rag-system/main.py
2025-05-01 12:21:47 -05:00

17 lines
671 B
Python

from loaders.pdf_loader import load_pdf
from loaders.web_loader import load_web_crawl
from vectordb.vector_store import add_documents
def main():
print("[1/2] Splitting and processing documents...")
# pdf_documents = load_pdf("data/verint-responsible-ethical-ai.pdf")
# web_documents = load_web(["https://excalibur.mgmresorts.com/en.html"])
web_documents = load_web_crawl("https://firecrawl.dev")
print("[2/2] Generating and storing embeddings...")
# add_documents(pdf_documents)
add_documents(web_documents)
print("Embeddings stored. You can now run the Streamlit app with:\n")
print(" streamlit run app/streamlit_app.py")
if __name__ == "__main__":
main()