glam/scripts/fix_ghcid_mismatches.py

#!/usr/bin/env python3
"""
Fix GHCID mismatches where ghcid_current doesn't match the filename.

Per AGENTS.md Rule on PID Stability, the ghcid_current MUST match the filename.
This script:
1. Finds all files where ghcid_current != filename
2. Updates ghcid_current to match filename
3. Updates ghcid_history to record the correction
4. Logs all changes for audit trail
"""

import os
import yaml
import re
from datetime import datetime, timezone
from pathlib import Path


def get_ghcid_from_file(filepath: Path) -> str | None:
    """Extract ghcid_current from a YAML file."""
    with open(filepath, 'r', encoding='utf-8') as f:
        content = f.read()
    
    # Use regex to find ghcid_current value
    match = re.search(r'ghcid_current:\s*[\'"]?([^\s\'"]+)[\'"]?', content)
    if match:
        return match.group(1)
    return None


def fix_ghcid_mismatch(filepath: Path, correct_ghcid: str, old_ghcid: str) -> bool:
    """
    Fix the ghcid_current in a file to match the filename.
    
    Returns True if file was modified, False otherwise.
    """
    with open(filepath, 'r', encoding='utf-8') as f:
        content = f.read()
    
    # Replace ghcid_current value
    # Handle both quoted and unquoted values
    patterns = [
        (rf"ghcid_current:\s*'{re.escape(old_ghcid)}'", f"ghcid_current: '{correct_ghcid}'"),
        (rf'ghcid_current:\s*"{re.escape(old_ghcid)}"', f'ghcid_current: "{correct_ghcid}"'),
        (rf'ghcid_current:\s*{re.escape(old_ghcid)}(?=\s|$)', f'ghcid_current: {correct_ghcid}'),
    ]
    
    new_content = content
    for pattern, replacement in patterns:
        new_content = re.sub(pattern, replacement, new_content, count=1)
        if new_content != content:
            break
    
    if new_content == content:
        print(f"  WARNING: Could not find ghcid_current to replace in {filepath.name}")
        return False
    
    # Add correction note to ghcid_history if it exists, or add a comment
    timestamp = datetime.now(timezone.utc).strftime('%Y-%m-%dT%H:%M:%SZ')
    
    # Check if there's a ghcid_history section
    if 'ghcid_history:' in new_content:
        # Add entry to existing history
        history_entry = f"""
  - ghcid_value: {old_ghcid}
    valid_from: null
    valid_to: '{timestamp}'
    reason: "Corrected: ghcid_current was incorrectly set to {old_ghcid}, should be {correct_ghcid} (filename mismatch fix)"
"""
        # Find ghcid_history and add entry after it
        new_content = re.sub(
            r'(ghcid_history:\s*\n)',
            r'\1' + history_entry,
            new_content,
            count=1
        )
    else:
        # Add a comment about the correction
        correction_comment = f"# GHCID Correction {timestamp}: Changed from {old_ghcid} to {correct_ghcid} (filename mismatch)\n"
        # Add after ghcid_current line
        new_content = re.sub(
            rf'(ghcid_current:\s*{re.escape(correct_ghcid)}[^\n]*\n)',
            r'\1' + correction_comment,
            new_content,
            count=1
        )
    
    with open(filepath, 'w', encoding='utf-8') as f:
        f.write(new_content)
    
    return True


def main():
    custodian_dir = Path('/Users/kempersc/apps/glam/data/custodian')
    
    print("=" * 70)
    print("GHCID Mismatch Fixer")
    print("=" * 70)
    print(f"Scanning: {custodian_dir}")
    print()
    
    mismatches = []
    
    # Find all mismatches
    for filepath in sorted(custodian_dir.glob('*.yaml')):
        filename_ghcid = filepath.stem  # filename without .yaml
        file_ghcid = get_ghcid_from_file(filepath)
        
        if file_ghcid and file_ghcid != filename_ghcid:
            mismatches.append({
                'filepath': filepath,
                'filename_ghcid': filename_ghcid,
                'file_ghcid': file_ghcid
            })
    
    print(f"Found {len(mismatches)} files with GHCID mismatches")
    print()
    
    if not mismatches:
        print("No mismatches to fix!")
        return
    
    # Display mismatches
    print("Mismatches to fix:")
    print("-" * 70)
    for m in mismatches:
        print(f"  {m['filepath'].name}")
        print(f"    Current ghcid_current: {m['file_ghcid']}")
        print(f"    Should be:             {m['filename_ghcid']}")
        print()
    
    # Fix them
    print("Fixing mismatches...")
    print("-" * 70)
    
    fixed_count = 0
    failed_count = 0
    
    for m in mismatches:
        print(f"Fixing: {m['filepath'].name}")
        if fix_ghcid_mismatch(m['filepath'], m['filename_ghcid'], m['file_ghcid']):
            print(f"  OK: {m['file_ghcid']} -> {m['filename_ghcid']}")
            fixed_count += 1
        else:
            print(f"  FAILED")
            failed_count += 1
    
    print()
    print("=" * 70)
    print(f"Summary: Fixed {fixed_count} files, Failed {failed_count} files")
    print("=" * 70)
    
    # Write audit log
    log_path = custodian_dir / 'ghcid_mismatch_fix_log.txt'
    with open(log_path, 'w', encoding='utf-8') as f:
        f.write(f"GHCID Mismatch Fix Log\n")
        f.write(f"Generated: {datetime.now(timezone.utc).isoformat()}\n")
        f.write(f"=" * 70 + "\n\n")
        f.write(f"Total mismatches found: {len(mismatches)}\n")
        f.write(f"Fixed: {fixed_count}\n")
        f.write(f"Failed: {failed_count}\n\n")
        f.write("Details:\n")
        f.write("-" * 70 + "\n")
        for m in mismatches:
            f.write(f"File: {m['filepath'].name}\n")
            f.write(f"  Old ghcid_current: {m['file_ghcid']}\n")
            f.write(f"  New ghcid_current: {m['filename_ghcid']}\n\n")
    
    print(f"\nAudit log written to: {log_path}")


if __name__ == '__main__':
    main()