FIX-#6367: Enable support for 'groupby.size()' in reshuffling groupby

[dchigarev]

What do these changes do?

This PR removes drop=False parameter overwriting for the groupby object used in groupby.size(). The overwriting was originally introduced by this commit bb8b23d more than 3 years ago in an attempt to fix groupby(as_index=False, ...).size(). Nowadays, this case works perfectly fine even without this drop=False overwriting.

The problem that this overwriting had brought, is that it confused implementations on whether the 'by' columns actually came from the grouping frame or not (drop=True means that the grouping column came from the same frame and drop=False means that the 'by' columns are foreign).

The experimental groupby implementation only allows for the 'by' columns to originate from the same frame, meaning that this drop=False overwriting never allowed us to run groupby.size() using the new implementation. This PR fixes that problem.

modin/modin/core/storage_formats/pandas/query_compiler.py

Lines 3421 to 3425 in f24ba6f

	if not is_all_column_names:
	raise NotImplementedError(
	"Reshuffling groupby is only supported when grouping on a column(s) of the same frame. "
	+ "https://github.com/modin-project/modin/issues/5926"
	)

• first commit message and PR title follow format outlined here

  NOTE: If you edit the PR title to match this format, you need to add another commit (even if it's empty) or amend your last commit for the CI job that checks the PR title to pick up the new PR title.

• passes flake8 modin/ asv_bench/benchmarks scripts/doc_checker.py
• passes black --check modin/ asv_bench/benchmarks scripts/doc_checker.py
• signed commit with git commit -s
• Resolves BUG: Reshuffling groupby doesn't work for groupby.size() method #6367
• tests added and passing
• module layout described at docs/development/architecture.rst is up-to-date

[dchigarev]

explicitly setting drop=False seems useless for the result of groupby.size(), all the tests are passing even without this tweak

[anmyachev]

LGTM

[vnlitvinov]

could you pls explain how this works?

[dchigarev]

modin_config is a fixture that modifies the modin.config variables just for this specific test suit.

The inderect=True parameter allows parametrization of the fixture, so the modify_config would take the parameters specified in the list, otherwise it won't be able to receive those parameters

[vnlitvinov]

what assert are we expecting there?

[dchigarev]

The test verifies that setting the ExperimentalGroupbyImpl to True will guarantee us to engage the reshuffling groupby implementation.

The patched method is the one that actually builds range-partitioning and executes functions over it, so if the code made it through to this method, we can assume that the reshuffling implementation was used, rather than the default MapReduce one.

[anmyachev]

@dchigarev please resolve conflicts

[anmyachev]

@vnlitvinov do you have any other comments?