cd: workflow for deploying ml pipeline

[Charlie-XIAO]

This PR adds the workflow for deploying the pipeline. See README under /deploy/ for more details. It also adds a GitHub action that triggers redeploy of ChromaDB, then reploy of app and pipeline as well.

[Charlie-XIAO]

I can confirm that the chromadb database indeed gets updated. It is a bit strange that I got this error from the remote frontend:

Status 500 (Internal Server Error); caused by:

  File "/veritastrial/main.py", line 139, in retrieve
    results = CHROMADB_COLLECTION.query(
              ^^^^^^^^^^^^^^^^^^^^^^^^^^

InvalidCollectionException: Collection c86ffd78-d7e9-49e2-97a5-fd06d6bbe7b9 does not exist.

Probably it's because the collection is determined on backend startup so if the old collection is removed and a new collection is created, even if they have the same name they are different. I think in a future PR I will change to only fix the client connection on startup but dynamically get the collection on query (I don't think it's going to be large overhead).

Also the Github workflows need to be tested.