Hide repeats in datetimeformatter

[muendlein]

All pull requests must have an associated issue in the issue tracker. If there
isn't one, please go open an issue describing the defect, deficiency or desired
feature. You can read more about our issue and PR processes in the
wiki.

• issues: adresses [FEATURE] datetime formatting: enhance default formatting #13186
• tests added / passed
• release document entry (if new feature or API change)

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 92.17%. Comparing base (e61219b) to head (eaf6838).
Report is 8 commits behind head on branch-3.5.

❗ Current head eaf6838 differs from pull request most recent head e1ef993. Consider uploading reports for the commit e1ef993 to get more accurate results

Additional details and impacted files

@@              Coverage Diff               @@
##           branch-3.5   #13777      +/-   ##
==============================================
+ Coverage       91.57%   92.17%   +0.59%     
==============================================
  Files             326      326              
  Lines           20737    20763      +26     
==============================================
+ Hits            18990    19138     +148     
+ Misses           1747     1625     -122     

[bryevdv]

A few initial comments @muendlein but I'm also not 100% clear on this change. Please add some before/after screenshots to aid quick comparison

[muendlein]

@bryevdv now with some screenshots

formatter = DatetimeTickFormatter(
        microseconds = "%fus",
        milliseconds = "%3Nms",
        seconds = "%T",
        minsec = "%T",
        minutes = "%H:%M",
        hourmin = "%H:%M",
        hours = "%H:%M",
        days = "%b %d",
        months = "%b %Y",
        years = "%Y",
        strip_leading_zeros = ["microseconds", "milliseconds", "seconds"],
        boundary_scaling = False,
        context_which = "all",
        context = DatetimeTickFormatter(
            microseconds="%T",
            milliseconds="%T",
            seconds="%b %d, %Y",
            minsec="%b %d, %Y",
            minutes="%b %d, %Y",
            hourmin="%b %d, %Y",
            hours="%b %d, %Y",
            days="%Y",
            months="",
            years="",
            boundary_scaling=False,
            hide_repeats=True,
        ),

hide_repeats=True

hide_repeats=False

[bryevdv]

@muendlein OK now the code makes more sense, but now I am confused by the help string:

When activated only the first value of repeating formatted ticks will be shown. Context is unaffected by this setting.

This seems to be exactly opposite of what this does, since it specifically hides the context portion of the ticks in your screenshots above. Perhaps you mean something like "Context configuration is unaffected by this setting" but I don't think that's enough information to let users predict what this setting will do. We might also want to consider making the name more descriptive in some way cc @tcmetzger

[muendlein]

@bryevdv I have added a bit more details to the help string in order to make it less confusing. Note: In my example above the hiding repeats is activated for the context directly and not the parent formatter.

[bryevdv]

@muendlein can you merge/rebase latest and we can see if the windows errors go away?

[bryevdv]

In considering the help string above I realized what I don't love about this approach. Ideally, we would hide ticks we don't want to see by removing the tick locations themselves altogether. "Removing" them by just replacing them with the empty strings seems like a bit of a kludge. But we don't pathways for that kind of feedback between the formatter and the ticker, and it's probably not worth trying to implement them, so I this approach will have to do.

[muendlein]

@bryevdv There seems to be an issue with the CI build.

[bryevdv]

@muendlein Yes it seems to be affecting all PRs at the moment. I'll get this merged once it is sorted.

edit: for reference

• Build job failing due to version mismatch #13825

[bryevdv]

@muendlein the broken setuptools releases have been yanked, hopefully just restarting CI will work now 🤞

[bryevdv]

Thanks @muendlein !