Retrieve Full-Texts from Sinequa Dev Servers

[CarsonDavis]

Description

The existing url import code that brings urls into cosmos cannot support our downstream ML tasks. To do that, we need full texts. Afaik, full texts cannot be retrieved via the query endpoint, only the sql endpoint.

Work was started previously on:

• https://github.com/NASA-IMPACT/COSMOS/issues/1016

However, this card was too broad, and we are breaking it into smaller chunks.

Implementation Considerations

• use the engine.sql endpoint to get all existing metadata from the dev servers
• use the engine.sql endpoint to get full_texts from the dev servers
• store the incoming full_text in a new CandidateURL field called full_text
• how will we do error handling?
• Tests: In order to really test the important bits, we would need to emulate a sinequa server, which we are not going to do. Therefore, it is probably not worth it to make any tests right now.
• We should be using tokens, similar to config_generation/minimum_api.py. The actual code will referecnce an environment variable. The token will be put into this file on local, and onto the server when we deploy. Sorry, it goes in .django local file.

Open Questions

• Credentials: for local development, we will use Li's server
• once it goes into staging, it should use existing environment variable?

Deliverable

• dropdown menu
• updated import script
• new field
• data migration

### Tasks
- [ ] https://github.com/NASA-IMPACT/COSMOS/issues/1075

[saifrk]

Please refer to the attached mind map for a visual representation of the changes incorporated in this task.