scripts/run_ingestion.py

# scripts/run_ingestion.py
import sys
import os

# Add the project root to the sys.path
# Assuming this script is in the project root or a 'scripts' subdir at root
sys.path.insert(0, os.path.abspath(os.path.join(os.path.dirname(__file__), '..')))

from src.ingestion_orchestrator.orchestrator import IngestionOrchestrator
from config.settings import DOCS_FOLDER # Use the configured docs folder
import logging

logger = logging.getLogger(__name__)

if __name__ == "__main__":
    # --- Financial Ministry Adaptation ---
    # Add argument parsing if needed (e.g., specify a different docs folder)
    # Implement more sophisticated scheduling if needed (e.g., run daily, weekly)
    # ------------------------------------
    logger.info("Starting the RAG ingestion process.")
    try:
        orchestrator = IngestionOrchestrator()
        orchestrator.run_ingestion_pipeline(docs_folder=DOCS_FOLDER) # Use configured folder
        logger.info("RAG ingestion process finished.")
    except Exception as e:
        logger.critical(f"RAG ingestion process failed: {e}")
        sys.exit(1) # Exit with an error code