc')
+            ->find();
+
+        if (!$row) {
+            $this->log('[CountryFill] no more pending experts, queue=' . $queue);
+            return false;
+        }
+
+        $data = [
+            'expert_id'   => intval($row['expert_id']),
+            'affiliation' => trim((string)$row['affiliation']),
+            'queue'       => $queue,
+            'chat_url'    => $chatUrl,
+        ];
+
+        $jobClass = 'app\api\job\FillExpertCountry@fire';
+        if ($delay > 0) {
+            Queue::later($delay, $jobClass, $data, $queue);
+        } else {
+            Queue::push($jobClass, $data, $queue);
+        }
+
+        return true;
+    }
+
+    /**
+     * 对单个专家执行国家解析（同步），由队列 Job FillExpertCountry 调用，也可直接调用测试。
+     */
+    public function fillExpertCountry($expertId, $affiliation, $chatUrl = '')
+    {
+        $affiliation = trim((string)$affiliation);
+        if ($affiliation === '') {
+            Db::name('expert')->where('expert_id', intval($expertId))->update(['country_id' => -1]);
+            return;
+        }
+
+        $defaultUrl = trim((string)Env::get('expert_country_chat_url', Env::get('citation_chat_url', 'http://chat.taimed.cn/v1/chat/completions')));
+        $url = ($chatUrl !== '') ? $chatUrl : $defaultUrl;
+
+        $resolver = new CountryResolverService([
+            'chat_url'   => $url,
+            'chat_model' => trim((string)Env::get('expert_country_chat_model', Env::get('citation_chat_model', 'gpt-4.1'))),
+            'api_key'    => trim((string)Env::get('expert_country_chat_api_key', Env::get('citation_chat_api_key', ''))),
+            'timeout'    => max(20, intval(Env::get('expert_country_chat_timeout', 60))),
+        ]);
+
+        $result = $resolver->resolve($affiliation);
+
+        $countryId = 0;
+        $enName = '';
+
+        if (!empty($result['code'])) {
+            $row = Db::name('country')->where('code', strtoupper(trim((string)$result['code'])))->find();
+            if ($row) {
+                $countryId = intval($row['country_id']);
+                $enName = (string)$row['en_name'];
+            }
+        }
+
+        if ($countryId === 0 && !empty($result['en_name'])) {
+            $row = Db::name('country')
+                ->whereRaw("LOWER(en_name) = ?", [strtolower(trim((string)$result['en_name']))])
+                ->find();
+            if ($row) {
+                $countryId = intval($row['country_id']);
+                $enName = (string)$row['en_name'];
+            }
+        }
+
+        if ($countryId > 0 && $enName !== '') {
+            Db::name('expert')->where('expert_id', intval($expertId))->update([
+                'country_id' => $countryId,
+                'country'    => $enName,
+            ]);
+        } else {
+            // country_id = -1 表示「已尝试但未识别」，避免链式执行时反复卡在同一条
+            Db::name('expert')->where('expert_id', intval($expertId))->update([
+                'country_id' => -1,
+            ]);
+            $this->log('[CountryFill] expert_id=' . $expertId . ' unresolved, code=' . ($result['code'] ?? '') . ' en_name=' . ($result['en_name'] ?? ''));
+        }
+    }
+
+    // ==================== Text Helpers ====================
+
+    private function extractEmailFromText($text)
+    {
+        if (empty($text)) return '';
+        if (preg_match('/[Ee]lectronic address:\s*([^\s;,]+@[^\s;,]+)/', $text, $m)) {
+            return strtolower(trim($m[1], '.'));
+        }
+        if (preg_match('/[Ee]-?mail:\s*([^\s;,]+@[^\s;,]+)/', $text, $m)) {
+            return strtolower(trim($m[1], '.'));
+        }
+        if (preg_match('/\b([a-zA-Z0-9._%+\-]+@[a-zA-Z0-9.\-]+\.[a-zA-Z]{2,})\b/', $text, $m)) {
+            return strtolower(trim($m[1], '.'));
+        }
+        return '';
+    }
+
+    private function extractEmailsFromNode($node, &$emails)
+    {
+        if ($node === null) return;
+        foreach ($node->children() as $child) {
+            if ($child->getName() === 'email') {
+                $email = strtolower(trim((string) $child));
+                if (!empty($email) && !in_array($email, $emails)) $emails[] = $email;
+            }
+            $this->extractEmailsFromNode($child, $emails);
+        }
+        $text = (string) $node;
+        if (preg_match_all('/\b([a-zA-Z0-9._%+\-]+@[a-zA-Z0-9.\-]+\.[a-zA-Z]{2,})\b/', $text, $matches)) {
+            foreach ($matches[1] as $email) {
+                $email = strtolower(trim($email, '.'));
+                if (!in_array($email, $emails)) $emails[] = $email;
+            }
+        }
+    }
+
+    private function cleanAffiliation($text)
+    {
+        $text = preg_replace('/\s*[Ee]lectronic address:\s*[^\s;,]+@[^\s;,]+/', '', $text);
+        $text = preg_replace('/\s*[Ee]-?mail:\s*[^\s;,]+@[^\s;,]+/', '', $text);
+        $text = preg_replace('/\s*\b[a-zA-Z0-9._%+\-]+@[a-zA-Z0-9.\-]+\.[a-zA-Z]{2,}\b/', '', $text);
+        return trim($text, " \t\n\r\0\x0B.,;");
+    }
+
+    private function xmlNodeToString($node)
+    {
+        if ($node === null) return '';
+        $xml  = $node->asXML();
+        $text = strip_tags($xml);
+        $text = html_entity_decode($text, ENT_QUOTES | ENT_XML1, 'UTF-8');
+        return trim(preg_replace('/\s+/', ' ', $text));
+    }
+
+    // ==================== Logging ====================
+
+    public function log($msg)
+    {
+        $line = date('Y-m-d H:i:s') . ' ' . $msg . PHP_EOL;
+        @file_put_contents($this->logFile, $line, FILE_APPEND);
+    }
+}

