pydantic 2.0

[apirogov]

Pydantic 2.0 will introduce many breaking changes, but could be up to 10x faster than current pydantic, because it will be backed by pydantic-core which is implemented in Rust.

A year or two from now, especially if we run into performance issues due to lots of pydantic models and validation going on, it could be a good time to tackle the migration.

This would affect mostly the schema plugins, where some "hacks" and mixins could become obsolete, and others would need to be ported/adapted/re-invented. While it could be a few weeks of work, it does not look like an insurmountable problem and could lead to a massive performance boost.

In any case this should wait until pydantic 2 adoption spreads out to the pydantic-heavy libraries, documentations are updated and the dust settles.

[apirogov]

Another thing to consider is the update of pydantic-yaml to 1.0, which also has breaking changes. Would make sense to address this migration in one go.