HeaderDiff doesn't recognize ignore_keywords for keyword_count

Astropy 4.2.1

Consider the following:

>>> h1 = fits.Header()
>>> h2 = fits.Header()
>>> h1['foo'] = 'bar'
>>> h2['foo'] = 'bar'
>>> h1['foo2'] = 'baz'
>>> diff = fits.HeaderDiff(h1, h2, ignore_keywords=['foo2'])
>>> diff.identical
False
>>> diff..diff_keyword_count
(2, 1)

Based on the documentation for HeaderDiff I would have expect these to be found to be identical. As shown above they are not identical because they have a different number of keywords, even though I’m trying to ignore one of them.

I did a quick check of some of the other *Diff methods and they seem to all suffer from the same issue.

Thanks.

[embray]

I could go digging deeper, but my suspicion is that ignore_keywords just isn’t taken into account in all cases. I think it’s probably used more when comparing just the values of individual keywords, but not the presence, or lack thereof, of some keyword.

I think your interpretation of how the option should work is valid, and patches would be welcome (though it might lead to breaking some things in some pipelines that have relied somehow on the existing behavior, but I’m open to only worrying about that if the case arises…)

[cmarmo]

Referencing here #5757 and #7763 for useful discussions.