init: Search Hub - 统一多搜索引擎聚合服务

providers/ai_provider.py

@@ -0,0 +1,137 @@
+"""AI 总结服务 — 基于 OpenCodeZen / OpenAI 兼容 API"""
+
+import json
+import time
+import requests
+from providers.base import BaseProvider
+
+
+def build_prompts(query, results):
+    search_text = ''
+    for i, r in enumerate(results, 1):
+        search_text += f"[{i}] 标题: {r.get('title', '')}\n"
+        search_text += f"    来源: {r.get('url', '')}\n"
+        search_text += f"    内容: {r.get('content', '')}\n\n"
+
+    system_prompt = (
+        '你是一个专业的搜索结果分析助手。'
+        '请根据用户搜索词和搜索结果，生成一份结构化的中文总结报告。'
+        '要求：\n'
+        '1. 先概括本次搜索的核心主题\n'
+        '2. 列出关键发现和重要信息点（分点说明）\n'
+        '3. 指出不同信息来源之间的共识与分歧（如有）\n'
+        '4. 给出综合结论\n'
+        '使用简洁的 Markdown 格式，不要过于冗长。'
+    )
+    user_prompt = f'## 搜索词\n{query}\n\n## 搜索结果\n{search_text}\n\n请根据以上搜索结果生成总结报告。'
+    return system_prompt, user_prompt
+
+
+class AIProvider(BaseProvider):
+    name = 'ai'
+    display_name = 'AI 总结'
+    needs_api_key = True
+    enabled = False
+    priority = 50
+
+    def __init__(self, config: dict):
+        super().__init__(config)
+        oc = config.get('opencodezen', {})
+        self.api_key = oc.get('api_key')
+        self.base_url = oc.get('base_url', 'https://opencode.ai/zen/go/v1').rstrip('/')
+        self.model = oc.get('model', 'deepseek-v4-flash')
+
+    def is_available(self) -> bool:
+        return bool(self.api_key)
+
+    def search(self, query: str, max_results: int = 5) -> list:
+        return []
+
+    def summarize(self, query: str, results: list) -> dict:
+        if not self.is_available():
+            return {'error': 'AI 总结未配置', 'summary': ''}
+
+        system_prompt, user_prompt = build_prompts(query, results)
+        url = f'{self.base_url}/chat/completions'
+        headers = {
+            'Authorization': f'Bearer {self.api_key}',
+            'Content-Type': 'application/json',
+        }
+        payload = {
+            'model': self.model,
+            'messages': [
+                {'role': 'system', 'content': system_prompt},
+                {'role': 'user', 'content': user_prompt},
+            ],
+            'max_tokens': 4096,
+            'temperature': 0.3,
+        }
+
+        start = time.time()
+        try:
+            resp = requests.post(url, json=payload, headers=headers, timeout=60)
+            elapsed = round(time.time() - start, 2)
+            if resp.status_code != 200:
+                return {'error': f'AI API 返回 {resp.status_code}', 'summary': '', 'elapsed': elapsed}
+            data = resp.json()
+            return {
+                'summary': data['choices'][0]['message']['content'],
+                'model': self.model,
+                'elapsed': elapsed,
+                'usage': data.get('usage', {}),
+            }
+        except Exception as e:
+            return {'error': str(e), 'summary': '', 'elapsed': round(time.time() - start, 2)}
+
+    def summarize_stream(self, query: str, results: list):
+        if not self.is_available():
+            yield f"event: error\ndata: {json.dumps({'error': 'AI 总结未配置'})}\n\n"
+            return
+
+        system_prompt, user_prompt = build_prompts(query, results)
+        url = f'{self.base_url}/chat/completions'
+        headers = {
+            'Authorization': f'Bearer {self.api_key}',
+            'Content-Type': 'application/json',
+        }
+        payload = {
+            'model': self.model,
+            'messages': [
+                {'role': 'system', 'content': system_prompt},
+                {'role': 'user', 'content': user_prompt},
+            ],
+            'max_tokens': 4096,
+            'temperature': 0.3,
+            'stream': True,
+        }
+
+        start = time.time()
+        try:
+            resp = requests.post(url, json=payload, headers=headers, stream=True, timeout=120)
+            if resp.status_code != 200:
+                yield f"event: error\ndata: {json.dumps({'error': f'AI API 返回 {resp.status_code}'})}\n\n"
+                return
+
+            for line in resp.iter_lines():
+                if not line:
+                    continue
+                line_str = line.decode('utf-8', errors='replace')
+                if not line_str.startswith('data: '):
+                    continue
+                data_str = line_str[6:]
+                if data_str.strip() == '[DONE]':
+                    break
+                try:
+                    chunk = json.loads(data_str)
+                    delta = chunk.get('choices', [{}])[0].get('delta', {})
+                    content = delta.get('content', '')
+                    if content:
+                        yield f"event: delta\ndata: {json.dumps({'content': content})}\n\n"
+                except json.JSONDecodeError:
+                    continue
+
+            elapsed = round(time.time() - start, 2)
+            yield f"event: meta\ndata: {json.dumps({'model': self.model, 'elapsed': elapsed})}\n\n"
+
+        except Exception as e:
+            yield f"event: error\ndata: {json.dumps({'error': str(e)})}\n\n"