ai_qc/backend/migrate_output_files.py

#!/usr/bin/env python3
"""
Migration script to move existing output files to client-specific folders
based on their profile information.
"""

import os
import json
import re
import shutil
from datetime import datetime
import time

def get_client_from_profile(profile_id):
    """Determine client from profile ID"""
    if not profile_id:
        return 'general'

    profile_lower = str(profile_id).lower()
    if 'loreal' in profile_lower or 'static' in profile_lower:
        return 'loreal'
    elif 'diageo' in profile_lower:
        return 'diageo'
    elif 'unilever' in profile_lower:
        return 'unilever'
    else:
        return 'general'

def extract_profile_from_html(html_path):
    """Extract profile information from HTML report"""
    try:
        with open(html_path, 'r', encoding='utf-8') as f:
            content = f.read()

        # Look for the profile_id in the JSON data at the end of the file
        json_match = re.search(r'var reportData = ({.*?});', content, re.DOTALL)
        if json_match:
            json_str = json_match.group(1)
            data = json.loads(json_str)
            return data.get('profile_id') or data.get('profiles', [None])[0]

        # Alternative: look for profile name in the HTML
        profile_match = re.search(r'Profile: ([^<]+)', content)
        if profile_match:
            return profile_match.group(1).strip().lower().replace(' ', '_')

    except Exception as e:
        print(f"Error extracting profile from {html_path}: {e}")

    return None

def migrate_files(output_folder='output-dev', max_age_days=14, dry_run=False):
    """Migrate existing files to client-specific folders"""
    if not os.path.exists(output_folder):
        print(f"Output folder {output_folder} does not exist")
        return

    cutoff_time = time.time() - (max_age_days * 24 * 60 * 60)
    migrated_count = 0
    deleted_count = 0
    error_count = 0

    print(f"{'DRY RUN - ' if dry_run else ''}Migrating files from {output_folder}")
    print(f"Files older than {max_age_days} days will be deleted")
    print("=" * 60)

    # Get all files in root output folder
    for filename in os.listdir(output_folder):
        file_path = os.path.join(output_folder, filename)

        # Skip directories
        if os.path.isdir(file_path):
            continue

        # Skip non-report files
        if not filename.endswith(('.html', '.json')):
            continue

        try:
            file_age = os.path.getctime(file_path)
            age_days = (time.time() - file_age) / (24 * 60 * 60)

            # Delete files older than max_age_days
            if file_age < cutoff_time:
                print(f"❌ DELETE: {filename} (age: {age_days:.1f} days)")
                if not dry_run:
                    os.remove(file_path)
                deleted_count += 1
                continue

            # Migrate files newer than max_age_days
            if filename.endswith('.html'):
                profile_id = extract_profile_from_html(file_path)
                client = get_client_from_profile(profile_id)

                print(f"📁 MIGRATE: {filename}")
                print(f"   Profile: {profile_id} → Client: {client}")

                # Create client folder
                client_folder = os.path.join(output_folder, client)
                if not dry_run:
                    os.makedirs(client_folder, exist_ok=True)

                # Move file
                dest_path = os.path.join(client_folder, filename)
                if not dry_run:
                    shutil.move(file_path, dest_path)
                    print(f"   Moved to: {client}/{filename}")

                migrated_count += 1

            elif filename.endswith('.json'):
                # Try to read JSON to get profile info
                try:
                    with open(file_path, 'r') as f:
                        data = json.load(f)
                    profile_id = data.get('profile_id') or data.get('profiles', [None])[0]
                    client = get_client_from_profile(profile_id)

                    print(f"📁 MIGRATE: {filename}")
                    print(f"   Profile: {profile_id} → Client: {client}")

                    client_folder = os.path.join(output_folder, client)
                    if not dry_run:
                        os.makedirs(client_folder, exist_ok=True)

                    dest_path = os.path.join(client_folder, filename)
                    if not dry_run:
                        shutil.move(file_path, dest_path)
                        print(f"   Moved to: {client}/{filename}")

                    migrated_count += 1
                except:
                    # If can't read JSON, move to general
                    client = 'general'
                    print(f"📁 MIGRATE: {filename} → {client} (default)")

                    client_folder = os.path.join(output_folder, client)
                    if not dry_run:
                        os.makedirs(client_folder, exist_ok=True)

                    dest_path = os.path.join(client_folder, filename)
                    if not dry_run:
                        shutil.move(file_path, dest_path)

                    migrated_count += 1

        except Exception as e:
            print(f"⚠️  ERROR processing {filename}: {e}")
            error_count += 1

    print("=" * 60)
    print(f"{'DRY RUN ' if dry_run else ''}Summary:")
    print(f"  Migrated: {migrated_count} files")
    print(f"  Deleted:  {deleted_count} files")
    print(f"  Errors:   {error_count} files")

    if dry_run:
        print("\nThis was a DRY RUN. No files were actually moved or deleted.")
        print("Run with --execute to perform the migration.")

if __name__ == '__main__':
    import sys

    # Parse arguments
    dry_run = '--execute' not in sys.argv
    output_folder = 'output-dev'

    if '--production' in sys.argv:
        output_folder = 'output'

    migrate_files(output_folder=output_folder, max_age_days=14, dry_run=dry_run)