application/common/PlagiarismService.php

@@ -0,0 +1,440 @@
+<?php
+
+namespace app\common;
+
+use think\Db;
+use think\Env;
+use think\Queue;
+use think\Exception;
+
+/**
+ * 查重业务层：把 TurnitinService 的低层调用包装成"按 article 查重"的高层流程，
+ * 并维护 t_plagiarism_check 状态机。
+ *
+ * 状态流：
+ *   submit()                 → state=1（上传中），入队 PlagiarismRun
+ *   PlagiarismRun.fire       → 上传 + 触发 similarity → state=2（比对中），入队 PlagiarismPoll
+ *   PlagiarismPoll.fire      → 轮询 status，完成后下载 PDF → state=3（完成）
+ *   任意环节抛异常           → state=4（失败），写 error_msg
+ */
+class PlagiarismService
+{
+    /**
+     * 报告 PDF 本地保存目录（相对于项目根，永久保留）
+     */
+    const REPORT_DIR = 'public/plagiarism';
+
+    /**
+     * 轮询间隔（秒）。Turnitin 一般 1-5 分钟出结果，30 秒一次比较合适
+     */
+    const POLL_INTERVAL = 30;
+
+    /**
+     * 最长轮询次数（30s × 60 = 30 分钟）
+     */
+    const MAX_POLL_ATTEMPTS = 60;
+
+    private $logFile;
+
+    public function __construct()
+    {
+        $this->logFile = ROOT_PATH . 'runtime' . DS . 'plagiarism_task.log';
+    }
+
+    // ---------- 顶层入口 ----------
+
+    /**
+     * 提交查重（入队，立即返回 check_id）
+     *
+     * @param int    $articleId   投稿 ID
+     * @param string $filePath    本地可读的 PDF/DOCX 绝对路径
+     * @param int    $triggeredBy 触发人 user_id（手工触发时编辑后台的 user_id）
+     * @param string $source      'manual' / 'auto_xxx'
+     * @return int   check_id
+     */
+    public function submit($articleId, $filePath, $triggeredBy = 0, $source = 'manual')
+    {
+        if (!is_file($filePath) || !is_readable($filePath)) {
+            throw new Exception("File not readable: {$filePath}");
+        }
+
+        $journalId = (int) Db::name('article')
+            ->where('article_id', $articleId)
+            ->value('journal_id');
+
+
+        $now = time();
+        $checkId = Db::name('plagiarism_check')->insertGetId([
+            'article_id'       => $articleId,
+            'journal_id'       => $journalId,
+            'triggered_by'     => $triggeredBy,
+            'trigger_source'   => $source,
+            'state'            => 1,           // 上传中
+            'source_file_name' => basename($filePath),
+            'source_file_size' => filesize($filePath) ?: 0,
+            'ctime'            => $now,
+            'utime'            => $now,
+        ]);
+        $this->log("submit service act");
+        // 入队执行：上传 + 触发 similarity
+        Queue::push(
+            'app\\api\\job\\PlagiarismRun',
+            ['check_id' => $checkId, 'file_path' => $filePath],
+            'PlagiarismRun'
+        );
+
+        return (int)$checkId;
+    }
+
+    /**
+     * Job 调用：上传文件到 Turnitin 并触发 similarity，然后入队 PlagiarismPoll
+     */
+    public function runUploadAndTrigger($checkId, $filePath)
+    {
+        $check = $this->mustGetCheck($checkId);
+        $this->log("runUploadAndTrigger is act0");
+        try {
+            $tii = new TurnitinService();
+
+            // 1. 创建 submission
+            $articleTitle = (string) Db::name('article')
+                ->where('article_id', $check['article_id'])
+                ->value('title');
+            if ($articleTitle === '') {
+                $articleTitle = 'Article #' . $check['article_id'];
+            }
+            $this->log("runUploadAndTrigger is act1");
+            $createResp = $tii->createSubmission([
+                'title'     => mb_substr($articleTitle, 0, 250),
+                'owner'     => 'editor_' . $check['triggered_by'],
+                'submitter' => 'editor_' . $check['triggered_by'],
+                'metadata'  => [
+                    'article_id' => (string)$check['article_id'],
+                    'check_id'   => (string)$check['check_id'],
+                ],
+            ]);
+            $submissionId = isset($createResp['id']) ? $createResp['id'] : '';
+            if ($submissionId === '') {
+                throw new Exception('Turnitin createSubmission returned empty id: ' . json_encode($createResp));
+            }
+
+            $this->updateCheck($checkId, [
+                'tii_submission_id' => $submissionId,
+                'raw_response'      => json_encode($createResp, JSON_UNESCAPED_UNICODE),
+            ]);
+            $this->log("runUploadAndTrigger is act2");
+            // 2. 上传文件
+            $tii->uploadFile($submissionId, $filePath, basename($filePath));
+
+            // 3. 触发 similarity
+            $simResp = $tii->triggerSimilarity($submissionId);
+
+            $this->updateCheck($checkId, [
+                'state'             => 2,                // 比对中
+                'tii_report_status' => 'PROCESSING',
+                'raw_response'      => json_encode($simResp, JSON_UNESCAPED_UNICODE),
+            ]);
+
+            $this->log("runUploadAndTrigger is act3");
+
+            // 4. 排队首次轮询（晚一点开始，让 Turnitin 先处理）
+            Queue::later(
+                self::POLL_INTERVAL,
+                'app\\api\\job\\PlagiarismPoll',
+                ['check_id' => $checkId, 'attempt' => 1],
+                'PlagiarismPoll'
+            );
+        } catch (\Throwable $e) {
+            $this->markFailed($checkId, '[upload] ' . $e->getMessage());
+            throw $e;
+        }
+    }
+
+    /**
+     * Job 调用：轮询 similarity 状态，完成后下载 PDF。未完成则重新入队。
+     */
+    public function runPollStatus($checkId, $attempt = 1)
+    {
+        $check = $this->mustGetCheck($checkId);
+        if (empty($check['tii_submission_id'])) {
+            $this->markFailed($checkId, '[poll] tii_submission_id empty');
+            return;
+        }
+
+        try {
+            $tii = new TurnitinService();
+            $statusResp = $tii->getSimilarityStatus($check['tii_submission_id']);
+            $status = isset($statusResp['status']) ? strtoupper($statusResp['status']) : '';
+
+            $this->updateCheck($checkId, [
+                'tii_report_status' => $status,
+                'attempts'          => $attempt,
+                'raw_response'      => json_encode($statusResp, JSON_UNESCAPED_UNICODE),
+            ]);
+
+            if ($status === 'COMPLETE') {
+                $score = isset($statusResp['overall_match_percentage'])
+                    ? floatval($statusResp['overall_match_percentage']) : 0;
+
+                // 下载 PDF + 取在线查看 URL
+                $localPdf = $this->downloadAndStorePdf($tii, $check['tii_submission_id'], $checkId);
+                $viewerInfo = $this->refreshViewerUrl($tii, $check['tii_submission_id']);
+
+                $this->updateCheck($checkId, [
+                    'state'                => 3,
+                    'similarity_score'     => $score,
+                    'pdf_local_path'       => $localPdf,
+                    'view_only_url'        => $viewerInfo['url'],
+                    'view_only_url_expire' => $viewerInfo['expire'],
+                    'error_msg'            => '',
+                ]);
+                return;
+            }
+
+            if ($status === 'ERROR') {
+                $errMsg = isset($statusResp['error_code']) ? (string)$statusResp['error_code'] : 'Turnitin reported ERROR';
+                $this->markFailed($checkId, '[poll] ' . $errMsg);
+                return;
+            }
+
+            // PROCESSING 或其它中间态：继续轮询
+            if ($attempt >= self::MAX_POLL_ATTEMPTS) {
+                $this->markFailed($checkId, '[poll] timeout after ' . $attempt . ' attempts');
+                return;
+            }
+            Queue::later(
+                self::POLL_INTERVAL,
+                'app\\api\\job\\PlagiarismPoll',
+                ['check_id' => $checkId, 'attempt' => $attempt + 1],
+                'plagiarism'
+            );
+        } catch (\Throwable $e) {
+            // 网络抖动不要直接 fail，给一定容错次数
+            if ($attempt < self::MAX_POLL_ATTEMPTS) {
+                Queue::later(
+                    self::POLL_INTERVAL,
+                    'app\\api\\job\\PlagiarismPoll',
+                    ['check_id' => $checkId, 'attempt' => $attempt + 1],
+                    'plagiarism'
+                );
+                $this->updateCheck($checkId, [
+                    'attempts'  => $attempt,
+                    'error_msg' => '[poll] transient: ' . $e->getMessage(),
+                ]);
+                return;
+            }
+            $this->markFailed($checkId, '[poll] exhausted: ' . $e->getMessage());
+            throw $e;
+        }
+    }
+
+    /**
+     * 重新生成在线查看 URL（已有的过期了用）
+     *
+     * @return array{url:string, expire:int, local_pdf:string}
+     */
+    public function refreshViewerUrlFor($checkId)
+    {
+        $check = $this->mustGetCheck($checkId);
+        if (empty($check['tii_submission_id'])) {
+            throw new Exception('check has no tii_submission_id');
+        }
+        $tii = new TurnitinService();
+        $info = $this->refreshViewerUrl($tii, $check['tii_submission_id']);
+        $this->updateCheck($checkId, [
+            'view_only_url'        => $info['url'],
+            'view_only_url_expire' => $info['expire'],
+        ]);
+        return [
+            'url'       => $info['url'],
+            'expire'    => $info['expire'],
+            'local_pdf' => $check['pdf_local_path'],
+        ];
+    }
+
+    // ---------- 内部 ----------
+
+    private function refreshViewerUrl($tii, $submissionId)
+    {
+        $resp = $tii->getViewerUrl($submissionId);
+        $url = '';
+        if (isset($resp['viewer_url'])) {
+            $url = (string)$resp['viewer_url'];
+        } elseif (isset($resp['url'])) {
+            $url = (string)$resp['url'];
+        }
+        // 默认 2 小时过期，保守起见
+        return ['url' => $url, 'expire' => time() + 7200];
+    }
+
+    /**
+     * 触发生成 + 轮询 + 下载 PDF 到本地，返回相对路径
+     */
+    private function downloadAndStorePdf($tii, $submissionId, $checkId)
+    {
+        // 1. 请求生成
+        $req = $tii->requestPdfReport($submissionId);
+        $pdfId = isset($req['id']) ? $req['id'] : '';
+        if ($pdfId === '') {
+            throw new Exception('requestPdfReport empty id: ' . json_encode($req));
+        }
+
+        // 2. 内联轮询 PDF 状态（最多 3 分钟，每 6 秒一次）
+        $maxLoops = 30;
+        for ($i = 0; $i < $maxLoops; $i++) {
+            $st = $tii->getPdfReportStatus($submissionId, $pdfId);
+            $stCode = isset($st['status']) ? strtoupper($st['status']) : '';
+            if ($stCode === 'SUCCESS') {
+                break;
+            }
+            if ($stCode === 'FAILED') {
+                throw new Exception('PDF report generation failed: ' . json_encode($st));
+            }
+            sleep(6);
+        }
+        // 3. 下载
+        $binary = $tii->downloadPdfReport($submissionId, $pdfId);
+        if (!is_string($binary) || strlen($binary) < 100) {
+            throw new Exception('downloaded pdf is empty/too small');
+        }
+
+        // 4. 落盘
+        $rootDir = ROOT_PATH ?: dirname(dirname(__DIR__));
+        $absDir  = rtrim($rootDir, '/\\') . DIRECTORY_SEPARATOR . self::REPORT_DIR;
+        if (!is_dir($absDir)) {
+            @mkdir($absDir, 0755, true);
+        }
+        $filename = sprintf('check_%d_%s.pdf', $checkId, date('Ymd_His'));
+        $absPath  = $absDir . DIRECTORY_SEPARATOR . $filename;
+        $bytes    = file_put_contents($absPath, $binary);
+        if ($bytes === false || $bytes < 100) {
+            throw new Exception('failed to save pdf to ' . $absPath);
+        }
+        return self::REPORT_DIR . '/' . $filename;
+    }
+
+    private function mustGetCheck($checkId)
+    {
+        $row = Db::name('plagiarism_check')->where('check_id', $checkId)->find();
+        if (!$row) {
+            throw new Exception("plagiarism_check #{$checkId} not found");
+        }
+        return $row;
+    }
+
+    private function updateCheck($checkId, array $data)
+    {
+        $data['utime'] = time();
+        Db::name('plagiarism_check')->where('check_id', $checkId)->update($data);
+    }
+
+    private function markFailed($checkId, $errMsg)
+    {
+        $this->log("markFailed act");
+        $this->updateCheck($checkId, [
+            'state'     => 4,
+            'error_msg' => mb_substr($errMsg, 0, 1000),
+        ]);
+    }
+
+    /**
+     * 从 t_article_file 找到投稿主稿（manuscirpt）的本地绝对路径。
+     * file_url 在系统里可能是 URL 或相对路径，调用方负责保证可读。
+     *
+     * @return string 文件绝对路径，找不到时抛异常
+     */
+    public function locateArticleManuscript($articleId)
+    {
+        $row = Db::name('article_file')
+            ->where('article_id', $articleId)
+            ->where('type_name', 'manuscirpt')   // 历史拼写
+            ->order('file_id desc')
+            ->find();
+        if (!$row || empty($row['file_url'])) {
+            throw new Exception("article #{$articleId} has no manuscirpt file");
+        }
+        return $this->resolveFileUrlToLocal($row['file_url']);
+    }
+
+    /**
+     * 把 file_url（可能是 http URL 或相对路径）解析成本地绝对路径。
+     * 不同环境部署可能有差异，这里用 .env 配置的 STATIC_ROOT 作前缀。
+     */
+    public function resolveFileUrlToLocal($fileUrl)
+    {
+        $fileUrl = trim((string)$fileUrl);
+        if ($fileUrl === '') {
+            throw new Exception('empty file_url');
+        }
+        // 已是绝对路径
+        if (preg_match('/^([a-zA-Z]:[\\\\\/]|\/)/', $fileUrl) && is_file($fileUrl)) {
+            return $fileUrl;
+        }
+
+        $staticRoot = trim((string)Env::get('plagiarism.static_root', ''));
+        $cdnPrefix  = trim((string)Env::get('plagiarism.cdn_prefix', ''));
+
+        // 是 http URL：先试着剥掉 cdn 前缀，映射到本地
+        if (preg_match('#^https?://#i', $fileUrl)) {
+            if ($cdnPrefix !== '' && stripos($fileUrl, $cdnPrefix) === 0) {
+                $rel = ltrim(substr($fileUrl, strlen($cdnPrefix)), '/');
+                $local = rtrim($staticRoot, '/\\') . DIRECTORY_SEPARATOR . $rel;
+                if (is_file($local)) {
+                    return $local;
+                }
+            }
+            // 实在不行，下载到 runtime/plagiarism/tmp 临时目录
+            return $this->downloadRemoteFile($fileUrl);
+        }
+
+        // 相对路径：拼 static_root
+        if ($staticRoot !== '') {
+            $local = rtrim($staticRoot, '/\\') . DIRECTORY_SEPARATOR . ltrim($fileUrl, '/\\');
+            if (is_file($local)) {
+                return $local;
+            }
+        }
+
+        throw new Exception("cannot resolve file_url to local path: {$fileUrl} (set [plagiarism] STATIC_ROOT/CDN_PREFIX in .env)");
+    }
+
+    private function downloadRemoteFile($url)
+    {
+        $rootDir = ROOT_PATH ?: dirname(dirname(__DIR__));
+        $tmpDir  = rtrim($rootDir, '/\\') . DIRECTORY_SEPARATOR . self::REPORT_DIR . DIRECTORY_SEPARATOR . 'tmp';
+        if (!is_dir($tmpDir)) {
+            @mkdir($tmpDir, 0755, true);
+        }
+        $ext = pathinfo(parse_url($url, PHP_URL_PATH), PATHINFO_EXTENSION) ?: 'pdf';
+        $local = $tmpDir . DIRECTORY_SEPARATOR . md5($url) . '_' . time() . '.' . $ext;
+
+        $ch = curl_init($url);
+        $fh = fopen($local, 'wb');
+        curl_setopt_array($ch, [
+            CURLOPT_FILE => $fh,
+            CURLOPT_FOLLOWLOCATION => true,
+            CURLOPT_TIMEOUT => 120,
+            CURLOPT_SSL_VERIFYPEER => false,
+        ]);
+        $ok = curl_exec($ch);
+        $code = curl_getinfo($ch, CURLINFO_HTTP_CODE);
+        curl_close($ch);
+        fclose($fh);
+        if (!$ok || $code !== 200 || filesize($local) < 100) {
+            @unlink($local);
+            throw new Exception("download failed url={$url} http={$code}");
+        }
+        return $local;
+    }
+
+    public function getCheck($checkId)
+    {
+        return Db::name('plagiarism_check')->where('check_id', $checkId)->find();
+    }
+
+    public function log($msg)
+    {
+        $line = date('Y-m-d H:i:s') . ' ' . $msg . PHP_EOL;
+        @file_put_contents($this->logFile, $line, FILE_APPEND);
+    }
+}

