Pydantic restricted to v1

[ghjklw]

Describe the bug

@benc-db, you mentioned that dbt asked you to not require Pydantic > 2, so you decided to restrict Pydantic to v1: #843

I am a bit confused as to why it was decided to constraint Pydantic to >=1.10.0, <2 instead of just >=1.10.0 which would have met dbt request. I am also confused about the reason for dbt request, as dbt itself does not publish any dependency on Pydantic.

Based on the date and other related issues/PR, I am guessing that it might be related to issues encountered with the release of Pydantic 2.10 and maybe its impact on Airflow:

• Known v2.10 Issues - Fixes in Progress pydantic/pydantic#10910
• Lower-bind pydantic to 2.10.1 apache/airflow#44284
• 2.10 Running airflow db init fails pydantic/pydantic#10924

However, these were promptly resolved. For example, Airflow itself requires Pydantic v2 and decided to only blacklist the version 2.10.0 of Pydantic.

Making the version 1.9.0 of dbt-databricks effectively incompatible with any project requiring Pydantic v2 seems to be fairly extreme, would it be possible to get any context to the request form dbt?

As a general rule, every project should publish only its constraints. This package has no incompatibility with Pydantic v2, and it should therefore not publish any such constraint ;) If another project has an issue with Pydantic v2, they should publish this constraint themselves to not impact more users than necessary.

This has in turned caused further problems with other packages that were forced to constrain the version of dbt-databricks: astronomer/astronomer-cosmos#1376

[benc-db]

The restriction to < 2 was actually to silence a bunch of warnings related to my internal usage of pydantic. Since I couldn't enforce that pydantic >= 2, I had to use v1 syntax, which in v2 leads to a bunch of warnings. As for why dbt asked me to allow v1, I believe it has to do with the fact that in dbt Cloud (at least in the testing environment), they install all of the adapters, and one of them requires V1. If anyone knows how to suppress the warnings, I would appreciate it. In the meantime, I can remove the upper bound.

[ghjklw]

Thanks a lot for the reply and loosening the restriction 😊

A bit difficult to answer your question about the warnings without more context. I see that youæ're using Ruff, have you tried https://docs.astral.sh/ruff/settings/#lint_flake8-type-checking_runtime-evaluated-base-classes:

[tool.ruff.lint.flake8-type-checking]
runtime-evaluated-base-classes = ["pydantic.BaseModel"]

For mypy, you can check out the Pydantic mypy plugin: https://docs.pydantic.dev/2.10/integrations/mypy/#configuring-the-plugin