src/scripts/4_title_embeddings.py

from src.title_embd.preprocessing import preprocess_titles
from src.title_embd.embed import generate_title_embeddings
from src.title_embd.index import create_title_faiss_index
import logging

logging.basicConfig(level=logging.INFO, format='%(asctime)s %(levelname)s %(message)s')

def run():
    logging.info("Step 1: Preprocessing document titles")
    title_data = preprocess_titles()
    logging.info("Step 1: Title preprocessing complete")

    logging.info("Step 2: Generating title embeddings")
    generate_title_embeddings(title_data)
    logging.info("Step 2: Title embeddings generated")

    logging.info("Step 3: Creating FAISS index for titles")
    create_title_faiss_index()
    logging.info("Step 3: FAISS index for titles created")

if __name__ == "__main__":
    run()