cern-sis / issues-inspire

0 stars 0 forks source link

Create embeddings with fulltext #487

Open miguelgrc opened 3 days ago

miguelgrc commented 3 days ago

For Pascal's thesis only the title and abstract of papers where added to the embeddings. We should now add the fulltext as well. The records in the vector DB should naturally have as metadata the Inspire ID to be able to retrieve extra information (citations, etc) if needed.