server/utils/BM25_to_Dict.py

def convert_bm25_to_dict(bm25_results):
    """
    Converts a list of LangChain Document objects from BM25 
    into standard dictionaries matching your Chroma format.
    """
    dict_results = []
    
    for i, doc in enumerate(bm25_results):
        # Extract the persistent chunk_id you generated via UUID
        chunk_id = doc.metadata.get('chunk_id')
        
        # Build the exact dictionary structure your pipeline expects
        bm25_dict = {
            'id': chunk_id,
            'content': doc.page_content,
            'metadata': doc.metadata,
            'similarity_score': 0.0,  # BM25 doesn't provide a normalized score
            'distance': 1.0,          # Maximum distance since it's not a vector match
            'rank': i + 1
        }
        
        dict_results.append(bm25_dict)
        
    return dict_results