feat: 添加算法目录和知识库 API 文档

- 新增 algorithm/ 目录
- 添加知识库 API 需求文档
- 添加相关截图

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

algorithm/main.py

@@ -0,0 +1,175 @@
+"""
+Algorithm Service - 文档解析、Embedding、LLM 调用服务
+"""
+from fastapi import FastAPI, HTTPException
+from pydantic import BaseModel
+from typing import Optional, List, Dict, Any
+import requests
+import os
+import json
+
+app = FastAPI(title="Algorithm Service")
+
+
+# ========== Models ==========
+
+class ParseRequest(BaseModel):
+    file_url: str
+    engine: str  # markitdown / docling
+    docling_url: Optional[str] = None
+
+
+class EmbeddingRequest(BaseModel):
+    input: str | List[str]
+    model: str
+
+
+class ChatMessage(BaseModel):
+    role: str
+    content: str
+
+
+class ChatRequest(BaseModel):
+    messages: List[ChatMessage]
+    model: str
+    temperature: Optional[float] = 0.7
+    api_key: Optional[str] = None
+    base_url: Optional[str] = None
+
+
+# ========== 文档解析 ==========
+
+@app.post("/parse")
+async def parse_document(req: ParseRequest):
+    """解析文档，支持 markitdown 和 docling"""
+    try:
+        if req.engine == "markitdown":
+            return await parse_with_markitdown(req.file_url)
+        elif req.engine == "docling":
+            return await parse_with_docling(req.file_url, req.docling_url)
+        else:
+            raise HTTPException(status_code=400, detail=f"Unsupported engine: {req.engine}")
+    except Exception as e:
+        return {"success": False, "error": str(e)}
+
+
+async def parse_with_markitdown(file_url: str) -> Dict[str, Any]:
+    """使用 markitdown 解析文档"""
+    try:
+        from markitdown import MarkItDown
+
+        md = MarkItDown()
+        result = md.convert(file_url)
+
+        # 简单分块（按段落分割）
+        content = result.text_content if hasattr(result, 'text_content') else str(result)
+        chunks = [c.strip() for c in content.split('\n\n') if c.strip()]
+
+        return {
+            "success": True,
+            "content": content,
+            "chunks": chunks[:100],  # 限制 chunk 数量
+            "total_pages": 1,
+            "metadata": {
+                "filename": file_url.split('/')[-1]
+            }
+        }
+    except ImportError:
+        raise HTTPException(status_code=500, detail="markitdown not installed. Run: pip install markitdown")
+    except Exception as e:
+        raise HTTPException(status_code=500, detail=f"Failed to parse with markitdown: {str(e)}")
+
+
+async def parse_with_docling(file_url: str, docling_url: Optional[str] = None) -> Dict[str, Any]:
+    """使用 docling 解析文档"""
+    if not docling_url:
+        raise HTTPException(status_code=400, detail="docling_url is required for docling engine")
+
+    try:
+        # 调用 docling 服务
+        response = requests.post(
+            f"{docling_url}/convert",
+            json={"url": file_url},
+            timeout=60
+        )
+
+        if response.status_code != 200:
+            raise HTTPException(status_code=500, detail=f"Docling service error: {response.text}")
+
+        result = response.json()
+
+        content = result.get("text", "")
+        chunks = [c.strip() for c in content.split('\n\n') if c.strip()]
+
+        return {
+            "success": True,
+            "content": content,
+            "chunks": chunks[:100],
+            "total_pages": result.get("num_pages", 1),
+            "metadata": {
+                "filename": file_url.split('/')[-1]
+            }
+        }
+    except requests.exceptions.RequestException as e:
+        raise HTTPException(status_code=500, detail=f"Failed to connect docling service: {str(e)}")
+
+
+# ========== Embedding ==========
+
+@app.post("/embedding")
+async def generate_embedding(req: EmbeddingRequest):
+    """生成 Embedding"""
+    try:
+        # TODO: 根据不同 provider 调用不同的 embedding 服务
+        # 目前返回模拟数据
+
+        texts = [req.input] if isinstance(req.input, str) else req.input
+
+        # 模拟 embedding 返回
+        embeddings = [[0.1] * 1536 for _ in texts]  # 1536 维向量
+
+        return {
+            "success": True,
+            "embeddings": embeddings,
+            "model": req.model
+        }
+    except Exception as e:
+        return {"success": False, "error": str(e)}
+
+
+# ========== Chat ==========
+
+@app.post("/chat")
+async def chat(req: ChatRequest):
+    """LLM 对话"""
+    try:
+        # TODO: 根据 model 和 base_url 调用实际的 LLM 服务
+        # 目前返回模拟数据
+
+        last_message = req.messages[-1].content if req.messages else ""
+
+        return {
+            "success": True,
+            "message": {
+                "role": "assistant",
+                "content": f"Echo: {last_message}"
+            },
+            "usage": {
+                "prompt_tokens": len(last_message),
+                "completion_tokens": 10
+            }
+        }
+    except Exception as e:
+        return {"success": False, "error": str(e)}
+
+
+# ========== Health Check ==========
+
+@app.get("/health")
+async def health():
+    return {"status": "ok"}
+
+
+if __name__ == "__main__":
+    import uvicorn
+    uvicorn.run(app, host="0.0.0.0", port=8081)