Vespa application creating an index of the CORD-19 dataset.
- Frontend: https://cord19.vespa.ai/
- API: https://api.cord19.vespa.ai/search/?query=sars-cov-2 - API doc
- Deploy a copy of this application by cloning this repo, and deploy it to Vespa Cloud or on your own.
- Accelerated retrieval using weakAnd
- Custom ranking, both BM25, and advanced neural ColBERT
- Document and query time inference using language models (ColBERT)
- Vespa Grouping to allow users to drill down into the result set
- Vespa approximate nearest neighbor search for similar articles functionality
- Vespa de-duping, using model inference to compute N * N document to document similarity, at query time to eliminate near duplicates
$ curl "https://api.cord19.vespa.ai/search/?query=sars-cov-2"
vespa-cloud/vespa-documentation-search