Vault

import os, re, shutil
from datetime import datetime
from nlp_classifier import extract_entities

def auto_organize_archive(root_path):
    audit_log = []
    for root, dirs, files in os.walk(root_path):
        for file in files:
            file_path = os.path.join(root, file)
            
            # NLP processing for unstructured files
            metadata = extract_entities(file_path)
            doc_type = metadata.get('type', 'Uncategorized')
            client = metadata.get('client_name', 'Unknown')
            date_obj = metadata.get('date', datetime.now())
            
            # Procedural directory layout: /Year/Client/Type/
            year = date_obj.strftime('%Y')
            new_dir = os.path.join('Z:\\Archivio_Aziendale', year, client, doc_type)
            os.makedirs(new_dir, exist_ok=True)
            
            # Clean string matching & regex sanitization
            clean_name = re.sub(r'[^A-Za-z0-9_.-]', '', f"{date_obj.strftime('%Y%m%d')}_{doc_type.upper()}_{client.replace(' ', '')}{os.path.splitext(file)[1]}")
            dest_path = os.path.join(new_dir, clean_name)
            
            shutil.move(file_path, dest_path)
            audit_log.append({'original': file_path, 'moved_to': dest_path, 'status': 'SUCCESS'})
            
    return {'processed': len(audit_log), 'anomalies': 0, 'log': audit_log}

output = auto_organize_archive('C:\\Scansioni_Inbox')