hamzaideators/cinerag
🎬 Discover movies effortlessly with CineRAG, a RAG system that combines natural language queries and intelligent retrieval for personalized recommendations.
What's novel
🎬 Discover movies effortlessly with CineRAG, a RAG system that combines natural language queries and intelligent retrieval for personalized recommendations.
Code Analysis
19 files read · 4 roundsA movie discovery RAG system that combines Elasticsearch BM25 and Qdrant vector search via Reciprocal Rank Fusion, optionally re-ranks with a cross-encoder, and generates LLM-powered recommendations with citations, served through a FastAPI backend and Streamlit UI.
Strengths
Clean separation of concerns across ingestion, retrieval, LLM, API, and UI layers with a well-structured hybrid retrieval pipeline (RRF + cross-encoder reranking) and a multi-provider LLM abstraction. The evaluation framework with both retrieval metrics (Recall, MRR, nDCG) and LLM-as-judge quality scoring is a thoughtful addition.
Weaknesses
The README is significantly misleading with fake download links and a generic 'Getting Started' that doesn't match the actual Docker-based setup. Several bugs exist (broken regex in UI humanize_answer, reviews dropped during ES indexing, unused custom analyzer), and the eval set is very small (5-10 queries) with no unit tests.
Score Breakdown
Signal breakdown
Innovation
Craft
Traction
Scope
Evidence
Commits
6
Contributors
2
Files
50
Active weeks
3
Repository
Language
Python
Stars
2
Forks
0
License
MIT