gateway/scripts/migrate_async_to_sync.py

#!/usr/bin/env python3
"""
Migration Script: Convert async def → def for route handlers that don't need async.

This fixes the event-loop blocking issue where synchronous psycopg2 DB operations
inside async def routes block the entire uvicorn event loop, preventing concurrent
request handling.

FastAPI behavior:
- `async def` routes → run directly on the event loop (blocks if sync code inside)
- `def` routes → run in a thread pool automatically (non-blocking)

Usage:
    python scripts/migrate_async_to_sync.py --dry-run     # Preview changes
    python scripts/migrate_async_to_sync.py               # Apply changes

Author: Auto-generated migration script
"""

import os
import re
import sys
import argparse
from pathlib import Path
from typing import Dict, List, Set, Tuple

# Base directory
GATEWAY_DIR = Path(__file__).parent.parent
ROUTES_DIR = GATEWAY_DIR / "modules" / "routes"
FEATURES_DIR = GATEWAY_DIR / "modules" / "features"
AUTH_DIR = GATEWAY_DIR / "modules" / "auth"


# =============================================================================
# Configuration: Functions that MUST stay async
# =============================================================================

# Key: relative file path from gateway dir
# Value: set of function names that must remain async def
_MUST_STAY_ASYNC: Dict[str, Set[str]] = {
    # --- routes/ ---
    "modules/routes/routeAdminAutomationEvents.py": {
        "sync_all_automation_events",       # await syncAutomationEvents(...)
    },
    "modules/routes/routeAdminRbacExport.py": {
        "import_global_rbac",               # await file.read()
        "import_mandate_rbac",              # await file.read()
    },
    "modules/routes/routeDataConnections.py": {
        "get_connections",                  # await token_refresh_service.refresh_expired_tokens(...)
    },
    "modules/routes/routeDataFiles.py": {
        "upload_file",                      # await file.read()
    },
    # These files have many genuinely async routes (httpx, external APIs) -- keep ALL async:
    "modules/routes/routeRealEstate.py": "__ALL__",
    "modules/routes/routeSharepoint.py": "__ALL__",
    "modules/routes/routeVoiceGoogle.py": "__ALL__",
    # Partial keeps in security routes (httpx.AsyncClient, request.json()):
    "modules/routes/routeSecurityGoogle.py": {
        "verify_google_token",              # await client.get(...)
        "auth_callback",                    # await verify_google_token(...), await client.get(...)
        "verify_token",                     # await verify_google_token(...)
        "refresh_token",                    # await request.json()
    },
    "modules/routes/routeSecurityMsft.py": {
        "auth_callback",                    # await client.get(...)
        "refresh_token",                    # await request.json()
    },
    # --- features/ ---
    "modules/features/automation/routeFeatureAutomation.py": {
        "execute_automation_route",          # await executeAutomation(...)
    },
    "modules/features/chatbot/routeFeatureChatbot.py": {
        "stream_chatbot_start",             # await chatProcess(...), contains async event_stream generator
        "event_stream",                     # await request.is_disconnected(), await asyncio.wait_for(...)
        "stop_chatbot",                     # await event_manager.emit_event(...)
    },
    "modules/features/neutralization/routeFeatureNeutralizer.py": {
        "process_sharepoint_files",         # await service.processSharepointFiles(...)
    },
    "modules/features/realestate/routeFeatureRealEstate.py": {
        "process_command",                  # await processNaturalLanguageCommand(...)
        "create_table_record",              # await create_project_with_parcel_data(...)
        "search_parcel",                    # await connector.search_parcel(...), connector._query_building_layer(...)
        "add_parcel_to_project",            # await connector.search_parcel(...)
    },
    "modules/features/trustee/routeFeatureTrustee.py": {
        "create_document",                  # await request.json()
        "upload_document",                  # await file.read()
    },
}

# Files to skip entirely (all routes must stay async)
_SKIP_FILES: Set[str] = {
    "modules/routes/routeRealEstate.py",
    "modules/routes/routeSharepoint.py",
    "modules/routes/routeVoiceGoogle.py",
}

