409 lines
16 KiB
Python
409 lines
16 KiB
Python
import os
|
|
import shutil
|
|
import subprocess
|
|
import tempfile
|
|
import asyncio
|
|
import logging
|
|
from pathlib import Path
|
|
from typing import Optional, Dict, Any
|
|
import json
|
|
|
|
from fastapi import FastAPI, HTTPException, BackgroundTasks, Request
|
|
from fastapi.staticfiles import StaticFiles
|
|
from fastapi.responses import HTMLResponse
|
|
from fastapi.templating import Jinja2Templates
|
|
from pydantic import BaseModel
|
|
import git
|
|
import requests
|
|
from dotenv import load_dotenv
|
|
|
|
# Load environment variables
|
|
load_dotenv()
|
|
|
|
# Configure logging
|
|
logging.basicConfig(
|
|
level=logging.INFO,
|
|
format='%(asctime)s - %(levelname)s - %(message)s',
|
|
handlers=[
|
|
logging.StreamHandler(),
|
|
logging.FileHandler('mcp_server.log')
|
|
]
|
|
)
|
|
logger = logging.getLogger(__name__)
|
|
|
|
app = FastAPI(title="MCP Server", description="AI-powered code editing server")
|
|
|
|
# Mount static files and templates
|
|
app.mount("/static", StaticFiles(directory="static"), name="static")
|
|
templates = Jinja2Templates(directory="templates")
|
|
|
|
# Models
|
|
class GiteaRequest(BaseModel):
|
|
repo_url: str
|
|
token: str # Gitea token instead of username/password
|
|
prompt: str
|
|
ai_model: str = "gemini" # gemini or openai
|
|
model_name: str = "gemini-1.5-pro" # specific model name
|
|
api_key: str # API key from frontend
|
|
|
|
class ProcessResponse(BaseModel):
|
|
task_id: str
|
|
status: str
|
|
message: str
|
|
|
|
# Global storage for task status
|
|
task_status = {}
|
|
|
|
class MCPServer:
|
|
def __init__(self):
|
|
self.repo_path = None
|
|
|
|
async def process_repository(self, task_id: str, request: GiteaRequest):
|
|
"""Main processing function"""
|
|
try:
|
|
logger.info(f"Task {task_id}: Starting process...")
|
|
task_status[task_id] = {"status": "processing", "message": "Starting process..."}
|
|
|
|
# Step 1: Clone repository
|
|
await self._clone_repository(task_id, request)
|
|
|
|
# Step 2: Analyze code with AI
|
|
await self._analyze_with_ai(task_id, request)
|
|
|
|
# Step 3: Commit and push changes
|
|
await self._commit_and_push(task_id, request)
|
|
|
|
logger.info(f"Task {task_id}: Successfully processed repository")
|
|
task_status[task_id] = {"status": "completed", "message": "Successfully processed repository"}
|
|
|
|
except Exception as e:
|
|
logger.error(f"Task {task_id}: Error - {str(e)}")
|
|
task_status[task_id] = {"status": "error", "message": str(e)}
|
|
# Do not delete the repo directory; keep for inspection
|
|
|
|
async def _clone_repository(self, task_id: str, request: GiteaRequest):
|
|
"""Clone repository from Gitea into a persistent directory"""
|
|
logger.info(f"Task {task_id}: Cloning repository...")
