1

2026-04-16 13:30:31 +08:00
parent a2338340f6
commit ae221e6be6
8 changed files with 838 additions and 91 deletions
--- a/application/common/CountryResolverService.php
+++ b/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];
+    }
+}
+