修改自动推广的相关任务

This commit is contained in:
wangjinlei
2026-05-07 11:45:55 +08:00
parent a68742d2c2
commit b1e978ed73
7 changed files with 1221 additions and 0 deletions

View File

@@ -0,0 +1,423 @@
<?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 = 'runtime/plagiarism';
/**
* 轮询间隔。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;
// ---------- 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,
]);
// 入队执行:上传 + 触发 similarity
Queue::push(
'app\\api\\job\\PlagiarismRun',
['check_id' => $checkId, 'file_path' => $filePath],
'plagiarism'
);
return (int)$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);
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'];
}
$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),
]);
// 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),
]);
// 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],
'plagiarism'
);
} 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)
{
$this->updateCheck($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', $articleId)
->where('type_name', 'manuscirpt') // 历史拼写
->order('article_file_id desc')
->find();
if (!$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();
}
}

View File

@@ -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 tokenBearer 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 $meta 必填字段:
* - title 论文标题
* - owner submission owner 标识符(自定义字符串,比如投稿系统 user_id
* - submitter 提交者标识符(同上)
* - eula (可选) ['version' => '...', 'language' => 'en-US', 'accepted_timestamp' => ISO8601]
* 如果 features-enabled 返回 require_eula=false 可省略
* 可选字段:
* - extract_text_only bool
* - metadata array 自定义键值,供后续追溯
*
* @return array 含 idsubmission UUID, status, owner, ...
*/
public function createSubmission($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',
$body
);
}
/**
* 触发生成 PDF 报告(异步,状态在另一个轮询里看)
* POST /submissions/{id}/similarity/pdf
*
* 返回 idpdf 报告 ID
*/
public function requestPdfReport($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;
}
}