Upgrade to use Pydantic 2.x.x

[tylerhutcherson]

With the shift in the ecosystem to Pydantic 2.x.x, we need to get redisvl to the same level.

Referenced here: https://github.com/RedisVentures/redisvl/issues/86

Fortunately, there is not a ton of complex usage of Pydantic; which makes this easy! Changes include:

• Shifting the storage classes to use Pydantic to avoid manual (and ugly) field validation. Now, the subclasses inherit from the base and params are all typed as expected.
• Shifting the vectorizer classes to use Pydantic for the same reasons.
• Bump and pin pydantic version in the packge requirements to a reasonable range pydantic>=2.0.0,<3
• use v1 shim for now for safety

[codecov-commenter]

Codecov Report

Attention: 27 lines in your changes are missing coverage. Please review.

Comparison is base (24955a2) 76.95% compared to head (183cb68) 77.16%.

❗ Current head 183cb68 differs from pull request most recent head 0569657. Consider uploading reports for the commit 0569657 to get more accurate results

Files	Patch %	Lines
redisvl/vectorize/text/openai.py	20.00%	8 Missing ⚠️
redisvl/vectorize/text/cohere.py	25.00%	6 Missing ⚠️
redisvl/vectorize/text/vertexai.py	25.00%	6 Missing ⚠️
redisvl/vectorize/text/huggingface.py	58.33%	5 Missing ⚠️
redisvl/storage.py	90.00%	1 Missing ⚠️
redisvl/vectorize/base.py	92.30%	1 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #101      +/-   ##
==========================================
+ Coverage   76.95%   77.16%   +0.21%     
==========================================
  Files          22       22              
  Lines        1380     1384       +4     
==========================================
+ Hits         1062     1068       +6     
+ Misses        318      316       -2     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[Spartee]

Undoubtly will cause some pain with the langchain integration. However, I think as long as there is no inheritance chain that utilizes pydantic v2 with langchain internals we should be fine. the langchain integration should just use langchain.pydantic_v1 for any instantiated classes. As well, any base classes we expect to use in that integration should also utilize langchain.pydantic_v1 if at all possible.

Also, as this is a library we expect to have people use as a dependency within their projects, I would expect to have a wider range of support for pydantic versions even if we have to test multiple versions (we shouldnt, but you never know).

nit: might think about breaking up some of these non-related changes just in case you need to roll this back if it causes major issues at some point down the road.

Up to you, otherwise looks good

[Spartee]

make check?

[Spartee]

may think about relaxing this a bit. Pinning to a patch version for this popular of a library is bound to cause some issues. more on this in the top level comment

[tylerhutcherson]

Will do. Good call

The base branch was changed.