Snappy - Spatially-Grounded Document Retrieval via Patch-to-Region Relevance Propagation

Read the Full Paper on arxiv.org: Spatially-Grounded Document Retrieval via Patch-to-Region Relevance Propagation

Snappy implements region-level document retrieval by unifying vision-language models with OCR through spatial coordinate mapping. Unlike traditional systems that return entire pages (VLMs) or lack semantic grounding (OCR-only), Snappy uses ColPali's patch-level similarity scores as spatial relevance filters over OCR-extracted regions; operating entirely at inference time without additional training.

Motivation

Vision-language models like ColPali achieve state-of-the-art document retrieval by embedding pages as images with fine-grained patch representations. However, they return entire pages as retrieval units, introducing irrelevant content into RAG context windows. Conversely, OCR systems extract structured text with bounding boxes but cannot assess which regions are relevant to a query.

Snappy bridges these paradigms through patch-to-region relevance propagation. The approach formalizes coordinate mapping between vision transformer patch grids (32×32) and OCR bounding boxes, repurposing ColPali's late interaction mechanism to generate interpretability maps. Patch similarity scores propagate to OCR regions via IoU-weighted intersection, enabling two-stage retrieval: efficient candidate retrieval using mean-pooled embeddings, followed by full-resolution region reranking.

This yields region-level granularity (return specific paragraphs, tables, or figures instead of entire pages), operates purely at inference time (no additional training), provides spatial interpretability (visual heatmaps showing which document regions match query tokens), and combines VLM semantic understanding with OCR structural precision in a production-ready system.

Demo 🎬

https://github.com/user-attachments/assets/95b37778-8dc3-4633-b4b2-639fc5017470

Quick start

Prerequisites: Docker with Compose.

1. Copy envs and set the essentials

cp .env.example .env

Set OPENAI_API_KEY (required for chat). Other defaults are ready for local use.

2. Start services

docker compose up -d

3. Open the UI - Frontend: http://localhost:3000 - Backend: http://localhost:8000/docs (OpenAPI)

Architecture

Configuration

Troubleshooting

• Progress stuck on upload/indexing: ensure Poppler is installed for PDF rasterization and check backend logs.
• Missing images: confirm LOCAL_STORAGE_PATH is accessible and check allowed domains in frontend/next.config.ts.
• Files lost after restart: ensure the snappy_storage Docker volume is properly mounted (configured by default in docker-compose.yml).
• OCR not running: ensure DEEPSEEK_OCR_ENABLED=true and /ocr/health reachable.
• Config not sticking: /config/update is runtime-only; edit .env for persistence.

Documentation

Core concepts — Late Interaction explains multi-vector retrieval, MaxSim scoring, and two-stage search. Spatial Grounding covers how spatial information flows from pixels to regions. Analysis discusses when to use vision vs text RAG.

System internals — Streaming Pipeline details how the indexer overlaps stages. Architecture provides deeper component and flow descriptions. Configuration is the full config reference.

Development — Service-specific guides are in frontend/README.md and backend/README.md. See CONTRIBUTING.md for contribution guidelines.

Star History

License

MIT