replace uses of agg("operation") with aggregate_func()

[igorsugak]

Summary: X-link: https://github.com/ctrl-labs/src2/pull/34567

Pandas-2.0.0 changed default behavior of many methods on DataFrame that take argument numeric_only (for example sum, mean, etc). In the past, if numeric_only, the default value was None, and resulted into non-numeric columns being silently dropped. Now, the default value of numeric_only is True. This has two consequences:

• a non-numeric columns that supports the aggregation function is now included in the result. For example, if the type of a column is str and aggregate operation is sum the result is all strings catenated. In the past this column was not included -- silently dropped.
• a non-numeric column that does not support the aggregation function now generates a runtime error. In the past it was silently dropped.

In an attempt to facilitate the upgrade, this rewrites uses of agg("aggregate_func") calls into aggregate_func(). Later this will enable using Pyre to detect places that requre explicit numeric_only argument to be compatible with Pandas-2.0.0. This is not a semantic change, the effect should be equivalent with the previous spelling.

This was generated with the following one-liner:

fbgr '\.agg\("[a-z]+"\)' -ls | xargs perl -pi -e 's,\.agg\("([a-z]+)"\),.\1(),'
arc f

Inspected all changes.

Differential Revision: D62195922

[facebook-github-bot]

This pull request was exported from Phabricator. Differential Revision: D62195922

[facebook-github-bot]

This pull request has been merged in facebookresearch/HolisticTraceAnalysis@727ae37c0534713f6c3ce4565eaeb6754405f80f.