RAGandLLM-MCP

Instructions

Implementation Reference

• src/ragandllm_mcp/server.py:58-76 (handler)

  Core handler function implementing the 'get_recipe' tool logic: constructs a JSON payload with user input, fish ID, and name, then POSTs to the /recipe2 API endpoint and returns the response.

  async def get_recipe(user_input: str, fish_id:str, fish_name: str) -> str:
      headers={
          "Content-Type":"application/json;charset=utf-8"
      }
      # 送信するJSONボディを組み立て
      body = {
          "UserInput": user_input,
          "FishID": fish_id,
          "FishName": fish_name
      }
      async with httpx.AsyncClient(timeout=80.0,verify=False) as client:
          response = await client.post(
              f"{API_BASE_URL}/recipe2",
              headers=headers,
              json=body
          )
          response.raise_for_status()
          data = response.json()
          return data

• src/ragandllm_mcp/server.py:102-119 (schema)

  JSON Schema defining the input parameters for the 'get_recipe' tool: UserInput, FishID, and FishName as required strings.

  inputSchema={
      "type": "object",
      "properties": {
          "UserInput": {
              "type": "string",
              "description": "ユーザのレシピに対する希望。例：夏バテ防止レシピ"
          },
          "FishID": {
              "type": "string",
              "description": "魚の画像アップロード後に得られた魚ID"
          },
          "FishName": {
              "type": "string",
              "description": "魚の画像アップロード後に得られた魚の名称"
          }
      },
      "required": ["UserInput","FishID","FishName"]
  }

• src/ragandllm_mcp/server.py:99-120 (registration)

  Registration of the 'get_recipe' tool in the list_tools() handler, including name, description, and input schema.

  types.Tool(
      name="get_recipe",
      description="ユーザプロンプトと前回取得した魚名、魚IDを元にレシピ生成",
      inputSchema={
          "type": "object",
          "properties": {
              "UserInput": {
                  "type": "string",
                  "description": "ユーザのレシピに対する希望。例：夏バテ防止レシピ"
              },
              "FishID": {
                  "type": "string",
                  "description": "魚の画像アップロード後に得られた魚ID"
              },
              "FishName": {
                  "type": "string",
                  "description": "魚の画像アップロード後に得られた魚の名称"
              }
          },
          "required": ["UserInput","FishID","FishName"]
      }
  ),

• src/ragandllm_mcp/server.py:195-231 (handler)

  Dispatch handler in @server.call_tool() for 'get_recipe': extracts arguments, calls the get_recipe function, formats response or error as TextContent.

  elif name == "get_recipe":
  
      if not isinstance(arguments, dict):
          raise ValueError("Invalid forecast arguments")
      
      userinput=arguments["UserInput"]
      fish_id=arguments["FishID"]
      fish_name=arguments["FishName"]
      try:
          answer= await get_recipe(userinput,fish_id,fish_name)
          return [
              types.TextContent(
                  type="text",
                  text=json.dumps(answer,ensure_ascii=False, indent=2)
                  #text=f"🎉 IRIS接続成功！\n📥 応答:：{msg}"
              )
          ]
      
      except Exception as e:
          error_details = {
              "error_type": type(e).__name__,
              "error_message": str(e),
          }
          return [
              types.TextContent(
                  type="text",
                  text=f"エラーが発生しました(recipe): {json.dumps(error_details, ensure_ascii=False, indent=2)}"
              )
          ]
      except httpx.HTTPError as e:
          logger.error(f"IRIS API error: {str(e)}")
          return [
              types.TextContent(
                  type="text",
                  text=f"エラーが発生しました(recipe): {str(e)}"
              )
          ]