feat: Initial commit of LLM Tool Proxy

app/main.py

@@ -0,0 +1,79 @@
+import os
+import sys
+from dotenv import load_dotenv
+
+# --- Explicit Debugging & Env Loading ---
+print(f"--- [DEBUG] Current Working Directory: {os.getcwd()}", file=sys.stderr)
+load_result = load_dotenv()
+print(f"--- [DEBUG] load_dotenv() result: {load_result}", file=sys.stderr)
+# ---
+
+import logging
+from fastapi import FastAPI, HTTPException, Depends
+from starlette.responses import StreamingResponse
+from .models import IncomingRequest, ProxyResponse
+from .services import process_chat_request, stream_llm_api, inject_tools_into_prompt
+from .core.config import get_settings, Settings
+
+# --- Logging Configuration ---
+logging.basicConfig(
+    level=logging.INFO,
+    format='%(asctime)s - %(name)s - %(levelname)s - %(message)s',
+    handlers=[
+        logging.FileHandler("llm_proxy.log"),
+        logging.StreamHandler()
+    ]
+)
+logger = logging.getLogger(__name__)
+# --- End of Logging Configuration ---
+
+app = FastAPI(
+    title="LLM Tool Proxy",
+    description="A proxy that intercepts LLM requests to inject and handle tool calls.",
+    version="1.0.0",
+)
+
+@app.on_event("startup")
+async def startup_event():
+    logger.info("Application startup complete.")
+    current_settings = get_settings()
+    logger.info(f"Loaded LLM API URL: {current_settings.REAL_LLM_API_URL}")
+
+@app.post("/v1/chat/completions")
+async def chat_completions(
+    request: IncomingRequest,
+    settings: Settings = Depends(get_settings)
+):
+    """
+    This endpoint mimics the OpenAI Chat Completions API and supports both
+    streaming (`stream=True`) and non-streaming (`stream=False`) responses.
+    """
+    if not settings.REAL_LLM_API_KEY or not settings.REAL_LLM_API_URL:
+        logger.error("REAL_LLM_API_KEY or REAL_LLM_API_URL is not configured.")
+        raise HTTPException(status_code=500, detail="LLM API Key or URL is not configured.")
+
+    # Prepare messages, potentially with tool injection
+    # This prepares the messages that will be sent to the LLM backend
+    messages_to_llm = request.messages
+    if request.tools:
+        messages_to_llm = inject_tools_into_prompt(request.messages, request.tools)
+
+    # Handle streaming request
+    if request.stream:
+        logger.info(f"Initiating streaming request with {len(messages_to_llm)} messages.")
+        generator = stream_llm_api(messages_to_llm, settings)
+        return StreamingResponse(generator, media_type="text/event-stream")
+
+    # Handle non-streaming request
+    try:
+        logger.info(f"Initiating non-streaming request with {len(messages_to_llm)} messages.")
+        response_message = await process_chat_request(messages_to_llm, request.tools, settings)
+        logger.info("Successfully processed non-streaming request.")
+        return ProxyResponse(message=response_message)
+    except Exception as e:
+        logger.exception("An unexpected error occurred during non-streaming request.")
+        raise HTTPException(status_code=500, detail=f"An unexpected error occurred: {str(e)}")
+
+@app.get("/")
+def read_root():
+    return {"message": "LLM Tool Proxy is running."}