ci: 👷 update numpy dependency to a simplified version constraint

[onuralpszr]

Description

Please include a summary of the change and which issue is fixed or implemented. Please also include relevant motivation and context (e.g. links, docs, tickets etc.).

List any dependencies that are required for this change.

Type of change

Please delete options that are not relevant.

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• This change requires a documentation update

How has this change been tested, please provide a testcase or example of how you tested the change?

YOUR_ANSWER

Any specific deployment considerations

For example, documentation changes, usability, usage/costs, secrets, etc.

Docs

• Docs updated? What were the changes:

[onuralpszr]

@LinasKo this change for helping inference to bump supervision version to latest and based on feedback I got @PawelPeczek-Roboflow asked me to relax numpy version for py38 and py39 and with this change it can install higher version than previously.

Side note : I also remove pandas-stubs from metrics and based on tests all fine I am moving stubs under other typing package for mypy.

[LinasKo]

1. If we're removing stubbs, we can remove the line entirely, rather than commenting it out.
2. What do you mean by a separate package for types? supervision-stubs, that we'd publish to PyPi?
3. To continue using pandas after calling metric_result.as_pandas(), you'd likely need to import pandas as pd. That line prompts installing stubbs. Would it not be more clear to the user if we installed it together with pandas? Or is that not a problem you see / not the ideal solution?

[onuralpszr]

1. If we're removing stubbs, we can remove the line entirely, rather than commenting it out.
2. What do you mean by a separate package for types? supervision-stubs, that we'd publish to PyPi?
3. To continue using pandas after calling metric_result.as_pandas(), you'd likely need to import pandas as pd. That line prompts installing stubbs. Would it not be more clear to the user if we installed it together with pandas? Or is that not a problem you see / not the ideal solution?

1 - ) I can remove entirely for sure +1
2 - ) I mean we can move other typing mypy packages for "dev" we don't publish to pypi at all
3 - ) Pandas itself enough to have it. Mypy stub package not needed. (I tested metrics and didn't see any requirement or error at all for pandas-stubs, We only need "pandas". Stubs for "mypy" only)

[LinasKo]

3. Okay. I've checked it, and the messages I see are mypy-specific. Sure - let's remove it for now. If we go the mypy route, we should offer something equivalent or offer stubs of our own, that also install pandas stubs.

I say this because as a user of mypy, having just installed supervision[metrics] to use pandas, seeing that I need to install more things is annoying.