arsalanafzal010/SmartRAG
📄 Enable smart conversations with documents, images, and audio files using this advanced Retrieval-Augmented Generation system.
What's novel
📄 Enable smart conversations with documents, images, and audio files using this advanced Retrieval-Augmented Generation system.
Code Analysis
8 files read · 3 roundsA local multimodal RAG chatbot that ingests PDFs, Word docs, images, and audio files, extracts text via OCR/transcription, stores embeddings in ChromaDB/FAISS, and answers questions using Ollama LLMs through a Streamlit UI.
Strengths
Clean layered architecture with abstract base classes, multiple vector store backends, and graceful degradation when optional dependencies (Whisper, Tesseract, BLIP) are missing. Good separation between processing, storage, and orchestration layers.
Weaknesses
No test suite at all, relevance filtering is a near-no-op (just takes top 3 chunks >50 chars), confidence scores are hardcoded, embeddings are generated one-at-a-time in a loop, and the 73KB chatbot_app.py is mostly CSS with minimal actual logic.
Score Breakdown
Signal breakdown
Innovation
Craft
Traction
Scope
Evidence
Commits
15
Contributors
2
Files
38
Active weeks
4
Repository
Language
Python
Stars
2
Forks
0
License
MIT