Catalog strptime

[deeplycloudy]

Some datasets, including various GOES datasets on the Unidata THREDDS catalog, don't have a time dimension, but that time dimension can be inferred from the filename. However, those GOES filenames have julian days in the timestamps, making it not possible to trivially parse them with the current regex approach in DatasetCollection.

This PR adds an additional regex tag strptime and an additional argument for supplying the datetime.strptime format string. This functionality increases the number of filenames from which time information can be harvested .

[zbruick]

Looks like there are some flake8 issues and we probably need a test added for this as well.

[deeplycloudy]

It looks like there is no existing test for regex= with filter_time_*, so do we really need two tests — one for the old behavior, and one for the new behavior with strptime?

[zbruick]

Good catch - yeah that would probably be best.

[deeplycloudy]

I added two tests: a standalone regex and regex with strptime. I don't understand the flake8 errors on travis, though I gather from above there may be some big-picture things driving those errors, and not this PR?

[zbruick]

Just some flake8 issues here. @dopplershift, do you know what the rest of the pydocstyle issues stem from? Any other feedback you have on this?

[dopplershift]

It's been a couple weeks of annoying package incompatibilities. pydocstyle released 4.0, which breaks flake8-docstrings. 😭 I'll push a commit to your branch that should fix it.

[dopplershift]

Well, that fixed the errors about pydocstyle.

[zbruick]

Looks like there's just a couple lint issues left to deal with on this and then it's good to go!

[deeplycloudy]

I think the remaining lint issues are handled. Looks like the only remaining Travis problem is a connection error in some part of the code not related to this PR?

[zbruick]

Thanks for your work on this! So the remaining issue is documented in #273 and doesn't have anything to do with this PR. I think this is good to merge, but I'll defer to @dopplershift.

[dopplershift]

Thanks @deeplycloudy !

[deeplycloudy]

My pleasure, @dopplershift and my thanks to you and @zbruick for a pleasant PR process!