added negative value support in pct_changes of generic.py

[paridhisharma24]

closes #56618

[asishm-wk]

Thanks for the PR. Few points:

1. We can't change the default behavior without a deprecation cycle here (and reading the comments of the issue, don't think the default needs to change either).
2. You'll need to add another argument/parameter to the pct_change method that indicates whether or not you want to take absolute value of the denominator (similar changes would be required in your tests as well)

[paridhisharma24]

@asishm-wk I am not able to figure out the issue with the failing check
Could you please help with the same

[asishm]

The error comes from mypy - if you click on the failed build.

pandas/core/generic.py:12152: error: Unsupported left operand type for / ("Self") [operator]
pandas/core/generic.py:12152: note: Right operand is of type "Self | DataFrame"

Looks like you removed the type ignore comment when you made the changes (needs to be there in both the if and else statements).

[asishm]

tests should likely go in pandas/tests/series/methods/test_pct_change.py and see if you can use an existing test and parametrize the new argument.

[asishm]

Suggested change

	rs = rs = (data - shifted) / abs(shifted) # type: ignore[operator]
	rs = (data - shifted) / abs(shifted) # type: ignore[operator]

[asishm]

is the import and setting the seed below required?

[paridhisharma24]

will check and remove if not required

[WillAyd]

I appreciate the PR but I am not sure it is worth expanding the API for this.

[mroeschke]

Thanks for the PR but it is a bit too early to craft a solution to the issue as it's still being actively discussed. Going to close but we can reopen if it has more acceptance from the core developers