application/common/ProductionArticleRefer.php

@@ -0,0 +1,303 @@
+<?php
+namespace app\common;
+use think\Db;
+use think\Env;
+class ProductionArticleRefer
+{
+
+    // DOI匹配核心正则（生产级优化：支持%字符、限制长度、单词边界断言）
+    // private const DOI_PATTERN = '/(?:doi[:\s]?|DOI[:\s]?)?\b10\.\d+(?:\.\d+)*\/[a-zA-Z0-9._\-!()%\/:;@$&+=?#[\]<>~`|^]+/i';
+
+    // 错误码与错误信息映射（标准化错误处理）
+    private const ERROR_CODES = [
+        'EMPTY_STRING' => 'Input string is empty (preprocessed)）',
+        'NO_MATCH' => 'No valid DOI detected',
+        'INVALID_AFTER_CLEAN' => 'No effective DOI after cleaning',
+        'FORCE_EXTRACT_FAILED' => 'Forced extraction still has no valid DOI',
+        'EXTRACTION_EXCEPTION' => 'Exception occurred during DOI extraction process',
+    ];
+
+    /**
+     * 获取未处理的参考文献
+     *
+     * @return void
+     */
+    public function top($aParam = []) {
+
+        //文章ID
+        $iArticleId = empty($aParam['article_id']) ? '' : $aParam['article_id'];
+        if(empty($iArticleId)){
+            return json_encode(array('status' => 2,'msg' => 'Please select an article'.json_encode($aParam) ));
+        }
+        // 获取生产文章ID
+        $iPArticleId = empty($aParam['p_article_id']) ? 0 : $aParam['p_article_id'];
+        if(empty($iPArticleId)) {
+            return json_encode(array('status' => 2,'msg' => 'Please select an production article'.json_encode($aParam) ));
+        }
+
+        //查询未处理过的数据
+        $aWhere = ['p_article_id' => $iPArticleId,'article_id' => $iArticleId,'state' => 0,'refer_doi' => ['<>',''],'is_deal' => 2];
+        $aResult = Db::name('production_article_refer')->field('article_id,p_article_id,p_refer_id,refer_doi')->where($aWhere)->select();
+        if(empty($aResult)){
+            return json_encode(array('status' => 2,'msg' => 'The reference data to be processed is empty'.json_encode($aParam)));
+        }
+
+        //数据处理
+        foreach ($aResult as $key => $value) {
+            if(empty($value['refer_doi'])){
+                continue;
+            }
+            //调用获取参考文献详情队列
+            \think\Queue::push('app\api\job\ArticleReferDetailQueue@fire', $value, 'ArticleReferDetailQueue');
+        }
+        return json_encode(['status' => 1,'msg' => 'Add to reference processing queue']);
+    }
+    /**
+     * 处理参考文献
+     *
+     * @return void
+     */
+    public function get($aParam = []) {
+        // 获取生产文章ID
+        $iPReferId = empty($aParam['p_refer_id']) ? 0 : $aParam['p_refer_id'];
+        if(empty($iPReferId)) {
+            return json_encode(array('status' => 2,'msg' => 'Please select a reference'.json_encode($aParam) ));
+        }
+        // 获取生产文章ID
+        $iPArticleId = empty($aParam['p_article_id']) ? 0 : $aParam['p_article_id'];
+        if(empty($iPArticleId)) {
+            return json_encode(array('status' => 2,'msg' => 'Please select an production article'.json_encode($aParam) ));
+        }
+        //查询未处理过的数据
+        $aWhere = ['p_refer_id' => $iPReferId,'p_article_id' => $iPArticleId,'state' => 0];
+        $aRefer = Db::name('production_article_refer')->field('refer_doi,refer_content')->where($aWhere)->find();
+        if(empty($aRefer)){
+            return json_encode(array('status' => 2,'msg' => 'No reference records found'.json_encode($aParam)));
+        }
+        if(empty($aRefer['refer_doi'])){
+            return json_encode(['status' => 4,'msg' => 'Reference DOI is empty'.json_encode($aParam)]);
+        }
+
+        //数据处理
+        $doi = str_replace('/', '%2F', $aRefer['refer_doi']);
+        $url = "https://citation.doi.org/format?doi=$doi&style=cancer-translational-medicine&lang=en-US";
+        $res = myGet($url);
+        $frag = trim(substr($res, strpos($res, '.') + 1));
+        if(empty($frag)){
+            $aUpdate = ['refer_frag' => $aRefer['refer_content'],'refer_type' => 'other','is_deal' => 1,'update_time' => time()];
+            $aWhere = ['p_refer_id' => $iPReferId];
+            $result = Db::name('production_article_refer')->where($aWhere)->limit(1)->update($aUpdate);
+            //写入通过AI获取参考文献详情队列
+            // \think\Queue::push('app\api\job\AiCheckReferByDoi@fire',$aParam,'AiCheckReferByDoi');
+            return json_encode(array('status' => 2,'msg' => 'The data obtained from the interface is empty'.$url));
+        }
+
+        //整理数据入库
+        $update = [];
+        if (mb_substr_count($frag, '.') != 3){
+            $f = $frag . " Available at: " . PHP_EOL . "https://doi.org/" . $aRefer['refer_doi'];
+            $update['refer_type'] = "other";
+            $update['refer_frag'] = $f;
+            $update['cs'] = 1;
+            //写入通过AI获取参考文献详情队列
+            // \think\Queue::push('app\api\job\AiCheckReferByDoi@fire',$aParam,'AiCheckReferByDoi');
+        } 
+        if (mb_substr_count($frag, '.') == 3){
+            $res = explode('.', $frag);
+            $update['author'] = prgeAuthor($res[0]);
+            $update['title'] = trim($res[1]);
+            $bj = bekjournal($res[2]);
+            $joura = formateJournal(trim($bj[0]));
+            $update['joura'] = $joura;
+            $is_js = 0;
+            if ($joura == trim($bj[0])) {
+            }
+            $update['refer_type'] = "journal";
+            $update['is_ja'] = $joura == trim($bj[0]) ? 0 : 1;
+            $update['dateno'] = str_replace(' ', '', str_replace('-', '–', trim($bj[1])));
+            //新增处理 期卷页码 20251127 start
+            if(!empty($update['dateno'])){
+                $sStr = $update['dateno'];
+                $aStr = explode(':', $sStr);
+                if(!empty($aStr[1])){
+                    $parts = explode('–', $aStr[1]);
+                    if(count($parts) == 2){
+                        $prefix = empty($parts[0]) ? 0 : intval($parts[0]);
+                        $suffix = empty($parts[1]) ? 0 : intval($parts[1]);
+                        if($prefix > $suffix){
+                            $prefixLen = strlen($prefix);
+                            $suffixLen = strlen($suffix);
+                            $missingLen = $prefixLen - $suffixLen;
+                            if ($missingLen > 0) {
+                                $fillPart = substr($prefix, 0, $missingLen);
+                                $newSuffix = $fillPart . $suffix;
+                                $update['dateno'] = $aStr[0].':'.$prefix.'–'.$newSuffix;
+                            }
+                        }
+                    }   
+                }
+                // if(empty($aStr[1])){
+                //     //写入通过AI获取参考文献详情队列
+                //     \think\Queue::push('app\api\job\AiCheckReferByDoi@fire',$aParam,'AiCheckReferByDoi');
+                // }
+            }
+            //新增处理 期卷页码 20251127 end
+            $update['doilink'] = strpos($aRefer['refer_doi'],"http")===false?"https://doi.org/" . $aRefer['refer_doi']:$aRefer['refer_doi'];
+            $update['cs'] = 1;
+        }
+        //数据库更新
+        if(empty($update)){
+            return json_encode(array('status' => 3,'msg' => 'Update data to empty'.$url.'====='.$frag));
+        }
+        $aWhere = ['p_refer_id' => $iPReferId];
+        $update += ['is_deal' => 1,'update_time' => time()];
+        $result = Db::name('production_article_refer')->where($aWhere)->limit(1)->update($update);
+        if($result === false){
+            return json_encode(array('status' => 3,'msg' => 'Update failed'.json_encode($update)));
+        }
+        return json_encode(['status' => 1,'msg' => 'Update successful']);
+    }
+
+    // /**
+    //  * 实例方法：提取单个DOI（核心逻辑，生产级优化）
+    //  * @param string $str 待检测字符串
+    //  * @param bool $standardize 是否标准化DOI（转小写）
+    //  * @param bool $forceExtract 是否强制提取（忽略微小格式瑕疵）
+    //  * @return array 提取结果（含错误码、错误信息、DOI）
+    //  */
+    // // public function extractDoiFromString(string $str, bool $standardize = true, bool $forceExtract = false): array
+    // // {
+    // //     // 初始化标准化结果
+    // //     $result = [
+    // //         'has_doi'    => false,
+    // //         'doi'        => null,
+    // //         'error_code' => null,
+    // //         'error_msg'  => null,
+    // //     ];
+
+    // //     try {
+    // //         // 严格类型校验（防止非字符串参数传入）
+    // //         if (!is_string($str)) {
+    // //             throw new InvalidArgumentException('输入参数必须为字符串类型', 1001);
+    // //         }
+    // //         // 字符串预处理（生产级：全角转半角、URL解码、HTML标签移除等）
+    // //         $processedStr = $this->preprocessString($str);
+    // //         if (trim($processedStr) === '') {
+    // //             $result['error_code'] = 'EMPTY_STRING';
+    // //             $result['error_msg'] = self::ERROR_CODES['EMPTY_STRING'];
+    // //             return $result;
+    // //         }
+
+    // //         // 性能优化：用preg_match仅匹配首个DOI，替代preg_match_all
+    // //         // 优化后的带前缀版正则
+    // //         $pattern = '/(?:doi[:\s]*|DOI[:\s]*)?\b10\.\d+(?:\.\d+)*\/[a-zA-Z0-9._\-!()%\/:;@$&+=?#[\]<>~`|^'"{},\\\\]+(?![\w?#])/i";
+    // //         if (!preg_match($pattern, $processedStr, $match)) {
+    // //             $result['error_code'] = 'NO_MATCH';
+    // //             $result['error_msg'] = self::ERROR_CODES['NO_MATCH'];
+    // //             return $result;
+    // //         }
+
+    // //         // 清洗并验证首个DOI
+    // //         $cleanDoi = $this->cleanAndValidateDoi($match[0], $standardize, $forceExtract);
+    // //         if ($cleanDoi !== null) {
+    // //             $result['has_doi'] = true;
+    // //             $result['doi'] = $cleanDoi;
+    // //         } else {
+    // //             // 根据是否强制提取设置错误信息
+    // //             $errorKey = $forceExtract ? 'FORCE_EXTRACT_FAILED' : 'INVALID_AFTER_CLEAN';
+    // //             $result['error_code'] = $errorKey;
+    // //             $result['error_msg'] = self::ERROR_CODES[$errorKey];
+    // //         }
+
+    // //     } catch (InvalidArgumentException $e) {
+    // //         // 业务异常：标准化错误码和信息
+    // //         $result['error_code'] = 'INVALID_PARAM';
+    // //         $result['error_msg'] = '参数错误：' . $e->getMessage();
+    // //     } catch (Exception $e) {
+    // //         // 系统异常：隐藏敏感信息，记录通用错误
+    // //         $result['error_code'] = 'EXTRACTION_EXCEPTION';
+    // //         $result['error_msg'] = self::ERROR_CODES['EXTRACTION_EXCEPTION'] . '：' . $e->getMessage();
+    // //     }
+
+    // //     return $result;
+    // // }
+
+    // // /**
+    // //  * 字符串预处理（生产级：覆盖所有编码/格式干扰场景）
+    // //  * @param string $str 原始字符串
+    // //  * @return string 预处理后的纯净字符串
+    // //  */
+    // // private function preprocessString(string $str): string
+    // // {
+    // //     // 1. 全角转半角（解决中文全角字符干扰，如１０．１００７／ｓ１１０４２－０２０－１０１０３－４）
+    // //     $str = $this->fullWidthToHalfWidth($str);
+    // //     // 2. 移除所有HTML标签（解决网页文本中DOI被<p>/<a>/<b>等标签包裹的问题）
+    // //     $str = strip_tags($str);
+    // //     // 3. URL解码（处理%2F等URL编码的特殊字符，如10.1007%2Fs11042-020-10103-4）
+    // //     $str = urldecode($str);
+    // //     // 4. 解码HTML实体（处理&amp;、&#47;等HTML实体编码）
+    // //     $str = html_entity_decode($str, ENT_QUOTES | ENT_HTML5, 'UTF-8');
+    // //     // 5. 移除不可见字符（换行、制表符、零宽空格、控制字符等）
+    // //     $str = preg_replace('/[\x00-\x1F\x7F\x{200B}-\x{200F}]/u', ' ', $str);
+    // //     // 6. 合并多个空格为单个（避免连续空格干扰正则匹配）
+    // //     $str = preg_replace('/\s+/', ' ', $str);
+
+    // //     return $str;
+    // // }
+
+    // // /**
+    // //  * 清洗并验证DOI（生产级：优化正则规则，严格长度校验）
+    // //  * @param string $match 原始正则匹配结果
+    // //  * @param bool $standardize 是否标准化DOI（转小写）
+    // //  * @param bool $forceExtract 是否强制提取
+    // //  * @return string|null 有效DOI或null
+    // //  */
+    // // private function cleanAndValidateDoi(string $match, bool $standardize, bool $forceExtract): ?string
+    // // {
+    // //     // 1. 移除DOI前缀（doi:/DOI:）和首尾空白字符
+    // //     $cleanDoi = preg_replace('/^doi[:\s]?|^DOI[:\s]?/i', '', trim($match));
+    // //     // 2. 移除尾部常见标点（避免DOI被标点包裹，如10.1007/s11042-020-10103-4.）
+    // //     $cleanDoi = rtrim($cleanDoi, '.,;(){}[]!?"\'');
+
+    // //     // 3. 严格的长度校验（DOI官方规范：6-200字符）
+    // //     $doiLength = strlen($cleanDoi);
+    // //     if ($doiLength < 6 || $doiLength > 200) {
+    // //         return null;
+    // //     }
+
+    // //     // 4. 验证规则（生产级优化：添加单词边界，避免匹配不完整DOI）
+    // //     // 基础规则：严格遵循官方规范，10.开头+包含/+/后有内容
+    // //     $basicRule = '/^10\.\d+\/.+$/D';
+    // //     // 宽松规则：强制提取时使用，添加单词边界，避免匹配被字符包裹的DOI
+    // //     $looseRule = '/\b10\.\d+\/[^\s%]{1,190}\b/';
+
+    // //     $validateRule = $forceExtract ? $looseRule : $basicRule;
+    // //     $isValid = preg_match($validateRule, $cleanDoi) === 1;
+
+    // //     // 5. 验证通过则标准化（转小写），否则返回null
+    // //     if ($isValid) {
+    // //         return $standardize ? strtolower($cleanDoi) : $cleanDoi;
+    // //     }
+
+    // //     return null;
+    // // }
+
+    // // /**
+    // //  * 辅助方法：全角转半角
+    // //  * @param string $str 包含全角字符的字符串
+    // //  * @return string 半角字符串
+    // //  */
+    // // private function fullWidthToHalfWidth(string $str): string
+    // // {
+    // //     $fullWidthChars = [
+    // //         '０' => '0', '１' => '1', '２' => '2', '３' => '3', '４' => '4',
+    // //         '５' => '5', '６' => '6', '７' => '7', '８' => '8', '９' => '9',
+    // //         '．' => '.', '／' => '/', '－' => '-', '％' => '%', '！' => '!',
+    // //         '（' => '(', '）' => ')', '：' => ':', '；' => ';', '，' => ',',
+    // //         '＂' => '"', '＇' => '\''
+    // //     ];
+
+    // //     return strtr($str, $fullWidthChars);
+    // // }
+}
+?>

