Technical upgrade on third party core libraries (pandas and pydantic)

[jose-turintech]

We're considering internally the migration of two core libraries to their up to date versions in order to follow up with changes on the python ecosystem.

This third party libraries are pandas and pydantic.
Both of this libraries have jumped from their corresponding 1.x version to their 2.x versions, which are now well stablished in the python ecosystem and addopted by most (if not all) relevant libraries that we use in our system.

This discussion thread is meant to hear team thoughts about this kind of migration (known problems, complications, recommendations etc) so we can all align and approach this the most comfortable way possible.

The thread is also meant to share information to all team transparently and in an asynchronous way.

Some resources that could make this transition simpler:

• pydantic : https://docs.pydantic.dev/latest/migration/
• pandas: https://www.shakudo.io/blog/pandas-2-upgrade-and-adapt-guide#h2-4

[mike-turintech]

Those are two very use cases that we should aim to use Artemis and see what are the current limitations and how we can solve such limitations.

[jose-turintech]

One inital good candidate to start evaluating this use-case is the following repo: https://github.com/turintech/evoml-python-packages

It is modularised and it has full CI/CD features, so probably it is a good fit for PR creation, determining success/error on Artemis processing.

We can try from engineering team to analyse it given the scope of this ticket, and ask any question in the process.

[jose-turintech]

http://artemis.turintech.ai/projects/4fd2e36a-68d2-40ad-8e58-0c35aaf1d607