Embedding Pipeline

Streamlit web app for generating vector embeddings from PDF documents and images and searching over them using Nomic's ColNomic Embed Multimodal 3B model.

Features

• Multi-PDF and image upload (PNG, JPG, JPEG, WebP) with batch or incremental embedding
• PDF page rendering at configurable DPI (72, 150, 300) via PyMuPDF
• Multi-vector embeddings with ColNomic Embed Multimodal 3B
• Cross-document text search with top-K and score threshold filtering
• Optional per-document search filtering
• Automatic device selection (MPS > CUDA > CPU)
• Per-document and combined JSON downloads with embeddings, DPI, and timing

Setup

uv sync
uv run streamlit run streamlit_app.py

Development

uv run ruff check .   # lint
uv run ruff format .  # format
uv run ty check       # typecheck
uv run pytest         # test