application/common/PromotionLlmService.php

@@ -0,0 +1,339 @@
+<?php
+
+namespace app\common;
+
+use think\Env;
+
+/**
+ * 推广邮件 LLM 服务
+ *
+ * 作用：基于 expert 的代表作 (paper_title)、expert 的研究领域、期刊信息 (name / scope / fields 等)，
+ *      通过一次 LLM 调用同时生成两段邮件内容：
+ *        - description    个性化赞美 + 邀稿段落（对应模板变量 {{ai_content_analysis}} / {{llm_description}}）
+ *        - advised_topics "我们特别关注 X / Y / Z 领域" 段落（对应模板变量 {{ai_advised_topics}}）
+ *
+ * 单次调用返回 JSON，避免两次 LLM 请求导致的延迟翻倍；任一段失败可独立兜底。
+ *
+ * 配置（.env 的 [promotion] 段）：
+ *   PROMOTION_LLM_URL               chat/completions 接口地址
+ *   PROMOTION_LLM_MODEL             模型名
+ *   PROMOTION_LLM_TIMEOUT           超时时间(秒)，默认 30
+ *   PROMOTION_LLM_FALLBACK          description 兜底文案
+ *   PROMOTION_LLM_ADVISED_FALLBACK  advised_topics 兜底文案
+ *
+ * 状态约定：
+ *   1 = LLM 成功生成
+ *   2 = LLM 调用失败 / 解析失败，使用兜底
+ *   3 = 前置条件不足（缺代表作、缺交集等），使用兜底
+ */
+class PromotionLlmService
+{
+    private $url;
+    private $model;
+    private $timeout;
+    private $apiKey;
+    private $fallback;
+    private $advisedFallback;
+
+    public function __construct()
+    {
+        $this->url      = trim((string)Env::get('promotion.promotion_llm_url', ''));
+        $this->model    = trim((string)Env::get('promotion.promotion_llm_model', ''));
+        $this->timeout  = max(5, intval(Env::get('promotion.promotion_llm_timeout', 30)));
+        $this->apiKey   = trim((string)Env::get('promotion.promotion_llm_api_key', ''));
+        $this->fallback = trim((string)Env::get('promotion.promotion_llm_fallback',
+            'Your recent work aligns closely with the scope of our journal, and we would be honored to consider a contribution from you.'));
+        $this->advisedFallback = trim((string)Env::get('promotion.promotion_llm_advised_fallback',
+            'We are especially interested in the research directions that align with your expertise, and warmly welcome your future submissions in these areas.'));
+    }
+
+    /**
+     * 一次 LLM 调用生成邮件两段内容。
+     *
+     * @param array $expert         expert 行（含 name / representative_work_title / fields 等）
+     * @param array $journal        journal 行（含 title / aims / databases 等）
+     * @param array $overlapFields  预先计算的 expert & journal 领域交集（可能为空）
+     * @param array $journalFields  journal（或 task/工厂）的目标领域
+     * @param array $expertFields   expert 的研究领域
+     * @return array [
+     *   'description'            => string,
+     *   'description_status'     => 1|2|3,
+     *   'advised_topics'         => string,
+     *   'advised_topics_status'  => 1|2|3,
+     * ]
+     */
+    public function generateEmailContent(
+        array $expert,
+        array $journal,
+        array $overlapFields = [],
+        array $journalFields = [],
+        array $expertFields = []
+    ): array {
+        $paperTitle  = trim((string)($expert['representative_work_title'] ?? ''));
+        $expertName  = trim((string)($expert['name'] ?? ''));
+        $journalName = trim((string)($journal['title'] ?? ''));
+
+        $overlapList  = $this->cleanList($overlapFields);
+        $journalList  = $this->cleanList($journalFields);
+        $expertList   = $this->cleanList($expertFields);
+
+        $hasDescInput    = ($paperTitle !== '' && $journalName !== '');
+        $hasAdvisedInput = (!empty($overlapList) || (!empty($journalList) && !empty($expertList)));
+
+        // 两段都缺输入时直接走双兜底
+        if (!$hasDescInput && !$hasAdvisedInput) {
+            return $this->allFallback(3, 3);
+        }
+
+        // LLM 未配置
+        if ($this->url === '' || $this->model === '') {
+            return $this->allFallback(
+                $hasDescInput ? 2 : 3,
+                $hasAdvisedInput ? 2 : 3
+            );
+        }
+
+        $expertField = trim((string)($expert['fields'] ?? ($expert['field'] ?? '')));
+        $journalAims = trim((string)($journal['aims'] ?? ''));
+        $journalDbs  = trim((string)($journal['databases'] ?? ''));
+
+        // advised_topics 目标数量：期望邮件里总是体现 3 个推荐方向
+        $targetTopicCount = 3;
+        $overlapCount = count($overlapList);
+        $needFill = max(0, $targetTopicCount - $overlapCount);
+
+        $system = 'You are an academic editorial assistant. '
+            . 'You will receive context about an author, their recent paper, and a target journal, '
+            . 'and you must produce TWO English paragraphs for an invitation email. '
+            . 'Output STRICT MINIFIED JSON ONLY with exactly these keys:'
+            . '{"description":"...","advised_topics":"..."} '
+            . 'Rules for "description": 2-3 sentences, <=50 words, warm and professional; '
+            . '(a) briefly appreciate the author\'s recent paper, '
+            . '(b) explain why it fits the journal\'s scope, '
+            . '(c) gently invite a future submission. '
+            . 'Rules for "advised_topics": 1-2 sentences, <=60 words; '
+            . 'Must recommend EXACTLY ' . $targetTopicCount . ' research directions the journal is particularly interested in, '
+            . 'and explicitly list these ' . $targetTopicCount . ' topics in the sentence. '
+            . 'Topic selection rules, in order: '
+            . '(i) include all items from "Overlap topics (exact match)" first, preserving their order; '
+            . '(ii) if overlap has fewer than ' . $targetTopicCount . ' items, fill the remaining slots '
+            . 'with topics that are semantically close to BOTH the journal\'s focus fields and the author\'s research fields '
+            . '(prefer concrete sub-topics over generic umbrella terms; '
+            . 'do NOT invent topics unrelated to either list); '
+            . '(iii) never repeat topics, keep wording concise (2-4 words each). '
+            . 'End the paragraph by inviting contributions leaning toward those directions. '
+            . 'No greetings, no signatures, no placeholders, no markdown, no code fences. '
+            . 'If a section genuinely cannot be produced, return an empty string for that key.';
+
+        $userLines = [];
+        $userLines[] = 'Author name: ' . ($expertName !== '' ? $expertName : '(unknown)');
+        if ($expertField !== '') {
+            $userLines[] = 'Author research field (raw): ' . $expertField;
+        }
+        if (!empty($expertList)) {
+            $userLines[] = 'Author research fields (list): ' . implode(', ', $expertList);
+        }
+        $userLines[] = 'Recent paper title: ' . ($paperTitle !== '' ? $paperTitle : '(none)');
+        $userLines[] = 'Target journal: ' . ($journalName !== '' ? $journalName : '(unknown)');
+        if (!empty($journalList)) {
+            $userLines[] = 'Journal focus fields: ' . implode(', ', $journalList);
+        }
+        if (!empty($overlapList)) {
+            $userLines[] = 'Overlap topics (exact match): ' . implode(', ', $overlapList)
+                . ' [count=' . $overlapCount . ', need_fill=' . $needFill . ']';
+        } else {
+            $userLines[] = 'Overlap topics (exact match): (none) '
+                . '[count=0, need_fill=' . $targetTopicCount
+                . '; please pick ' . $targetTopicCount
+                . ' semantically related topics from the two field lists above]';
+        }
+        if ($journalAims !== '') {
+            $userLines[] = 'Journal aims & scope: ' . mb_substr($journalAims, 0, 500);
+        }
+        if ($journalDbs !== '') {
+            $userLines[] = 'Journal indexing: ' . mb_substr($journalDbs, 0, 200);
+        }
+        $userLines[] = 'Advised topics required count: ' . $targetTopicCount;
+        $userLines[] = 'Return only minified JSON {"description":"...","advised_topics":"..."}.';
+        $user = implode("\n", $userLines);
+
+        $payload = [
+            'model'       => $this->model,
+            'temperature' => 0.4,
+            'messages'    => [
+                ['role' => 'system', 'content' => $system],
+                ['role' => 'user',   'content' => $user],
+            ],
+        ];
+
+        $content = $this->postChat($payload);
+        if ($content === null) {
+            return $this->allFallback(
+                $hasDescInput ? 2 : 3,
+                $hasAdvisedInput ? 2 : 3
+            );
+        }
+
+        $parsed = $this->parseJson($content);
+        if ($parsed === null) {
+            return $this->allFallback(
+                $hasDescInput ? 2 : 3,
+                $hasAdvisedInput ? 2 : 3
+            );
+        }
+
+        $desc    = $this->cleanContent((string)($parsed['description'] ?? ''));
+        $advised = $this->cleanContent((string)($parsed['advised_topics'] ?? ''));
+
+        $descStatus = 1;
+        if ($desc === '') {
+            $desc = $this->fallback;
+            $descStatus = $hasDescInput ? 2 : 3;
+        }
+
+        $advisedStatus = 1;
+        if ($advised === '') {
+            $advised = $this->advisedFallback;
+            $advisedStatus = $hasAdvisedInput ? 2 : 3;
+        }
+
+        return [
+            'description'           => $desc,
+            'description_status'    => $descStatus,
+            'advised_topics'        => $advised,
+            'advised_topics_status' => $advisedStatus,
+        ];
+    }
+
+    /**
+     * 兼容旧接口：单段 description 生成（仍走 generateEmailContent，取第一段）
+     *
+     * @return array ['status' => 1|2|3, 'text' => string]
+     */
+    public function generateDescription(array $expert, array $journal): array
+    {
+        $r = $this->generateEmailContent($expert, $journal);
+        return [
+            'status' => intval($r['description_status']),
+            'text'   => (string)$r['description'],
+        ];
+    }
+
+    /**
+     * 调用 chat/completions 接口，返回 content 字符串；失败返回 null。
+     */
+    private function postChat(array $payload)
+    {
+        $ch = curl_init();
+        curl_setopt($ch, CURLOPT_URL, $this->url);
+        curl_setopt($ch, CURLOPT_POST, true);
+        curl_setopt($ch, CURLOPT_POSTFIELDS, json_encode($payload, JSON_UNESCAPED_UNICODE));
+        curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
+        curl_setopt($ch, CURLOPT_CONNECTTIMEOUT, min(10, $this->timeout));
+        curl_setopt($ch, CURLOPT_TIMEOUT, $this->timeout);
+
+        $headers = ['Content-Type: application/json'];
+        if ($this->apiKey !== '') {
+            $headers[] = 'Authorization: Bearer ' . $this->apiKey;
+        }
+        curl_setopt($ch, CURLOPT_HTTPHEADER, $headers);
+
+        $raw = curl_exec($ch);
+        if ($raw === false) {
+            curl_close($ch);
+            return null;
+        }
+        $httpCode = intval(curl_getinfo($ch, CURLINFO_HTTP_CODE));
+        curl_close($ch);
+        if ($httpCode < 200 || $httpCode >= 300) {
+            return null;
+        }
+
+        $data = json_decode($raw, true);
+        if (!is_array($data)) return null;
+
+        if (isset($data['choices'][0]['message']['content'])) {
+            return (string)$data['choices'][0]['message']['content'];
+        }
+        if (isset($data['content'])) {
+            return (string)$data['content'];
+        }
+        return null;
+    }
+
+    /**
+     * 从模型原始输出里提取 JSON 对象；失败返回 null。
+     */
+    private function parseJson(string $raw)
+    {
+        $raw = trim($raw);
+        if ($raw === '') return null;
+        $raw = preg_replace('/^```[a-zA-Z]*\s*|```$/m', '', $raw);
+        $raw = trim($raw);
+
+        // 直接 decode
+        $obj = json_decode($raw, true);
+        if (is_array($obj)) return $obj;
+
+        // 抓出第一个 {...} 块再 decode
+        if (preg_match('/\{.*\}/s', $raw, $m)) {
+            $obj = json_decode($m[0], true);
+            if (is_array($obj)) return $obj;
+        }
+        return null;
+    }
+
+    /**
+     * 清洗 LLM 输出：去除 markdown 包裹、多余空白、首尾引号、过长截断。
+     */
+    private function cleanContent(string $text): string
+    {
+        $text = trim($text);
+        $text = preg_replace('/^```[a-zA-Z]*\s*|```$/m', '', $text);
+        $text = trim($text);
+        $text = trim($text, "\"' \t\n\r\0\x0B");
+        $text = preg_replace('/\s+/', ' ', $text);
+        if (mb_strlen($text) > 800) {
+            $text = mb_substr($text, 0, 800);
+        }
+        return trim($text);
+    }
+
+    /**
+     * 对领域列表做 trim / 去空 / 去重，保留首次出现顺序。
+     */
+    private function cleanList(array $list): array
+    {
+        $out = [];
+        $seen = [];
+        foreach ($list as $item) {
+            $v = trim((string)$item);
+            if ($v === '') continue;
+            $key = strtolower($v);
+            if (isset($seen[$key])) continue;
+            $seen[$key] = true;
+            $out[] = $v;
+        }
+        return $out;
+    }
+
+    private function allFallback(int $descStatus, int $advisedStatus): array
+    {
+        return [
+            'description'           => $this->fallback,
+            'description_status'    => $descStatus,
+            'advised_topics'        => $this->advisedFallback,
+            'advised_topics_status' => $advisedStatus,
+        ];
+    }
+
+    public function getFallback(): string
+    {
+        return $this->fallback;
+    }
+
+    public function getAdvisedFallback(): string
+    {
+        return $this->advisedFallback;
+    }
+}