# Helper functions that are fake-async (async def but no await inside)
# These will be converted from async def -> def
_FAKE_ASYNC_HELPERS: Dict[str, Set[str]] = {
    "modules/features/trustee/routeFeatureTrustee.py": {"_validateInstanceAccess", "_validateInstanceAdmin"},
    "modules/features/realestate/routeFeatureRealEstate.py": {"_validateInstanceAccess"},
    "modules/features/chatbot/routeFeatureChatbot.py": {"_validateInstanceAccess"},
    "modules/routes/routeNotifications.py": {"_handleInvitationAction"},
}

# Calls to these functions should have 'await' removed after they become sync
_REMOVE_AWAIT_CALLS: Set[str] = {
    "_validateInstanceAccess",
    "_validateInstanceAdmin",
    "_handleInvitationAction",
}


# =============================================================================
# Migration Logic
# =============================================================================

def _getRelativePath(filePath: Path) -> str:
    """Get path relative to gateway dir."""
    try:
        return str(filePath.relative_to(GATEWAY_DIR)).replace("\\", "/")
    except ValueError:
        return str(filePath)


def _shouldSkipFile(relPath: str) -> bool:
    """Check if file should be skipped entirely."""
    return relPath in _SKIP_FILES or _MUST_STAY_ASYNC.get(relPath) == "__ALL__"


def _mustStayAsync(relPath: str, funcName: str) -> bool:
    """Check if a specific function must stay async."""
    keepSet = _MUST_STAY_ASYNC.get(relPath, set())
    if keepSet == "__ALL__":
        return True
    return funcName in keepSet


def _isFakeAsyncHelper(relPath: str, funcName: str) -> bool:
    """Check if a function is a fake-async helper that should be converted."""
    helpers = _FAKE_ASYNC_HELPERS.get(relPath, set())
    return funcName in helpers


def _processFile(filePath: Path, dryRun: bool = True) -> Dict[str, any]:
    """Process a single file and convert async def → def where appropriate."""
    relPath = _getRelativePath(filePath)
    
    if _shouldSkipFile(relPath):
        return {"file": relPath, "skipped": True, "reason": "all routes must stay async"}
    
    with open(filePath, "r", encoding="utf-8") as f:
        originalContent = f.read()
    
    content = originalContent
    changes = []
    
    # Step 1: Find all async def functions and convert eligible ones
    # Pattern matches: async def function_name(
    asyncDefPattern = re.compile(r'^(\s*)async def (\w+)\s*\(', re.MULTILINE)
    
    convertedFunctions = set()
    
    for match in asyncDefPattern.finditer(originalContent):
        indent = match.group(1)
        funcName = match.group(2)
        
        # Check if this function must stay async
        if _mustStayAsync(relPath, funcName):
            changes.append(f"  KEEP async: {funcName} (must stay async)")
            continue
        
        # Convert async def → def
        convertedFunctions.add(funcName)
        changes.append(f"  CONVERT: async def {funcName} -> def {funcName}")
    
    # Apply conversions
    for funcName in convertedFunctions:
        # Replace "async def funcName(" with "def funcName("
        # Be careful to match the exact function definition
        pattern = re.compile(
            r'^(\s*)async def ' + re.escape(funcName) + r'\s*\(',
            re.MULTILINE
        )
        content = pattern.sub(
            lambda m: f'{m.group(1)}def {funcName}(',
            content
        )
    
    # Step 2: Remove 'await' from calls to converted functions
    # This handles: await _validateInstanceAccess(...) → _validateInstanceAccess(...)
    # And also: result = await someConvertedFunc(...) → result = someConvertedFunc(...)
    for funcName in _REMOVE_AWAIT_CALLS:
        if funcName in convertedFunctions or _isFakeAsyncHelper(relPath, funcName):
            awaitPattern = re.compile(
                r'(\s*)(.*)await\s+' + re.escape(funcName) + r'\s*\(',
                re.MULTILINE
            )
            newContent = awaitPattern.sub(
                lambda m: f'{m.group(1)}{m.group(2)}{funcName}(',
                content
            )
            if newContent != content:
                changes.append(f"  REMOVE await: await {funcName}(...) -> {funcName}(...)")
                content = newContent
    
    # Step 3: Check for any remaining 'await' in converted functions
    # This catches cases where a converted function still has await calls
    remainingAwaits = []
    lines = content.split('\n')
    currentFunc = None
    funcIndent = 0
    
    for i, line in enumerate(lines):
        # Track current function
        defMatch = re.match(r'^(\s*)def (\w+)\s*\(', line)
        asyncDefMatch = re.match(r'^(\s*)async def (\w+)\s*\(', line)
        
        if defMatch and defMatch.group(2) in convertedFunctions:
            currentFunc = defMatch.group(2)
            funcIndent = len(defMatch.group(1))
        elif defMatch or asyncDefMatch:
            currentFunc = None
        elif currentFunc and line.strip() and not line[0].isspace():
            currentFunc = None
        
