NO-SNOW: Add modin as dependency for AST tests.

[sfc-gh-jrose]

1. Which Jira issue is this PR addressing? Make sure that there is an accompanying issue to your PR.

   Fixes NO-SNOW

2. Fill out the following pre-review checklist:

   • I am adding a new automated test(s) to verify correctness of my new code
     • If this test skips Local Testing mode, I'm requesting review from @snowflakedb/local-testing
   • I am adding new logging messages
   • I am adding a new telemetry message
   • I am adding new credentials
   • I am adding a new dependency
   • If this is a new feature/behavior, I'm adding the Local Testing parity changes.
   • I acknowledge that I have ensured my changes to be thread-safe. Follow the link for more information: Thread-safe Developer Guidelines

3. Please describe how your code solves the related issue.

   One of the AST tests has been failing in daily-precommits due to requiring modin. Modin was not a tox dependency so I've added it.

[sfc-gh-lspiegelberg]

Can you point out which test is failing due to Modin? Modin doesn't work with python 3.8, so adding it as a dependency may break those gates immediately.

[sfc-gh-lspiegelberg]

ast should not have a dependency on modin, we do not use anything from it for the AST tests to the best of my knowledge. Do you have a link/CI run to point why this fix is needed?

[sfc-gh-jrose]

This is the test that breaks:
https://github.com/snowflakedb/snowpark-python/blob/main/tests/ast/data/Dataframe.to_snowpark_pandas.test
The function it is testing imports modin:
https://github.com/snowflakedb/snowpark-python/blob/main/src/snowflake/snowpark/dataframe.py#L1243

[sfc-gh-jrose]

Here's an example failure: https://github.com/snowflakedb/snowpark-python/actions/runs/12926763404/job/36050719429

[sfc-gh-lspiegelberg]

Thank you for the pointers. What baffles me that this fails, I don't recall the tests failing when the code was checked in. Did something change? To me the snowpark_pandas test is an integration test with the snowpark pandas api, it may make sense to separate these types of tests out. Quickly adding modin may fix this, but I am concerned about 3.8 (despite its EOL). Let's involve @sfc-gh-vbudati here, given she authored the originally code.

@sfc-gh-vbudati: Do you have an idea/suggestion how we could avoid adding modin as dependency and still test the to_snowpark_pandas API? Maybe we can convert this to an integration test and run only with python3.9+?

[sfc-gh-vbudati]

To my knowledge, the test was working when it was checked in. The error here (https://github.com/snowflakedb/snowpark-python/actions/runs/12926763404/job/36050719429) looks like it's because modin is not being imported.

@sfc-gh-lspiegelberg the to_snowpark_pandas.test won't run unless the Python version is 3.9+ (look at load_test_cases() in test_ast_driver.py. So it should be fine to import modin after a check to see that the Python version is 3.9+.