application/common/PubmedService.php

@@ -0,0 +1,237 @@
+<?php
+
+namespace app\common;
+
+/**
+ * PubMed 工具类（E-utilities）
+ *
+ * 功能：
+ * - DOI -> PMID
+ * - PMID -> 文章结构化信息（title/abstract/mesh/publication_types/year/journal）
+ *
+ * 说明：
+ * - 默认使用 runtime 文件缓存，避免重复请求 NCBI
+ */
+class PubmedService
+{
+    private $base = 'https://eutils.ncbi.nlm.nih.gov/entrez/eutils/';
+    private $timeout = 20;
+    private $tool = 'tmrjournals';
+    private $email = '';
+
+    public function __construct(array $config = [])
+    {
+        if (isset($config['base'])) $this->base = rtrim((string)$config['base'], '/') . '/';
+        if (isset($config['timeout'])) $this->timeout = max(5, intval($config['timeout']));
+        if (isset($config['tool'])) $this->tool = (string)$config['tool'];
+        if (isset($config['email'])) $this->email = (string)$config['email'];
+    }
+
+    /**
+     * DOI -> PMID（优先用 [DOI]，命中不到再用 [AID]）
+     */
+    public function doiToPmid(string $doi): ?string
+    {
+        $doi = trim($doi);
+        if ($doi === '') return null;
+
+        $cacheKey = 'doi2pmid_' . sha1(strtolower($doi));
+        $cached = $this->cacheGet($cacheKey, 30 * 86400);
+        if (is_string($cached) && $cached !== '') {
+            return $cached;
+        }
+
+        $pmid = $this->esearch($doi . '[DOI]');
+        if (!$pmid) {
+            $pmid = $this->esearch($doi . '[AID]');
+        }
+        if ($pmid) {
+            $this->cacheSet($cacheKey, $pmid);
+            return $pmid;
+        }
+        return null;
+    }
+
+    /**
+     * PMID -> 文章信息（title/abstract/mesh/publication_types/year/journal）
+     */
+    public function fetchByPmid(string $pmid): ?array
+    {
+        $pmid = trim($pmid);
+        if ($pmid === '') return null;
+
+        $cacheKey = 'pmid_' . $pmid;
+        $cached = $this->cacheGet($cacheKey, 30 * 86400);
+        if (is_array($cached)) return $cached;
+
+        $url = $this->base . 'efetch.fcgi?' . http_build_query([
+            'db' => 'pubmed',
+            'id' => $pmid,
+            'retmode' => 'xml',
+            'tool' => $this->tool,
+            'email' => $this->email,
+        ]);
+
+        $xml = $this->httpGet($url);
+        if (!is_string($xml) || trim($xml) === '') return null;
+
+        $data = $this->parseEfetchXml($xml);
+        if (!$data) return null;
+
+        $this->cacheSet($cacheKey, $data);
+        return $data;
+    }
+
+    /**
+     * DOI -> PubMed 信息（含 abstract/mesh）
+     */
+    public function fetchByDoi(string $doi): ?array
+    {
+        $pmid = $this->doiToPmid($doi);
+        if (!$pmid) return null;
+        $info = $this->fetchByPmid($pmid);
+        if (!$info) return null;
+        $info['pmid'] = $pmid;
+        $info['doi'] = $doi;
+        return $info;
+    }
+
+    // ----------------- Internals -----------------
+
+    private function esearch(string $term): ?string
+    {
+        $url = $this->base . 'esearch.fcgi?' . http_build_query([
+            'db' => 'pubmed',
+            'retmode' => 'json',
+            'retmax' => 1,
+            'term' => $term,
+            'tool' => $this->tool,
+            'email' => $this->email,
+        ]);
+
+        $res = $this->httpGet($url);
+        $json = json_decode((string)$res, true);
+        $ids = $json['esearchresult']['idlist'] ?? [];
+        if (!empty($ids[0])) return (string)$ids[0];
+        return null;
+    }
+
+    private function parseEfetchXml(string $xml): ?array
+    {
+        libxml_use_internal_errors(true);
+        $doc = new \DOMDocument();
+        if (!$doc->loadXML($xml)) {
+            return null;
+        }
+        $xp = new \DOMXPath($doc);
+
+        $title = $this->xpText($xp, '//PubmedArticle//ArticleTitle');
+
+        $abstractParts = [];
+        $absNodes = $xp->query('//PubmedArticle//Abstract//AbstractText');
+        if ($absNodes) {
+            foreach ($absNodes as $n) {
+                $label = $n->attributes && $n->attributes->getNamedItem('Label')
+                    ? trim($n->attributes->getNamedItem('Label')->nodeValue)
+                    : '';
+                $txt = trim($n->textContent);
+                if ($txt === '') continue;
+                $abstractParts[] = $label ? ($label . ': ' . $txt) : $txt;
+            }
+        }
+        $abstract = trim(implode("\n", $abstractParts));
+
+        $mesh = [];
+        $meshNodes = $xp->query('//PubmedArticle//MeshHeadingList//MeshHeading//DescriptorName');
+        if ($meshNodes) {
+            foreach ($meshNodes as $n) {
+                $t = trim($n->textContent);
+                if ($t !== '') $mesh[] = $t;
+            }
+        }
+        $mesh = array_values(array_unique($mesh));
+
+        $pubTypes = [];
+        $ptNodes = $xp->query('//PubmedArticle//PublicationTypeList//PublicationType');
+        if ($ptNodes) {
+            foreach ($ptNodes as $n) {
+                $t = trim($n->textContent);
+                if ($t !== '') $pubTypes[] = $t;
+            }
+        }
+        $pubTypes = array_values(array_unique($pubTypes));
+
+        $journal = $this->xpText($xp, '//PubmedArticle//Journal//Title');
+
+        $year = '';
+        $year = $this->xpText($xp, '//PubmedArticle//JournalIssue//PubDate//Year');
+        if ($year === '') {
+            $medlineDate = $this->xpText($xp, '//PubmedArticle//JournalIssue//PubDate//MedlineDate');
+            if (preg_match('/(19\\d{2}|20\\d{2})/', $medlineDate, $m)) {
+                $year = $m[1];
+            }
+        }
+
+        if ($title === '' && $abstract === '') {
+            return null;
+        }
+
+        return [
+            'title' => $title,
+            'abstract' => $abstract,
+            'mesh_terms' => $mesh,
+            'publication_types' => $pubTypes,
+            'journal' => $journal,
+            'year' => $year,
+        ];
+    }
+
+    private function xpText(\DOMXPath $xp, string $query): string
+    {
+        $n = $xp->query($query);
+        if ($n && $n->length > 0) {
+            return trim($n->item(0)->textContent);
+        }
+        return '';
+    }
+
+    private function httpGet(string $url): string
+    {
+        $ch = curl_init();
+        curl_setopt($ch, CURLOPT_URL, $url);
+        curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
+        curl_setopt($ch, CURLOPT_TIMEOUT, $this->timeout);
+        curl_setopt($ch, CURLOPT_SSL_VERIFYPEER, false);
+        curl_setopt($ch, CURLOPT_HTTPHEADER, [
+            'User-Agent: TMRjournals-PubMed/1.0'
+        ]);
+        $res = curl_exec($ch);
+        curl_close($ch);
+        return is_string($res) ? $res : '';
+    }
+
+    private function cacheDir(): string
+    {
+        return rtrim(ROOT_PATH, '/') . '/runtime/pubmed_cache';
+    }
+
+    private function cacheGet(string $key, int $ttlSeconds)
+    {
+        $file = $this->cacheDir() . '/' . $key . '.json';
+        if (!is_file($file)) return null;
+        $mtime = filemtime($file);
+        if (!$mtime || (time() - $mtime) > $ttlSeconds) return null;
+        $raw = @file_get_contents($file);
+        $decoded = json_decode((string)$raw, true);
+        return $decoded;
+    }
+
+    private function cacheSet(string $key, $value): void
+    {
+        $dir = $this->cacheDir();
+        if (!is_dir($dir)) @mkdir($dir, 0777, true);
+        $file = $dir . '/' . $key . '.json';
+        @file_put_contents($file, json_encode($value, JSON_UNESCAPED_UNICODE));
+    }
+}
+