|
|
task_status[task_id] = {"status": "processing", "message": "Cloning repository..."}
|
|
|
|
# Extract repo name from URL
|
|
repo_name = request.repo_url.split('/')[-1].replace('.git', '')
|
|
# Persistent directory under /app/data
|
|
data_dir = "/app/data"
|
|
os.makedirs(data_dir, exist_ok=True)
|
|
self.repo_path = os.path.join(data_dir, f"{repo_name}_{task_id}")
|
|
try:
|
|
os.chmod(data_dir, 0o777) # Give full permissions to the data dir
|
|
logger.info(f"Task {task_id}: Created/using data directory: {self.repo_path}")
|
|
except Exception as e:
|
|
logger.warning(f"Task {task_id}: Could not set permissions on data dir: {e}")
|
|
|
|
# Clone repository using git command with credentials
|
|
try:
|
|
# Use git command with credentials in URL
|
|
auth_url = request.repo_url.replace('://', f'://{request.token}@')
|
|
|
|
result = subprocess.run(
|
|
['git', 'clone', auth_url, self.repo_path],
|
|
capture_output=True,
|
|
text=True,
|
|
timeout=300 # 5 minutes timeout
|
|
)
|
|
|
|
if result.returncode != 0:
|
|
logger.error(f"Task {task_id}: Git clone error - {result.stderr}")
|
|
raise Exception(f"Failed to clone repository: {result.stderr}")
|
|
|
|
logger.info(f"Task {task_id}: Successfully cloned repository to {self.repo_path}")
|
|
|
|
except subprocess.TimeoutExpired:
|
|
raise Exception("Repository cloning timed out after 5 minutes")
|
|
except Exception as e:
|
|
raise Exception(f"Failed to clone repository: {str(e)}")
|
|
|
|
async def _analyze_with_ai(self, task_id: str, request: GiteaRequest):
|
|
"""Analyze code with AI model and apply changes"""
|
|
logger.info(f"Task {task_id}: Analyzing code with AI...")
|
|
task_status[task_id] = {"status": "processing", "message": "Analyzing code with AI..."}
|
|
|
|
if request.ai_model == "gemini":
|
|
await self._use_gemini_cli(task_id, request.prompt, request.api_key, request.model_name)
|
|
elif request.ai_model == "openai":
|
|
await self._use_openai_ai(task_id, request.prompt, request.api_key, request.model_name)
|
|
else:
|
|
raise Exception(f"Unsupported AI model: {request.ai_model}")
|
|
|
|
async def _use_gemini_cli(self, task_id: str, prompt: str, api_key: str, model_name: str):
|
|
"""Use Gemini CLI for code analysis and editing"""
|
|
try:
|
|
# Check if Gemini CLI is installed
|
|
try:
|
|
subprocess.run(["gemini", "--version"], check=True, capture_output=True)
|
|
logger.info(f"Task {task_id}: Gemini CLI is available")
|
|
except (subprocess.CalledProcessError, FileNotFoundError):
|
|
raise Exception("Gemini CLI is not installed. Please install it first: https://github.com/google/generative-ai-go/tree/main/cmd/gemini")
|
|
|
|
# Read all code files
|
|
code_content = self._read_code_files()
|
|
logger.info(f"Task {task_id}: Read {len(code_content)} characters of code content")
|
|
|
|
# Create AI prompt
|
|
ai_prompt = f"""
|
|
Analyze the following codebase and make the requested changes:
|
|
|
|
USER REQUEST: {prompt}
|
|
|
|
CODEBASE:
|
|
{code_content}
|
|
|
|
Please provide:
|
|
1. A summary of what changes need to be made
|
|
2. The specific file changes in the format:
|
|
FILE: filename.py
|
|
CHANGES:
|
|
[describe changes or provide new code]
|
|
|
|
Be specific about which files to modify and what changes to make.
|
|
"""
|
|
|
|
# Set API key as environment variable for Gemini CLI
|
|
env = os.environ.copy()
|
|
env['GEMINI_API_KEY'] = api_key
|
|
|
|
logger.info(f"Task {task_id}: Calling Gemini CLI with model: {model_name}")
|
|
|
|
# Call Gemini CLI with specific model, passing prompt via stdin
|
|
result = subprocess.run(
|
|
["gemini", "generate", "--model", model_name],
|
|
input=ai_prompt,
|
|
capture_output=True,
|
|
text=True,
|
|
env=env,
|
|
cwd=self.repo_path,
|
|
timeout=600 # 10 minutes timeout
|
|
)
|
|
|
|
if result.returncode != 0:
|
|
logger.error(f"Task {task_id}: Gemini CLI error - {result.stderr}")
|
|
raise Exception(f"Gemini CLI error: {result.stderr}")
|
|
|
|
logger.info(f"Task {task_id}: Gemini CLI response received ({len(result.stdout)} characters)")
|
|
logger.info(f"Task {task_id}: Gemini CLI raw response:\n{result.stdout}")
|
|
# Store the raw AI response for frontend display
|
|
task_status[task_id]["ai_response"] = result.stdout
|
|
# Parse and apply changes
|
|
await self._apply_ai_changes(result.stdout, task_id)
|
|
|
|
except subprocess.TimeoutExpired:
|
|
raise Exception("Gemini CLI request timed out after 10 minutes")
|
|
except Exception as e:
|
|
raise Exception(f"Gemini CLI error: {str(e)}")
|
|
|
|
async def _use_openai_ai(self, task_id: str, prompt: str, api_key: str, model_name: str):
|
|
"""Use OpenAI for code analysis and editing"""
|
|
try:
|
|
from openai import OpenAI
|
|
|
|
# Configure OpenAI with API key from frontend
|
|
client = OpenAI(api_key=api_key)
|
|
|
|
# Read all code files
|
|
code_content = self._read_code_files()
|
|
logger.info(f"Task {task_id}: Read {len(code_content)} characters of code content")
|
|
|
|
# Create AI prompt
|
|
ai_prompt = f"""
|
|
Analyze the following codebase and make the requested changes:
|
|
|
|
USER REQUEST: {prompt}
|
|
|
|
CODEBASE:
|
|
{code_content}
|
|
|
|
Please provide:
|
|
1. A summary of what changes need to be made
|
|
2. The specific file changes in the format:
|
|
FILE: filename.py
|
|
CHANGES:
|
|
[describe changes or provide new code]
|
|
|
|
Be specific about which files to modify and what changes to make.
