MNT: flake8 and datetime fixes for Travis

[zbruick]

Travis had a lot of flake8 and a few assertion errors due to changes in how Pandas handles datetimes/Timestamps. I also updated the contributor's guide to be more in line with the MetPy version to encourage the usage of the Conda environment.

Note: I closed #268 and replaced with this identical PR due to a mistake I made with git in the previous branch.

[zbruick]

Not sure what the remaining Travis errors are now.

[dopplershift]

Thanks for running these down, just some stylistic things that I think would make these changes a bit shorter.

[zbruick]

All requested changes made. Added the two commits and pushed up.

[dopplershift]

Looks good to me!

[dopplershift]

Ah, but sadly timezone is not available in Python 2.7. 😭 In the http_util, there's an existing UTC class you should be able to use instead.

[zbruick]

Have I said how much I can't wait till 2.7 gets laid to rest? I'll hunt this down and get it compatible with 2 and 3.

[zbruick]

http.util.utc worked. Rebased and pushed.

[dopplershift]

At the rate we're working on Siphon, the next release could very well be the last release we do for 2.7.

[lesserwhirls]

And we'll celebrate the heck out of that release, too.

[zbruick]

Can y'all take a look at the Travis and AppVeyor errors to see what's up? I'd like to get this cleaned up and then get #272 in early this week so that the new GLM python gallery example can be merged successfully.

[dopplershift]

Trying to re-run.

[dopplershift]

@zbruick One problem is that a previous work-around has us installing sphinx-gallery from git, which is now failing on Python 2.7. It looks like sphinx-gallery has been updated since, so can you try removing the line manually installing sphinx-gallery? (It's already listed in setup.py.)

[zbruick]

Removed the sphinx line in setup.py and pushed.

[dopplershift]

Sorry, I was unclear. The line in setup.py should stay, the line in .travis.yml doing pip install of sphinx-gallery from git needs to go.

[dopplershift]

We also have a problem where the ncss/NCSS_Example.py is failing, which I can reproduce locally. It's hitting live data, so it's possible things changed out from under us. The error:

Traceback (most recent call last):
  File "NCSS_Example.py", line 66, in <module>
    ax.plot(temp[:].squeeze(), press_vals, 'r', linewidth=2)
  File "/Users/rmay/miniconda3/envs/py37/lib/python3.7/site-packages/matplotlib/axes/_axes.py", line 1666, in plot
    lines = [*self._get_lines(*args, data=data, **kwargs)]
  File "/Users/rmay/miniconda3/envs/py37/lib/python3.7/site-packages/matplotlib/axes/_base.py", line 225, in __call__
    yield from self._plot_args(this, kwargs)
  File "/Users/rmay/miniconda3/envs/py37/lib/python3.7/site-packages/matplotlib/axes/_base.py", line 391, in _plot_args
    x, y = self._xy_from_xy(x, y)
  File "/Users/rmay/miniconda3/envs/py37/lib/python3.7/site-packages/matplotlib/axes/_base.py", line 270, in _xy_from_xy
    "have shapes {} and {}".format(x.shape, y.shape))
ValueError: x and y must have same first dimension, but have shapes (34,) and (31,)

seems like we may have been doing something done before.

[zbruick]

Oh that looks like a FV3 error - those dimensions bring back the memories. Should be able to implement the fix we used previously.

[zbruick]

The docs are failing on the IGRA2_example with a timeout error, but the ftp link is still valid and the example runs fine on 3.7 (need to try 3.6 and 2.7 which we're actually using here). Tried rerunning Travis, but that didn't fix it.

[zbruick]

Just tried 3.6 and 2.7, and was unable to recreate the timeout issue locally for that example.

[zbruick]

Trying an update to Travis to allow FTP testing (see https://docs.travis-ci.com/user/common-build-problems/#ftpsmtpother-protocol-does-not-work). Will remove if the FTP connection still doesn't work.

[zbruick]

So that didn't work. This data is also accessible by HTTP, and it's probably worth migrating over to that from FTP, as it will likely be dropped eventually and Travis recommends use of HTTP whenever possible (https://blog.travis-ci.com/2018-07-23-the-tale-of-ftp-at-travis-ci). I'll open an issue for this - in the meantime, what should we do?

[dopplershift]

This PR didn’t cause the problem, so I’ll just merge over the failures. Rewriting IGRA to use http before the next release is the real answer.