application/common/QueueJob.php

@@ -152,6 +152,54 @@ class QueueJob
         $job->delete();
     }
 
+    /**
+     * 简化版异常处理（无 Redis 锁场景）。
+     *
+     * 适用于一对一类型的 job：每个 job 处理一个自然唯一的资源（如 log_id/expert_id），
+     * 不需要 Redis 锁也不会重复消费。
+     *
+     * 行为：
+     *   - 致命 DB 错误（连接断开等）→ release 任务回队列 + exit(1) 让 supervisor 重启进程
+     *   - 普通异常 → 按 attempts 限次重试（默认最多 3 次，超过 delete）
+     *
+     * @param \Exception|\Throwable $e
+     * @param \think\queue\Job $job
+     * @param string $context  日志中显示的上下文标识（如 "log_id=123" / "task_id=45"）
+     */
+    public function handleException($e, $job, $context = '')
+    {
+        $msg = $e->getMessage();
+        $this->log("队列异常 | {$context} | msg={$msg}");
+
+        if ($e instanceof \Exception && $this->isFatalDatabaseError($e)) {
+            $this->log("致命 DB 错误，释放任务并退出进程让 supervisor 重启 | {$context}");
+            $job->release(60);
+            exit(1);
+        }
+        // 兼容 \Throwable（PHP7 fatal 等）
+        if (!($e instanceof \Exception)) {
+            $low = strtolower((string)$msg);
+            $hits = ['mysql server has gone away', 'lost connection to mysql', 'gone away', 'sqlstate[hy000]'];
+            foreach ($hits as $kw) {
+                if (strpos($low, $kw) !== false) {
+                    $this->log("致命 DB 错误（Throwable），释放任务并退出进程 | {$context}");
+                    $job->release(60);
+                    exit(1);
+                }
+            }
+        }
+
+        $attempts = method_exists($job, 'attempts') ? $job->attempts() : 1;
+        if ($attempts >= $this->maxRetries) {
+            $this->log("超过最大重试次数({$this->maxRetries})，删除任务 | {$context}");
+            $job->delete();
+            return;
+        }
+        $delay = $this->getRetryDelay($msg);
+        $this->log("{$delay} 秒后重试({$attempts}/{$this->maxRetries}) | {$context}");
+        $job->release($delay);
+    }
+
     // 判断是否为需要重启的致命数据库错误
     private function isFatalDatabaseError(\Exception $e)
     {

application/common/QueueRedis.php

@@ -80,6 +80,25 @@ class QueueRedis
             return null;
         }
     }
+
+    /**
+     * 删除一个或多个 Redis 键（用于重检前清除队列任务 completed 标记）
+     */
+    public function deleteRedisKeys(array $keys)
+    {
+        $keys = array_values(array_filter($keys, function ($k) {
+            return $k !== null && $k !== '';
+        }));
+        if (empty($keys)) {
+            return true;
+        }
+        try {
+            $this->connect()->del(...$keys);
+            return true;
+        } catch (\Exception $e) {
+            return false;
+        }
+    }
     
     // 安全释放锁（仅当值匹配时删除）
     public function releaseRedisLock($key, $value)

application/common/Scopus.php

@@ -0,0 +1,134 @@
+<?php
+namespace app\api\controller;
+use app\api\controller\Base;
+
+
+class Scopus extends Base
+{
+
+    private $sApiKey = '434dcaf0e787b8676745c97687184f71';
+
+    // 获取h指数
+    public function getHindex($aParam = [])
+    {
+
+        $aParam = empty($aParam) ? $this->request->post() : $aParam;
+        // 获取输入参数
+        $sAuthorName = empty($aParam['author_name']) ? '' : $aParam['author_name'];
+        $affiliation = empty($aParam['affiliation']) ? '' : $aParam['affiliation'];
+
+        // 验证输入
+        if (empty($sAuthorName) || empty($affiliation)) {
+            return json_encode(['status' => 2,'msg' => '']);
+        }
+
+
+
+        $url = "https://api.elsevier.com/content/search/author?query=affil(beijing)&count=1";
+        
+        $ch = curl_init();
+        curl_setopt_array($ch, [
+            CURLOPT_URL => $url,
+            CURLOPT_HTTPHEADER => [
+                'X-ELS-APIKey: ' . $this->sApiKey,
+                'Accept: application/json'
+            ],
+            CURLOPT_RETURNTRANSFER => true,
+            CURLOPT_TIMEOUT => 15,
+            CURLOPT_FAILONERROR => true
+        ]);
+
+        $response = curl_exec($ch);
+var_dump(curl_getinfo($ch));exit;
+        // 第一步：检索作者ID
+        $url = "https://api.elsevier.com/content/search/author";
+        $params = [
+            'query' => "AUTHFIRST({$sAuthorName}) AND AFFIL({$affiliation})",
+            'count' => 1
+        ];
+var_dump($this->sApiKey);exit();
+        $ch = curl_init();
+        curl_setopt_array($ch, [
+            CURLOPT_URL => $url . '?' . http_build_query($params),
+            CURLOPT_HTTPHEADER => [
+                'X-ELS-APIKey: ' . $this->sApiKey,
+                'Accept: application/json'
+            ],
+            CURLOPT_RETURNTRANSFER => true,
+            // CURLOPT_TIMEOUT => 15,
+            CURLOPT_FAILONERROR => true
+        ]);
+
+        $response = curl_exec($ch);
+echo '<pre>';var_dump($response,curl_getinfo($ch));exit;
+        if (curl_errno($ch)) {
+            throw new \Exception('Scopus API错误: ' . curl_error($ch));
+        }
+
+        $httpCode = curl_getinfo($ch, CURLINFO_HTTP_CODE);
+        curl_close($ch);
+
+        if ($httpCode !== 200) {
+            throw new \Exception("Scopus API返回错误: HTTP {$httpCode}");
+        }
+
+        $data = json_decode($response, true);
+        if (empty($data['search-results']['entry'][0]['dc:identifier'])) {
+            throw new \Exception('未找到匹配的作者');
+        }
+
+        // 提取作者ID（从"AUTHOR_ID:7004212771"中分离）
+        $identifier = $data['search-results']['entry'][0]['dc:identifier'];
+        $authorId = explode(':', $identifier)[1];
+
+        // 第二步：获取作者指标
+        $url = "https://api.elsevier.com/content/author/author_id/{$authorId}?view=METRICS";
+        $ch = curl_init();
+        curl_setopt_array($ch, [
+            CURLOPT_URL => $url,
+            CURLOPT_HTTPHEADER => [
+                'X-ELS-APIKey: ' . $apiKey,
+                'Accept: application/json'
+            ],
+            CURLOPT_RETURNTRANSFER => true,
+            CURLOPT_TIMEOUT => 15,
+            CURLOPT_FAILONERROR => true
+        ]);
+
+        $response = curl_exec($ch);
+        if (curl_errno($ch)) {
+            throw new \Exception('指标API错误: ' . curl_error($ch));
+        }
+
+        $httpCode = curl_getinfo($ch, CURLINFO_HTTP_CODE);
+        curl_close($ch);
+
+        if ($httpCode !== 200) {
+            throw new \Exception("指标API返回错误: HTTP {$httpCode}");
+        }
+
+        $data = json_decode($response, true);
+        if (empty($data['author-retrieval-response'][0]['h-index'])) {
+            throw new \Exception('未找到h指数数据');
+        }
+
+        // 提取其他作者信息
+        $author = $data['author-retrieval-response'][0];
+        $metrics = [
+            'h_index' => $author['h-index'],
+            'citation_count' => $author['citation-count'] ?? '未知',
+            'document_count' => $author['document-count'] ?? '未知',
+            'coauthors_count' => $author['coauthor-count'] ?? '未知',
+            'affiliation_current' => $author['affiliation-current']['affiliation']['ip-doc']['afdispname'] ?? '未知'
+        ];
+
+        return json([
+            'status' => 'success',
+            'author_id' => $authorId,
+            'author_name' => $author['author-profile']['preferred-name']['ce:indexed-name'] ?? '未知',
+            'metrics' => $metrics
+        ]);
+
+        
+    }
+}

