TYP: @final for some Index methods

[jbrockmendel]

cc @simonjayhawkins

Tangential question: when I run flake8 it passes, but the pre-commit hook often gives me a bunch of messages like pandas/core/indexes/base.py:956:51: F821 undefined name 'ignore'. Is this something I can fix by pinning a version or something?

[simonjayhawkins]

Tangential question: when I run flake8 it passes, but the pre-commit hook often gives me a bunch of messages like pandas/core/indexes/base.py:956:51: F821 undefined name 'ignore'. Is this something I can fix by pinning a version or something?

not sure. I don't have pre-comit setup. (One thing that maybe relevant is that pyflakes is currently installed in the dev environment separately to allow mypy error codes #35311 since we have flake8 pinned <3.8.0 #34150)

[simonjayhawkins]

before we have a dependence on pandas._vendored.typing_extensions, I think we need some sort of tests, see #36000 (comment)

When trying to use Literal and Final types, i'm getting errors like

error: Variable "pandas._vendored.typing_extensions.Final" is not valid as a type  [valid-type]

not sure what i'm doing wrong.

[WillAyd]

Lgtm - I think this is nice to do

[jbrockmendel]

Adding @final incorrectly to get_loc and running mypy, it isn't flagging it as incorrect. It gets flagged correctly if I import final from typing directlyinstead ofpandas._vendored.typing_extensions`. Also if I import from a non-vendored typing_extensions it seems to work.

It also isn't flagging private methods with incorrect final tag, regardless of where i get final from

[jbrockmendel]

since this doesnt appear to have the desired effects, closing to clear the queue

[jbrockmendel]

@WillAyd @simonjayhawkins do the problems this had have any impact on the original decision to vendor vs depend?