[全量] 初始化项目代码、配置、文档及Agent协同harness

2026-04-02 11:36:05 +08:00
parent 0553309cdf
commit 87e571d9ec
1133 changed files with 221948 additions and 0 deletions
--- a/langchain-chat/server/chat/relevant_articles.py
+++ b/langchain-chat/server/chat/relevant_articles.py
@@ -0,0 +1,108 @@
+import re
+from fastapi import Body, HTTPException
+from fastapi.concurrency import run_in_threadpool
+from typing import List
+import logging
+from configs import (
+    SELF_TOP_K,
+    SCORE_THRESHOLD,
+)
+from configs.kb_config import DEFAULT_JOURNAL_BASE
+from server.knowledge_base.kb_doc_api import search_docs
+from server.utils import BaseResponse
+from pydantic import BaseModel, Field  # 用于请求模型验证
+
+# 配置日志记录
+logger = logging.getLogger(__name__)
+
+# 请求模型验证（可选增强）
+class ArticleRequest(BaseModel):
+    file_name: str = Field(..., min_length=1, description="文件标题", example="智能制造技术的发展与应用")
+    key_words: str = Field(..., description="逗号分隔的关键词", example="智能制造，工业互联网，物联网")
+
+async def relevant_articles(
+    request: ArticleRequest = Body(...)  # 使用Pydantic模型进行输入验证
+):
+    """
+        相关文献搜索：\n
+        - 输入：文件标题、关键词\n
+        - 输出：相关文献列表\n
+        返回值数据结构：\n
+            {\n
+                "code": 状态码（200表示成功，404表示未找到相关文档，500表示服务出现错误）,\n
+                "msg": 消息描述,\n
+                "data": {\n
+                    "total": 相关文献的总数,\n
+                    "articles": [\n
+                        {\n
+                            "doc_id": 文件ID\n
+                            "title": 标题\n
+                            "author_info": 作者\n
+                            "publish_date": 发表时间\n
+                            "data_source": 链接\n
+                        },\n
+                    ]\n
+                }\n
+            }\n
+    """
+    try:
+        # 构造规范化查询（标题重复两次加强权重，关键词空格分隔）
+        keywords_list = [kw.strip() for kw in request.key_words.split(',') if kw.strip()]
+        processed_keywords = ' '.join(set(keywords_list)).strip("暂无关键词")  # 去重处理
+        
+        # 构造带权重的查询字符串
+        query = f"{request.file_name}{processed_keywords}"
+        logger.info(f"相关文献检索query：{query}")
+        # 合并所有知识库的搜索结果
+        all_docs = []
+        for knowledge_name in [DEFAULT_JOURNAL_BASE]:
+            docs = await run_in_threadpool(
+                search_docs,
+                query=query,
+                fileName=[],
+                knowledge_base_name=knowledge_name,
+                top_k=SELF_TOP_K,
+                score_threshold=SCORE_THRESHOLD
+            )
+            for doc in docs:
+                if doc.metadata.get("publish_date", None) == 17010101:
+                    doc.metadata.pop("publish_date", None)
+            all_docs.extend(docs)
+        docs = all_docs[:5]
+        
+        # ========== 结果后处理 ==========
+        if not docs:
+            logger.info(f"未找到相关文档：{query}")
+            return BaseResponse(code=404, msg="未找到相关文档")
+        
+        # 提取链接
+        # pattern = r'http[s]?://(?:[a-zA-Z]|[0-9]|[$-_@.&+]|[!*\(\),]|(?:%[0-9a-fA-F][0-9a-fA-F]))+'
+        formatted_docs = []
+        for doc in docs:
+            # data_source = doc.metadata.get("data_source", "")
+            # links = re.findall(pattern, data_source)
+            formatted_doc = {
+                "doc_id": doc.metadata.get("ID", ""),
+                "title": doc.metadata.get("title", ""),
+                "author_info": doc.metadata.get("author_info", ""),
+                "publish_date": doc.metadata.get("publish_date", ""),
+                "data_source": doc.metadata.get("data_source", ""),
+            }
+            formatted_docs.append(formatted_doc)
+        
+        logger.info(f"成功返回 {len(docs)} 篇相关文档")
+        return BaseResponse(
+            code=200,
+            msg="成功检索到相关文档",
+            data={
+                "total": len(formatted_docs),
+                "articles": formatted_docs,
+            }
+        )
+
+    except Exception as e:
+        logger.error(f"文档检索失败: {str(e)}", exc_info=True)
+        return BaseResponse(
+            code=500,
+            msg=f"文档检索服务暂时不可用，错误信息：{str(e)}"
+        )