        # Check for remaining awaits in converted functions
        if currentFunc and 'await ' in line:
            remainingAwaits.append(f"  WARNING: Remaining 'await' in {currentFunc} at line {i+1}: {line.strip()}")
    
    # Build result
    result = {
        "file": relPath,
        "skipped": False,
        "convertedCount": len(convertedFunctions),
        "convertedFunctions": sorted(convertedFunctions),
        "changes": changes,
        "warnings": remainingAwaits,
        "modified": content != originalContent,
    }
    
    # Write file if not dry run and content changed
    if not dryRun and content != originalContent:
        with open(filePath, "w", encoding="utf-8") as f:
            f.write(content)
        result["written"] = True
    else:
        result["written"] = False
    
    return result


def _discoverRouteFiles() -> List[Path]:
    """Discover all route files to process."""
    files = []
    
    # Standard routes
    if ROUTES_DIR.exists():
        for f in sorted(ROUTES_DIR.glob("route*.py")):
            files.append(f)
    
    # Feature routes
    if FEATURES_DIR.exists():
        for f in sorted(FEATURES_DIR.glob("*/routeFeature*.py")):
            files.append(f)
    
    return files


def _main():
    parser = argparse.ArgumentParser(
        description="Migrate async def → def for FastAPI routes with sync DB operations"
    )
    parser.add_argument(
        "--dry-run",
        action="store_true",
        default=False,
        help="Preview changes without writing files (default: apply changes)"
    )
    parser.add_argument(
        "--file",
        type=str,
        default=None,
        help="Process only a specific file (relative to gateway dir)"
    )
    args = parser.parse_args()
    
    dryRun = args.dry_run
    
    print("=" * 70)
    print(f"  FastAPI Route Migration: async def -> def")
    print(f"  Mode: {'DRY RUN (preview only)' if dryRun else 'APPLY CHANGES'}")
    print("=" * 70)
    print()
    
    # Discover files
    if args.file:
        targetFile = GATEWAY_DIR / args.file.replace("/", os.sep)
        if not targetFile.exists():
            print(f"ERROR: File not found: {targetFile}")
            sys.exit(1)
        files = [targetFile]
    else:
        files = _discoverRouteFiles()
    
    print(f"Found {len(files)} route files to analyze\n")
    
    totalConverted = 0
    totalWarnings = 0
    totalModified = 0
    allResults = []
    
    for filePath in files:
        result = _processFile(filePath, dryRun=dryRun)
        allResults.append(result)
        
        if result.get("skipped"):
            print(f"[SKIP]  {result['file']} - SKIPPED ({result.get('reason', '')})")
            continue
        
        converted = result.get("convertedCount", 0)
        warnings = result.get("warnings", [])
        modified = result.get("modified", False)
        
        if converted == 0 and not warnings:
            continue
        
        totalConverted += converted
        totalWarnings += len(warnings)
        if modified:
            totalModified += 1
        
        status = "[DONE] WRITTEN" if result.get("written") else ("[PLAN] WOULD WRITE" if modified else "---")
        print(f"{status}  {result['file']} ({converted} functions)")
        
        for change in result.get("changes", []):
            print(f"    {change}")
        
        for warning in warnings:
            print(f"    [WARN]  {warning}")
        
        print()
    
    # Summary
    print("=" * 70)
    print(f"  SUMMARY")
    print(f"  Files analyzed:    {len(files)}")
    print(f"  Files modified:    {totalModified}")
    print(f"  Functions converted: {totalConverted}")
    print(f"  Warnings:          {totalWarnings}")
    if dryRun:
        print(f"\n  This was a DRY RUN. Run without --dry-run to apply changes.")
    else:
        print(f"\n  Changes applied. Restart the server to take effect.")
    print("=" * 70)
    
    # Return exit code based on warnings
    if totalWarnings > 0:
        print(f"\n[WARN]  There are {totalWarnings} warnings - review before deploying!")
        return 1
    return 0


if __name__ == "__main__":
    sys.exit(_main())