backend/services/github_service.py

import asyncio
import sqlite3
import logging
import re
import time
from backend.services.usb_vault import get_secret
from backend.services.usb_monitor import get_db_path
from backend.ws.agent_ws import ws_manager

# Ensure DB table exists
def init_github_db():
    try:
        with sqlite3.connect(get_db_path()) as conn:
            conn.execute('PRAGMA journal_mode=WAL')
            conn.execute("""
                CREATE TABLE IF NOT EXISTS commit_history (
                    id INTEGER PRIMARY KEY AUTOINCREMENT,
                    commit_hash TEXT,
                    message TEXT,
                    timestamp DATETIME DEFAULT CURRENT_TIMESTAMP
                )
            """)
            conn.commit()
    except Exception as e:
        logging.error(f"Failed to init github db: {e}")

init_github_db()

# Mutex to prevent concurrent commits
commit_lock = asyncio.Lock()

async def run_git(cmd: str, cwd: str = None) -> tuple[int, str, str]:
    """Runs a git command and returns (returncode, stdout, stderr)."""
    proc = await asyncio.create_subprocess_shell(
        cmd,
        cwd=cwd or '.',
        stdout=asyncio.subprocess.PIPE,
        stderr=asyncio.subprocess.PIPE
    )
    stdout, stderr = await proc.communicate()
    return proc.returncode, stdout.decode().strip(), stderr.decode().strip()

async def execute_commit(message: str) -> dict:
    """Executes a git commit and push with safety checks, queueing, and rate-limit backoff."""
    async with commit_lock:
        logging.info("GITHUB_ENGINE: Acquired commit lock.")
        
        # 1. Fetch token from Vault
        token = get_secret("GITHUB_TOKEN")
        if not token:
            return {"status": "error", "message": "GitHub Token not found in vault. Please inject it securely via EXE."}

        # 2. Incremental Diff Compute
        # Check if there are unstaged changes or untracked files
        await run_git("git add -A")
        code, stdout, stderr = await run_git("git diff --cached --name-only")
        
        if not stdout:
            return {"status": "skipped", "message": "No incremental changes detected. Aborting commit."}
            
        logging.info(f"GITHUB_ENGINE: Staged files for commit:\n{stdout}")
        
        # 3. Commit
        code, c_out, c_err = await run_git(f'git commit -m "{message}"')
        if code != 0:
            return {"status": "error", "message": f"Commit failed: {c_err or c_out}"}
            
        # Extract commit hash
        hash_code, h_out, h_err = await run_git("git rev-parse HEAD")
        commit_hash = h_out if hash_code == 0 else "UNKNOWN_HASH"
        
        # 4. Push with Rate Limit Backoff
        # Using a dummy remote URL structure that includes the PAT.
        # Ensure your git remote is configured correctly if relying on system git.
        # For simplicity, we just run git push.
        
        max_retries = 3
        attempt = 0
        push_success = False
        
        while attempt < max_retries:
            attempt += 1
            logging.info(f"GITHUB_ENGINE: Pushing to remote (Attempt {attempt})...")
            p_code, p_out, p_err = await run_git("git push")
            
            if p_code == 0:
                push_success = True
                break
                
            # 5. Rate-Limit 403/429 Backoff Parsing
            # GitHub usually responds with HTTP 403 or 429 and a "retry-after" header embedded in the stderr proxy response.
            combined_err = (p_err + " " + p_out).lower()
            if "403" in combined_err or "429" in combined_err or "rate limit" in combined_err:
                retry_match = re.search(r"retry-after[:\s]+(\d+)", combined_err)
                sleep_time = int(retry_match.group(1)) if retry_match else 60
                
                logging.warning(f"GITHUB_ENGINE: Rate limit hit. Backing off for {sleep_time} seconds.")
                await ws_manager.broadcast("github:rate_limit", {"wait_seconds": sleep_time})
                await asyncio.sleep(sleep_time)
            else:
                return {"status": "error", "message": f"Push failed: {p_err}"}

        if not push_success:
            return {"status": "error", "message": "Push failed after max retries."}
            
        # 6. SQLite History Logging
        try:
            with sqlite3.connect(get_db_path()) as conn:
                conn.execute('PRAGMA journal_mode=WAL')
                conn.execute("INSERT INTO commit_history (commit_hash, message) VALUES (?, ?)", (commit_hash, message))
                conn.commit()
        except Exception as e:
            logging.error(f"Failed to log commit history: {e}")
            
        # 7. WebSocket Event Emitted
        await ws_manager.broadcast("github:commit_success", {
            "hash": commit_hash,
            "message": message,
            "timestamp": time.time()
        })
        
        logging.info(f"GITHUB_ENGINE: Commit and push successful! Hash: {commit_hash}")
        return {"status": "success", "hash": commit_hash, "message": "Successfully committed and pushed to remote."}