application/common/TurnitinService.php

@@ -0,0 +1,322 @@
+<?php
+
+namespace app\common;
+
+use think\Env;
+use think\Exception;
+
+/**
+ * Turnitin Core API (TCA) REST 客户端封装。
+ *
+ * 适用 Crossref Similarity Check 通道（product_name=Crossref）以及标准 TCA 接入。
+ *
+ * 鉴权：Authorization: Bearer <API_KEY>
+ *       X-Turnitin-Integration-Name / X-Turnitin-Integration-Version 用于审计
+ *
+ * .env 配置（[turnitin] 段）:
+ *   BASE_URL              形如 https://crossref-12345.turnitin.com/api/v1（不带尾斜杠）
+ *   API_KEY               生成的 Bearer token
+ *   INTEGRATION_NAME      Scope Name（创建 integration 时填的名字）
+ *   INTEGRATION_VERSION   自定义版本号，便于审计 e.g. 1.0.0
+ *
+ * API 文档：https://developers.turnitin.com/docs/tca
+ *
+ * 注意：
+ *   - 所有方法返回原始 decode 后的数组；HTTP 错误抛 Exception
+ *   - 不做任何业务层逻辑（业务层在 PlagiarismService 里）
+ *   - 不缓存 token（Bearer 不需要登录，每次请求自带）
+ */
+class TurnitinService
+{
+    private $baseUrl;
+    private $apiKey;
+    private $integrationName;
+    private $integrationVersion;
+    private $timeout = 60;
+
+    public function __construct()
+    {
+        $this->baseUrl            = rtrim(trim((string)Env::get('turnitin.base_url', '')), '/');
+        $this->apiKey             = trim((string)Env::get('turnitin.api_key', ''));
+        $this->integrationName    = trim((string)Env::get('turnitin.integration_name', 'tmr'));
+        $this->integrationVersion = trim((string)Env::get('turnitin.integration_version', '1.0.0'));
+
+        if ($this->baseUrl === '' || $this->apiKey === '') {
+            throw new Exception('Turnitin not configured: missing BASE_URL or API_KEY in .env [turnitin] section');
+        }
+    }
+
+    // ==================== Public API ====================
+
+    /**
+     * 探活 / 拿账户能力
+     * GET /features-enabled
+     */
+    public function featuresEnabled()
+    {
+        return $this->request('GET', '/features-enabled');
+    }
+
+    /**
+     * 创建 submission（拿到 id 之后才能上传文件）
+     * POST /submissions
+     *
+     * @param array $meta 必填字段:
+     *   - title           论文标题
+     *   - owner           submission owner 标识符（自定义字符串，比如投稿系统 user_id）
+     *   - submitter       提交者标识符（同上）
+     *   - eula            (可选) ['version' => '...', 'language' => 'en-US', 'accepted_timestamp' => ISO8601]
+     *                     如果 features-enabled 返回 require_eula=false 可省略
+     *   可选字段:
+     *   - extract_text_only  bool
+     *   - metadata           array  自定义键值，供后续追溯
+     *
+     * @return array 含 id（submission UUID）, status, owner, ...
+     */
+    public function createSubmission($meta)
+    {
+        return $this->request('POST', '/submissions', $meta);
+    }
+
+    /**
+     * 上传文件到 submission
+     * PUT /submissions/{id}/original/{filename}
+     *
+     * @param string $submissionId
+     * @param string $filePath  本地 PDF/DOCX 路径
+     * @param string $filename  传给 Turnitin 的文件名（用于报告显示）
+     * @return array
+     */
+    public function uploadFile($submissionId, $filePath, $filename = '')
+    {
+        if (!is_file($filePath) || !is_readable($filePath)) {
+            throw new Exception("File not found or not readable: {$filePath}");
+        }
+        if ($filename === '') {
+            $filename = basename($filePath);
+        }
+        $body = file_get_contents($filePath);
+
+        return $this->request(
+            'PUT',
+            '/submissions/' . urlencode($submissionId) . '/original/' . rawurlencode($filename),
+            $body,
+            [
+                'Content-Type'              => 'binary/octet-stream',
+                'Content-Disposition'       => 'inline; filename="' . $filename . '"',
+            ]
+        );
+    }
+
+    /**
+     * 触发 similarity 比对
+     * PUT /submissions/{id}/similarity
+     *
+     * @param string $submissionId
+     * @param array  $opts
+     *   - generation_settings.search_repositories  默认 ['INTERNET','PUBLICATION','CROSSREF','CROSSREF_POSTED_CONTENT','SUBMITTED_WORK']
+     *   - generation_settings.submission_auto_excludes  bool
+     *   - view_settings.exclude_quotes / exclude_bibliography / exclude_citations / exclude_abstract / exclude_methods  bool
+     *   - indexing_settings.add_to_index  bool 是否把本文加进 SUBMITTED_WORK 索引（一般 true）
+     * @return array
+     */
+    public function triggerSimilarity($submissionId, $opts = [])
+    {
+        $body = array_merge([
+            'generation_settings' => [
+                'search_repositories' => ['INTERNET', 'PUBLICATION', 'CROSSREF', 'CROSSREF_POSTED_CONTENT', 'SUBMITTED_WORK'],
+                'submission_auto_excludes' => true,
+                'auto_exclude_self_matching_scope' => 'GROUP_CONTEXT',
+            ],
+            'view_settings' => [
+                'exclude_quotes'        => true,
+                'exclude_bibliography'  => true,
+                'exclude_citations'     => true,
+            ],
+            'indexing_settings' => [
+                'add_to_index' => true,
+            ],
+        ], $opts);
+
+        return $this->request(
+            'PUT',
+            '/submissions/' . urlencode($submissionId) . '/similarity',
+            $body
+        );
+    }
+
+    /**
+     * 查询 similarity 状态
+     * GET /submissions/{id}/similarity
+     *
+     * 返回 status: PROCESSING / COMPLETE / ERROR
+     * COMPLETE 时返回 overall_match_percentage / time_requested / time_generated
+     */
+    public function getSimilarityStatus($submissionId)
+    {
+        return $this->request(
+            'GET',
+            '/submissions/' . urlencode($submissionId) . '/similarity'
+        );
+    }
+
+    /**
+     * 取在线查看报告的临时 URL
+     * POST /submissions/{id}/viewer-url
+     *
+     * 返回 viewer_url（数小时有效）
+     *
+     * @param array $viewer 可选 viewer 设置 e.g. ['viewer_default_permission_set' => 'INSTRUCTOR']
+     */
+    public function getViewerUrl($submissionId, $viewer = [])
+    {
+        $body = array_merge([
+            'viewer_default_permission_set' => 'INSTRUCTOR',
+            'similarity' => [
+                'default_mode' => 'MATCH_OVERVIEW',
+                'view_settings' => ['save_changes' => true],
+                'modes' => ['match_overview' => true, 'all_sources' => true],
+            ],
+            'locale' => 'en-US',
+        ], $viewer);
+
+        return $this->request(
+            'POST',
+            '/submissions/' . urlencode($submissionId) . '/viewer-url',
+            $body
+        );
+    }
+
+    /**
+     * 触发生成 PDF 报告（异步，状态在另一个轮询里看）
+     * POST /submissions/{id}/similarity/pdf
+     *
+     * 返回 id（pdf 报告 ID）
+     */
+    public function requestPdfReport($submissionId, $opts = [])
+    {
+        $body = array_merge([
+            'locale' => 'en-US',
+        ], $opts);
+
+        return $this->request(
+            'POST',
+            '/submissions/' . urlencode($submissionId) . '/similarity/pdf',
+            $body
+        );
+    }
+
+    /**
+     * 查询 PDF 报告状态
+     * GET /submissions/{id}/similarity/pdf/{pdf_id}/status
+     *
+     * status: PENDING / SUCCESS / FAILED
+     */
+    public function getPdfReportStatus($submissionId, $pdfId)
+    {
+        return $this->request(
+            'GET',
+            '/submissions/' . urlencode($submissionId) . '/similarity/pdf/' . urlencode($pdfId) . '/status'
+        );
+    }
+
+    /**
+     * 下载 PDF 报告内容（status=SUCCESS 后才可调用）
+     * GET /submissions/{id}/similarity/pdf/{pdf_id}
+     *
+     * 返回 raw PDF binary 字符串；调用方负责落盘
+     */
+    public function downloadPdfReport($submissionId, $pdfId)
+    {
+        return $this->request(
+            'GET',
+            '/submissions/' . urlencode($submissionId) . '/similarity/pdf/' . urlencode($pdfId),
+            null,
+            [],
+            true   // raw response (不 json_decode)
+        );
+    }
+
+    // ==================== Internal HTTP layer ====================
+
+    /**
+     * 统一 HTTP 调用
+     *
+     * @param string      $method      GET/POST/PUT/DELETE
+     * @param string      $path        以 / 开头的相对路径，会拼到 baseUrl 后
+     * @param mixed       $body        array 时按 JSON 编码；string 时直接当 raw body
+     * @param array       $extraHeaders 额外 header
+     * @param bool        $rawResponse  true=返回 raw 字符串；false=json_decode
+     * @return mixed
+     * @throws Exception
+     */
+    private function request($method, $path, $body = null, $extraHeaders = [], $rawResponse = false)
+    {
+        $url = $this->baseUrl . $path;
+
+        $headers = [
+            'Authorization: Bearer ' . $this->apiKey,
+            'X-Turnitin-Integration-Name: '    . $this->integrationName,
+            'X-Turnitin-Integration-Version: ' . $this->integrationVersion,
+        ];
+
+        $payload = null;
+        if ($body !== null) {
+            if (is_array($body)) {
+                $payload = json_encode($body, JSON_UNESCAPED_UNICODE | JSON_UNESCAPED_SLASHES);
+                $headers[] = 'Content-Type: application/json';
+            } else {
+                $payload = $body;
+                if (!isset($extraHeaders['Content-Type'])) {
+                    $headers[] = 'Content-Type: application/octet-stream';
+                }
+            }
+        }
+        foreach ($extraHeaders as $k => $v) {
+            $headers[] = $k . ': ' . $v;
+        }
+
+        $ch = curl_init();
+        curl_setopt_array($ch, [
+            CURLOPT_URL            => $url,
+            CURLOPT_CUSTOMREQUEST  => strtoupper($method),
+            CURLOPT_RETURNTRANSFER => true,
+            CURLOPT_HTTPHEADER     => $headers,
+            CURLOPT_TIMEOUT        => $this->timeout,
+            CURLOPT_CONNECTTIMEOUT => 15,
+            CURLOPT_SSL_VERIFYPEER => true,
+            CURLOPT_SSL_VERIFYHOST => 2,
+        ]);
+        if ($payload !== null) {
+            curl_setopt($ch, CURLOPT_POSTFIELDS, $payload);
+        }
+
+        $resp     = curl_exec($ch);
+        $httpCode = curl_getinfo($ch, CURLINFO_HTTP_CODE);
+        $err      = curl_error($ch);
+        curl_close($ch);
+
+        if ($resp === false) {
+            throw new Exception("Turnitin curl error: {$err} (url={$url})");
+        }
+        if ($httpCode < 200 || $httpCode >= 300) {
+            // 把响应体的前 1k 也带上方便排错
+            $excerpt = mb_substr((string)$resp, 0, 1000);
+            throw new Exception("Turnitin HTTP {$httpCode} {$method} {$path}: {$excerpt}");
+        }
+
+        if ($rawResponse) {
+            return $resp;
+        }
+        // 部分响应可能是 204 No Content
+        if ($resp === '' || $resp === null) {
+            return [];
+        }
+        $data = json_decode($resp, true);
+        if (json_last_error() !== JSON_ERROR_NONE) {
+            // 不是 JSON 也直接抛回原文
+            return $resp;
+        }
+        return $data;
+    }
+}

application/common/UnsubscribeService.php

@@ -0,0 +1,105 @@
+<?php
+
+namespace app\common;
+
+use think\Env;
+
+/**
+ * 退订工具：生成/校验签名 token，构造退订 URL（支持 expert / user 两类受众）。
+ *
+ * .env 配置（[unsubscribe] 段）：
+ *   UNSUBSCRIBE_SECRET    用于签名的密钥（必须自行设置一个长随机字符串）
+ *   UNSUBSCRIBE_BASE_URL  退订入口 URL，指向 Unsubscribe 控制器的 index 方法，例如：
+ *                         https://api.tmrjournals.com/api/Unsubscribe/index
+ *
+ * URL 格式：
+ *   {UNSUBSCRIBE_BASE_URL}?k={kind}&id={audience_id}&t={sha256(kind|id|email_lower|secret)}
+ *
+ *   kind:
+ *     expert  → 外部专家库 t_expert.expert_id
+ *     user    → 系统内部用户 t_user.user_id（编委 / 主编 / 青年编委等）
+ *
+ * 设计要点：
+ *   - 不依赖 session：纯 HMAC 风格签名，token 永久有效（直到密钥更换或邮箱变更）。
+ *   - 收到请求后服务端按 kind+id 查邮箱，重新计算签名比对，hash_equals 防时序攻击。
+ *   - 邮箱大小写敏感性：统一 strtolower 后再签名/校验。
+ */
+class UnsubscribeService
+{
+    const KIND_EXPERT = 'expert';
+    const KIND_USER   = 'user';
+
+    /**
+     * 读取签名密钥；未配置时使用一个固定默认值（仅供本地测试，生产必须覆盖）。
+     */
+    public static function getSecret()
+    {
+        $secret = trim((string)Env::get('unsubscribe.unsubscribe_secret', ''));
+        if ($secret === '') {
+            $secret = 'tmrjournals-default-unsubscribe-secret-change-me';
+        }
+        return $secret;
+    }
+
+    /**
+     * 读取退订入口 URL（指向 Unsubscribe/index）；未配置返回空串。
+     */
+    public static function getBaseUrl()
+    {
+        return rtrim(trim((string)Env::get('unsubscribe.unsubscribe_base_url', '')), '/');
+    }
+
+    /**
+     * 规范化 kind；未知值降级为 expert（保持向后兼容）。
+     */
+    public static function normalizeKind($kind)
+    {
+        $kind = strtolower(trim((string)$kind));
+        return $kind === self::KIND_USER ? self::KIND_USER : self::KIND_EXPERT;
+    }
+
+    /**
+     * 生成签名 token。
+     */
+    public static function buildToken($kind, $audienceId, $email)
+    {
+        $kind       = self::normalizeKind($kind);
+        $audienceId = intval($audienceId);
+        $email      = strtolower(trim((string)$email));
+        return hash('sha256', $kind . '|' . $audienceId . '|' . $email . '|' . self::getSecret());
+    }
+
+    /**
+     * 校验签名 token；恒定时间比较，防时序攻击。
+     */
+    public static function verifyToken($kind, $audienceId, $email, $token)
+    {
+        $kind       = self::normalizeKind($kind);
+        $audienceId = intval($audienceId);
+        $email      = strtolower(trim((string)$email));
+        $token      = strtolower(trim((string)$token));
+        if ($audienceId <= 0 || $email === '' || $token === '') {
+            return false;
+        }
+        $expected = self::buildToken($kind, $audienceId, $email);
+        return hash_equals($expected, $token);
+    }
+
+    /**
+     * 构造完整退订 URL（用于邮件模板变量 {{unsubscribe_url}}）。
+     * BASE_URL 未配置或参数无效时返回空串。
+     */
+    public static function buildUrl($kind, $audienceId, $email)
+    {
+        $kind       = self::normalizeKind($kind);
+        $audienceId = intval($audienceId);
+        $email      = trim((string)$email);
+        if ($audienceId <= 0 || $email === '') return '';
+        $base = self::getBaseUrl();
+        if ($base === '') return '';
+        return $base
+            . '?k=' . $kind
+            . '&id=' . $audienceId
+            . '&t='  . self::buildToken($kind, $audienceId, $email);
+    }
+}

application/config.php

@@ -172,8 +172,11 @@ return [
         // 日志保存目录
         'path'  => LOG_PATH,
         // 日志记录级别
-        'level' => ['log'],
-        "max_files"=>3,
+        'level' => ['error', 'warning', 'info', 'notice', 'debug'],
+        'file_size'   => 1024 * 1024 * 2, // 2MB
+        'max_files'=>30,
+        'apart_day'   => true,
+        'format'      => '[%s][%s] %s',
     ],
 