|
|
"""
|
|
|
|
logger.info(f"Task {task_id}: Calling OpenAI with model: {model_name}")
|
|
|
|
# Get AI response
|
|
response = client.chat.completions.create(
|
|
model=model_name,
|
|
messages=[
|
|
{"role": "system", "content": "You are a code analysis and editing assistant."},
|
|
{"role": "user", "content": ai_prompt}
|
|
]
|
|
)
|
|
|
|
logger.info(f"Task {task_id}: OpenAI response received")
|
|
|
|
# Parse and apply changes
|
|
await self._apply_ai_changes(response.choices[0].message.content, task_id)
|
|
|
|
except ImportError:
|
|
raise Exception("OpenAI library not installed. Run: pip install openai")
|
|
except Exception as e:
|
|
raise Exception(f"OpenAI error: {str(e)}")
|
|
|
|
def _read_code_files(self) -> str:
|
|
"""Read all code files in the repository"""
|
|
code_content = ""
|
|
file_count = 0
|
|
|
|
for root, dirs, files in os.walk(self.repo_path):
|
|
# Skip .git directory
|
|
if '.git' in dirs:
|
|
dirs.remove('.git')
|
|
|
|
for file in files:
|
|
if file.endswith(('.py', '.js', '.ts', '.jsx', '.tsx', '.html', '.css', '.json', '.md')):
|
|
file_path = os.path.join(root, file)
|
|
try:
|
|
with open(file_path, 'r', encoding='utf-8') as f:
|
|
content = f.read()
|
|
relative_path = os.path.relpath(file_path, self.repo_path)
|
|
code_content += f"\n\n=== {relative_path} ===\n{content}\n"
|
|
file_count += 1
|
|
except Exception as e:
|
|
logger.warning(f"Could not read {file_path}: {e}")
|
|
|
|
logger.info(f"Read {file_count} code files")
|
|
return code_content
|
|
|
|
async def _apply_ai_changes(self, ai_response: str, task_id: str):
|
|
"""Apply changes suggested by AI"""
|
|
logger.info(f"Task {task_id}: Applying AI suggestions...")
|
|
task_status[task_id] = {"status": "processing", "message": "Applying AI suggestions..."}
|
|
|
|
# Parse AI response for file changes
|
|
# This is a simplified parser - you might want to make it more robust
|
|
lines = ai_response.split('\n')
|
|
current_file = None
|
|
current_changes = []
|
|
files_modified = 0
|
|
|
|
for line in lines:
|
|
if line.startswith('FILE:'):
|
|
if current_file and current_changes:
|
|
await self._apply_file_changes(current_file, '\n'.join(current_changes))
|
|
files_modified += 1
|
|
current_file = line.replace('FILE:', '').strip()
|
|
current_changes = []
|
|
elif line.startswith('CHANGES:') or line.strip() == '':
|
|
continue
|
|
elif current_file:
|
|
current_changes.append(line)
|
|
|
|
# Apply last file changes
|
|
if current_file and current_changes:
|
|
await self._apply_file_changes(current_file, '\n'.join(current_changes))
|
|
files_modified += 1
|
|
|
|
logger.info(f"Task {task_id}: Applied changes to {files_modified} files")
|
|
|
|
async def _apply_file_changes(self, filename: str, changes: str):
|
|
"""Apply changes to a specific file"""
|
|
file_path = os.path.join(self.repo_path, filename)
|
|
|
|
if os.path.exists(file_path):
|
|
# For now, we'll append the changes to the file
|
|
# In a real implementation, you'd want more sophisticated parsing
|
|
with open(file_path, 'a', encoding='utf-8') as f:
|
|
f.write(f"\n\n# AI Generated Changes:\n{changes}\n")
|
|
logger.info(f"Applied changes to file: {filename}")
|
|
|
|
async def _commit_and_push(self, task_id: str, request: GiteaRequest):
|
|
"""Commit and push changes back to Gitea"""
|
|
logger.info(f"Task {task_id}: Committing and pushing changes...")
