README.md

dandiset-search-service

Running locally

To run the services, firt set the environment variables:

export QDRANT_HOST=
export QDRANT_PORT=
export QDRANT_COLLECTION_NAME=
export QDRANT_VECTOR_SIZE=
export QDRANT_API_KEY=
export OPENAI_API_KEY=
export DANDI_API_KEY=

Running with docker compose pulling images from github packages:

docker compose up

Running with docker compose building images locally:

docker compose -f docker-compose.dev.yaml up --build

Services will be available at:

• http://localhost:5173 - Frontend
• http://localhost:8000 - REST API
• http://localhost:8000/docs - REST API docs
• http://localhost:6333 - Qdrant API
• http://localhost:6333/dashboard - Qdrant dashboard

When services are up, populate the Qdrant vector store with vectorized data from DANDI:

python populate_qdrant.py