Real-time RAG pipeline for Moroccan news — RSS feeds ingested via Kafka, stored in PostgreSQL + pgvector (HNSW index), vectorized with multilingual-e5-small, and queried through a FastAPI RAG service powered by Gemini 1.5 Flash. Fully containerized with Docker, deployed on GitHub Codespace.
python docker rss postgres kafka spark postgresql gemini rag fastapi streamlit pgvector github-codespace
-
Updated
Aug 31, 2026 - Python