CI: Include FutureWarning from numpy in npdev build

[mroeschke]

• Tests added and passed if fixing a bug or adding a new feature

xref #46767 (comment)

cc @jbrockmendel

[mroeschke]

@jbrockmendel just to confirm, do we want to check FutureWarnings that pandas issues or numpy? I don't think I fully understand the use case here yet.

[jbrockmendel]

just to confirm, do we want to check FutureWarnings that pandas issues or numpy?

pandas

I don't think I fully understand the use case here yet.

based on test_fillna_datelike:

df = DataFrame(
            {
                "Date": [NaT, Timestamp("2014-1-1")],
                "Date2": [Timestamp("2013-1-1"), NaT],
            }
        )

>>> result = df.fillna(value={"Date": df["Date2"]})
<stdin>:1: FutureWarning: In a future version, `df.iloc[:, i] = newvals` will attempt to set the values inplace instead of always setting a new array. To retain the old behavior, use either `df[df.columns[i]] = newvals` or, if columns are non-unique, `df.isetitem(i, newvals)`

The warning issued here isn't because the user needs to update their code, but because we haven't updated our internal usage. i.e. it indicates something we need to change on our end, and shouldn't be shown to the user. So getting it should fail the CI.

[mroeschke]

Ah gotcha, that makes sense. I think failing the CI for pandas warning should happen for all builds not just the npdev one so I'll work on adding that.

[mroeschke]

I'll try to add separate PRs addressing these warnings too (I see some pyarrow ones as well). This PR may be quite large if we tackle them all in this PR. I think running pytest with -W error:::pandas should catch them.

This PR though independently should be a positive change; we will also now fail the npdev build if there are FutureWarnings from numpy