Upgrade vectorize support to Astra's Public Preview available service #33
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
zzzming
approved these changes
May 27, 2024
jordanrfrazier
approved these changes
May 28, 2024
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
I have extended the vectorize support to cover for what has been launched in public preview this week.
Namely, this means bumping astrapy to access the latest parameters "service options", and the various ways of passing secrets to embedding services:
A further param ( <--> the header) is added to the vector store and the helper "util" class to achieve that (plus further ValueError for new unacceptable combinations of parameters)
Much has changed in the tests, which end up unavoidably convoluted for testing these things according to what services are available on which environments: the "main" vectorize is now openai, so that "regular" astra have it. This requires
(a) an openai key - I have not entered any in the CI yet
(b) which must be also set as a secret in the KMS for that account (same)
These two things I have yet to fix on the repo CI, but so far I tested locally and everything is perfect!
The above cover for the kms+header usages.
What remains is the no-auth (=nvidia nemo) way. Right now, this is available only if your Astra org is manually flagged to try that. That is the reason I switched the standard vectorize testing service to openai. However, I have access to an org "in the club", so I could test that vectorize flavour as well: all green.