     // +----------------------------------------------------------------------

application/database.php

@@ -57,4 +57,6 @@ return [
     'datetime_format' => 'Y-m-d H:i:s',
     // 是否需要进行SQL性能分析
     'sql_explain'     => false,
+    // 断线重连（长驻 worker 必开）：MySQL wait_timeout 关闭连接后，下一次查询自动重连
+    'break_reconnect' => true,
 ];

sql/add_country_scope_to_promotion_task.sql

@@ -0,0 +1,3 @@
+ALTER TABLE t_promotion_task
+  ADD COLUMN target_partitions VARCHAR(32) NOT NULL DEFAULT '' COMMENT '目标分区，逗号分隔，如 1,2,3；空=不限',
+  ADD COLUMN target_country_ids VARCHAR(255) NOT NULL DEFAULT '' COMMENT '单独指定的国家ID，逗号分隔；空=不额外指定';

sql/add_llm_advised_topics_to_promotion_email_log.sql

@@ -0,0 +1,7 @@
+-- 为 t_promotion_email_log 增加第二段 AI 生成字段（推荐投稿方向）
+-- llm_advised_topics:         大模型生成的"我们特别关注 X/Y/Z 领域"段落，对应模板变量 {{ai_advised_topics}}
+-- llm_advised_topics_status:  0未处理 1成功 2失败兜底 3跳过(无可用领域)
+
+ALTER TABLE `t_promotion_email_log`
+  ADD COLUMN `llm_advised_topics` TEXT NULL COMMENT '大模型生成的推荐投稿方向文案' AFTER `llm_description`,
+  ADD COLUMN `llm_advised_topics_status` TINYINT NOT NULL DEFAULT 0 COMMENT 'AI 推荐方向状态: 0未处理 1成功 2失败兜底 3跳过' AFTER `llm_advised_topics`;

sql/add_llm_fields_to_promotion_email_log.sql

@@ -0,0 +1,7 @@
+-- 为 t_promotion_email_log 增加 LLM 生成字段
+-- llm_description: 大模型生成的个性化描述（用于模板变量 {{llm_description}}）
+-- llm_status:      0=未处理 1=成功 2=失败已回退兜底 3=跳过(缺代表作)
+
+ALTER TABLE `t_promotion_email_log`
+  ADD COLUMN `llm_description` TEXT NULL COMMENT '大模型生成的个性化描述' AFTER `body_prepared`,
+  ADD COLUMN `llm_status` TINYINT NOT NULL DEFAULT 0 COMMENT 'LLM 处理状态: 0未处理 1成功 2失败兜底 3跳过' AFTER `llm_description`;

sql/add_plagiarism_check_table.sql

@@ -0,0 +1,44 @@
+-- 查重任务表（Turnitin TCA / Crossref Similarity Check）
+--
+-- 一篇 article 可重复触发多次查重；同一 article 的最近一次显示在编辑详情页。
+-- state 流转：0 待上传 → 1 上传中 → 2 比对中 → 3 完成 → 4 失败
+-- 失败可重新触发，会创建新行（保留历史）
+--
+-- 报告永久保留：pdf_local_path 指向 runtime/plagiarism/ 下的本地 PDF；
+-- view_only_url 是 Turnitin 临时签名（数小时过期），过期需重新生成
+
+DROP TABLE IF EXISTS `t_plagiarism_check`;
+CREATE TABLE `t_plagiarism_check` (
+  `check_id`              INT NOT NULL AUTO_INCREMENT,
+  `article_id`            INT NOT NULL DEFAULT 0           COMMENT '关联投稿 t_article.article_id',
+  `journal_id`            INT NOT NULL DEFAULT 0           COMMENT '所属期刊（冗余便于按期刊统计）',
+  `triggered_by`          INT NOT NULL DEFAULT 0           COMMENT '触发人 user_id（手工触发时编辑的 user_id）',
+  `trigger_source`        VARCHAR(32) NOT NULL DEFAULT 'manual' COMMENT 'manual/auto_initial_review/...',
+  `state`                 TINYINT NOT NULL DEFAULT 0       COMMENT '0待上传 1上传中 2比对中 3完成 4失败',
+
+  -- Turnitin 端的实体 ID
+  `tii_submission_id`     VARCHAR(64) NOT NULL DEFAULT ''  COMMENT 'Turnitin submission UUID',
+  `tii_report_status`     VARCHAR(32) NOT NULL DEFAULT ''  COMMENT 'PROCESSING/COMPLETE/ERROR',
+
+  -- 结果
+  `similarity_score`      DECIMAL(5,2) NOT NULL DEFAULT 0  COMMENT '总相似度 %（如 12.34）',
+  `view_only_url`         VARCHAR(1024) NOT NULL DEFAULT '' COMMENT '在线查看报告 URL（临时签名）',
+  `view_only_url_expire`  INT NOT NULL DEFAULT 0           COMMENT '在线查看 URL 过期时间戳',
+  `pdf_local_path`        VARCHAR(255) NOT NULL DEFAULT '' COMMENT '本地缓存的 PDF 报告相对路径',
+
+  -- 文件元数据（上传时记录，便于追踪）
+  `source_file_name`      VARCHAR(255) NOT NULL DEFAULT '' COMMENT '原始 PDF 文件名',
+  `source_file_size`      INT NOT NULL DEFAULT 0           COMMENT '原始 PDF 字节数',
+
+  -- 调试与重试
+  `attempts`              INT NOT NULL DEFAULT 0           COMMENT '总轮询/重试次数',
+  `error_msg`             VARCHAR(1024) NOT NULL DEFAULT '' COMMENT '失败原因',
+  `raw_response`          MEDIUMTEXT                       COMMENT '最近一次 Turnitin API 原始返回（调试用）',
+
+  `ctime`                 INT NOT NULL DEFAULT 0,
+  `utime`                 INT NOT NULL DEFAULT 0,
+  PRIMARY KEY (`check_id`),
+  KEY `idx_article` (`article_id`, `state`),
+  KEY `idx_state` (`state`),
+  KEY `idx_tii_submission` (`tii_submission_id`)
+) ENGINE=InnoDB DEFAULT CHARSET=utf8mb3 COMMENT='Turnitin 查重任务表';

sql/add_unsubscribed_to_expert.sql

@@ -0,0 +1,9 @@
+-- 退订功能：t_expert 增加退订标记
+-- 用户已手动新增 unsubscribed 字段，以下为完整建议建表语句（含索引）
+-- 0=正常 1=已退订；选人时统一过滤 unsubscribed=0
+
+ALTER TABLE `t_expert`
+  ADD COLUMN `unsubscribed` TINYINT NOT NULL DEFAULT 0 COMMENT '0正常 1已退订' AFTER `state`;
+
+ALTER TABLE `t_expert`
+  ADD INDEX `idx_unsubscribed` (`unsubscribed`);

sql/add_user_id_to_promotion_email_log.sql

@@ -0,0 +1,11 @@
+-- 推广邮件日志：增加内部受众 user_id 列
+-- expert_type=5（外部 expert 库）：expert_id 有值，user_id=0
+-- expert_type∈{1,2,3,4}（内部主编/编委/青年编委/作者）：user_id 有值，expert_id=0
+
+ALTER TABLE `t_promotion_email_log`
+  ADD COLUMN `user_id` INT NOT NULL DEFAULT 0
+    COMMENT '内部受众 user_id；expert_type∈{1..4} 时填，=5 时为 0'
+    AFTER `expert_id`;
+
+ALTER TABLE `t_promotion_email_log`
+  ADD INDEX `idx_user_send_time` (`user_id`, `send_time`);

test_plagiarism_e2e.php

@@ -0,0 +1,102 @@
+<?php
+/**
+ * Turnitin TCA 端到端连通性测试。
+ *
+ * 用法（在项目根执行）:
+ *   php test_plagiarism_e2e.php features            # 探活
+ *   php test_plagiarism_e2e.php submit <article_id> # 用 article 主稿提交查重（手工触发）
+ *   php test_plagiarism_e2e.php submit-file <pdf>   # 用本地 PDF 提交（不绑定 article）
+ *   php test_plagiarism_e2e.php status <check_id>   # 查询状态
+ *   php test_plagiarism_e2e.php list <article_id>   # 列出某 article 的查重记录
+ *   php test_plagiarism_e2e.php viewer <check_id>   # 取在线查看 URL
+ *
+ * 说明：
+ *   submit-file 不会真正落库（仅用于联通验证），它会用 article_id=0 走完整套流程。
+ *   submit       会写入 t_plagiarism_check，并把 check_id 打回，再用 status 自己轮询。
+ */
+
+define('IS_CLI', true);
+
+require __DIR__ . '/thinkphp/start.php';
+
+use think\Db;
+use app\common\PlagiarismService;
+use app\common\TurnitinService;
+
+if ($argc < 2) {
+    echo "Usage: php test_plagiarism_e2e.php <command> [args...]\n";
+    exit(1);
+}
+$cmd = $argv[1];
+
+try {
+    switch ($cmd) {
+        case 'features': {
+            $tii = new TurnitinService();
+            print_r($tii->featuresEnabled());
+            break;
+        }
+        case 'submit': {
+            if ($argc < 3) {
+                echo "Usage: ... submit <article_id>\n";
+                exit(1);
+            }
+            $articleId = intval($argv[2]);
+            $svc = new PlagiarismService();
+            $local = $svc->locateArticleManuscript($articleId);
+            echo "manuscript local path: {$local}\n";
+            $checkId = $svc->submit($articleId, $local, 0, 'cli_test');
+            echo "submitted, check_id = {$checkId}\n";
+            echo "now run: php think queue:work --queue plagiarism --tries=1 -v\n";
+            break;
+        }
+        case 'submit-file': {
+            if ($argc < 3) {
+                echo "Usage: ... submit-file <pdf_path>\n";
+                exit(1);
+            }
+            $path = $argv[2];
+            if (!is_file($path)) {
+                echo "file not exists: {$path}\n";
+                exit(1);
+            }
+            $svc = new PlagiarismService();
+            $checkId = $svc->submit(0, $path, 0, 'cli_test_file');
+            echo "submitted, check_id = {$checkId}\n";
+            break;
+        }
+        case 'status': {
+            if ($argc < 3) {
+                echo "Usage: ... status <check_id>\n";
+                exit(1);
+            }
+            $row = Db::name('plagiarism_check')->where('check_id', intval($argv[2]))->find();
+            print_r($row);
+            break;
+        }
+        case 'list': {
+            if ($argc < 3) {
+                echo "Usage: ... list <article_id>\n";
+                exit(1);
+            }
+            $rows = Db::name('plagiarism_check')->where('article_id', intval($argv[2]))->order('check_id desc')->select();
+            print_r($rows);
+            break;
+        }
+        case 'viewer': {
+            if ($argc < 3) {
+                echo "Usage: ... viewer <check_id>\n";
+                exit(1);
+            }
+            $svc = new PlagiarismService();
+            print_r($svc->refreshViewerUrlFor(intval($argv[2])));
+            break;
+        }
+        default:
+            echo "unknown command: {$cmd}\n";
+            exit(1);
+    }
+} catch (\Throwable $e) {
+    echo "ERROR: " . $e->getMessage() . "\n" . $e->getTraceAsString() . "\n";
+    exit(1);
+}