|
|
task_status[task_id] = {"status": "processing", "message": "Committing and pushing changes..."}
|
|
|
|
try:
|
|
repo = git.Repo(self.repo_path)
|
|
|
|
# Add all changes
|
|
repo.git.add('.')
|
|
|
|
# Check if there are changes to commit
|
|
if repo.is_dirty():
|
|
# Commit changes
|
|
repo.index.commit("AI-generated code updates")
|
|
logger.info(f"Task {task_id}: Changes committed")
|
|
|
|
# Push changes
|
|
origin = repo.remote(name='origin')
|
|
origin.push()
|
|
logger.info(f"Task {task_id}: Changes pushed to remote")
|
|
else:
|
|
logger.info(f"Task {task_id}: No changes to commit")
|
|
# Remove the cloned repo directory after push
|
|
if self.repo_path and os.path.exists(self.repo_path):
|
|
shutil.rmtree(self.repo_path)
|
|
logger.info(f"Task {task_id}: Removed cloned repo directory {self.repo_path}")
|
|
except Exception as e:
|
|
raise Exception(f"Failed to commit and push changes: {str(e)}")
|
|
|
|
# Create MCP server instance
|
|
mcp_server = MCPServer()
|
|
|
|
@app.get("/", response_class=HTMLResponse)
|
|
async def read_root(request: Request):
|
|
"""Serve the frontend"""
|
|
return templates.TemplateResponse("index.html", {"request": request})
|
|
|
|
@app.post("/process", response_model=ProcessResponse)
|
|
async def process_repository(request: GiteaRequest, background_tasks: BackgroundTasks):
|
|
"""Process repository with AI"""
|
|
import uuid
|
|
|
|
task_id = str(uuid.uuid4())
|
|
logger.info(f"Starting new task: {task_id}")
|
|
|
|
# Start background task
|
|
background_tasks.add_task(mcp_server.process_repository, task_id, request)
|
|
|
|
return ProcessResponse(
|
|
task_id=task_id,
|
|
status="started",
|
|
message="Processing started"
|
|
)
|
|
|
|
@app.get("/status/{task_id}")
|
|
async def get_status(task_id: str):
|
|
"""Get status of a processing task"""
|
|
if task_id not in task_status:
|
|
raise HTTPException(status_code=404, detail="Task not found")
|
|
return task_status[task_id]
|
|
|
|
@app.get("/health")
|
|
async def health_check():
|
|
"""Health check endpoint"""
|
|
return {"status": "healthy", "message": "MCP Server is running"}
|
|
|
|
if __name__ == "__main__":
|
|
import uvicorn
|
|
uvicorn.run(app, host="0.0.0.0", port=8000)
|
|
|
|
# AI Generated Changes:
|
|
```
|
|
```python
|
|
--- a/main.py
|
|
+++ b/main.py
|
|
@@ -490,9 +490,6 @@
|
|
origin = repo.remote(name='origin')
|
|
origin.push()
|
|
logger.info(f"Task {task_id}: Changes pushed to remote")
|
|
- # Remove the cloned repo directory after push
|
|
- if self.repo_path and os.path.exists(self.repo_path):
|
|
- shutil.rmtree(self.repo_path)
|
|
- logger.info(f"Task {task_id}: Removed cloned repo directory {self.repo_path}")
|
|
except Exception as e:
|
|
raise Exception(f"Failed to commit and push changes: {str(e)}")
|
|
```
|
|
```
|