import yaml
import os

def load_yaml(path):
    with open(path, 'r') as f:
        return yaml.safe_load(f) or {}

def save_yaml(data, path):
    with open(path, 'w') as f:
        yaml.dump(data, f, sort_keys=False, allow_unicode=True)

def merge_slot_fixes():
    full_path = 'slot_fixes_backup_full.yaml'
    current_path = 'slot_fixes_current.yaml'
    target_path = 'schemas/20251121/linkml/modules/slots/slot_fixes.yaml'

    full_data = load_yaml(full_path)
    current_data = load_yaml(current_path)

    full_list = full_data.get('fixes', []) if isinstance(full_data, dict) else full_data
    current_list = current_data.get('fixes', []) if isinstance(current_data, dict) else current_data

    # Normalize structure if it's a list at root vs dict with 'fixes' key
    # The standard seems to be a list, but let's check
    if not isinstance(full_list, list):
        # Maybe it's a list at root
        full_list = full_data if isinstance(full_data, list) else []
    
    if not isinstance(current_list, list):
        current_list = current_data if isinstance(current_data, list) else []

    print(f"Loaded {len(full_list)} entries from backup.")
    print(f"Loaded {len(current_list)} entries from current file.")

    # Map current entries by original_slot_id for easy lookup
    current_map = {}
    for entry in current_list:
        if 'original_slot_id' in entry:
            current_map[entry['original_slot_id']] = entry
        elif 'orignal_slot_id' in entry: # Handle typo found in previous edits
             current_map[entry['orignal_slot_id']] = entry

    # Merge strategy:
    # 1. Start with full backup list
    # 2. Update matching entries with current version
    # 3. Track which current entries were used
    
    merged_list = []
    processed_ids = set()

    for entry in full_list:
        slot_id = entry.get('original_slot_id') or entry.get('orignal_slot_id')
        
        if slot_id and slot_id in current_map:
            # Use the CURRENT version (contains recent updates)
            merged_list.append(current_map[slot_id])
            processed_ids.add(slot_id)
        else:
            # Use the BACKUP version (restoring deleted entry)
            merged_list.append(entry)

    # 4. Add any NEW entries from current that weren't in backup
    for entry in current_list:
        slot_id = entry.get('original_slot_id') or entry.get('orignal_slot_id')
        if slot_id and slot_id not in processed_ids:
            print(f"Adding new entry not in backup: {slot_id}")
            merged_list.append(entry)

    print(f"Total entries in merged file: {len(merged_list)}")

    # Save result
    # Check if the original file format was a list or dict
    if isinstance(full_data, dict) and 'fixes' in full_data:
        output_data = {'fixes': merged_list}
    else:
        output_data = merged_list

    save_yaml(output_data, target_path)
    print("Successfully restored slot_fixes.yaml")

if __name__ == "__main__":
    merge_slot_fixes()