feat(api): openai compliant annotations and vector_content retrieval

[gphorvath]

Description

Ensures our annotations pass conformance with OpenAI's API spec and provides a method of getting chunk/vector information via the LeapfrogAI namespace.

BREAKING CHANGES

Minor: Changed the format of annotation replacement text to match OpenAI spec.

CHANGES

• Adds get_vector endpoint that can be used to retrieve rows from the vector_content table (contains chunk text, embeddings, etc.)
• Adds crud operation for reading the rows of the vector_content table
• Refactors composer and converter to use vector_content instead of file_id's to make it easier to return vector_id
• Adds metadata type to help with converting dict --> object --> dict
• Move Vector type out of crud into typedef
• Adds a test client for testing api via rest calls (used for new tests in this PR)
• Adds integration tests for leapfrogai/vector_stores
• Updates tools/annotations conformance test to match on filename or source (sometimes OpenAI still returns source for some reason)

Related Issue

Addresses #766

Checklist before merging

• Tests, documentation, ADR added or updated as needed
• Followed the Contributor Guide Steps

[netlify]

✅ Deploy Preview for leapfrogai-docs canceled.

Name	Link
🔨 Latest commit	e3626ca
🔍 Latest deploy log	https://app.netlify.com/sites/leapfrogai-docs/deploys/66facdaa0970540008560710

[justinthelaw]

As long as @andrewrisse is good with this API, then I am! Everything looks good to me.

[jalling97]

Confirmed in a test UDS deployment that the vector endpoint works as expected and annotations are properly updated. Thanks for doing the legwork on this @gphorvath!

The base branch was changed.