feat: add plot tornado function and state mappers from spot check

[merrielle]

This code is adding the most useful bits from the most recent spot check notebooks. It has plotting code to gracefully show a horizontal bar graph with both positive and negative values. It also adds some dictionaries to help map state abbreviations to state names and also loadzones to states.
EDIT: The dictionaries have been moved to PreREISE.

Ideally the PR should be quick and easy. :)

[kasparm]

The following test fails for me. Maybe I am doing something wrong:

postreise/plot/multi/tests/test_plot_pie.py ....                                                                                             [ 72%]
postreise/plot/multi/tests/test_plot_shortfall.py .........                                                                                  [ 83%]
postreise/plot/tests/test_plot_shadowprice.py .....FFFF                                                                                      [ 94%]
postreise/process/test/test_transfer.py ..                                                                                                   [ 96%]
postreise/tests/test_mocks.py ...                                                                                                            [100%]

===================================================================== FAILURES =====================================================================
_______________________________________________________ test_get_bus_legend_bars_and_labels ________________________________________________________

    def test_get_bus_legend_bars_and_labels():
        mock_split_points = [-1, 1, 2, 3, 4]
>       bars, labels = _get_bus_legend_bars_and_labels(mock_split_points, [''])
E       ValueError: too many values to unpack (expected 2)

postreise/plot/tests/test_plot_shadowprice.py:113: ValueError
________________________________________________ test_get_bus_legend_bars_and_labels_with_rounding _________________________________________________

    def test_get_bus_legend_bars_and_labels_with_rounding():
        mock_split_points = [-1, 1, 1.9999, 3, 4]
>       bars, labels = _get_bus_legend_bars_and_labels(mock_split_points, [''])
E       ValueError: too many values to unpack (expected 2)

postreise/plot/tests/test_plot_shadowprice.py:121: ValueError
_________________________________________ test_get_bus_legend_bars_and_labels_clamps_large_numbers_on_end __________________________________________

    def test_get_bus_legend_bars_and_labels_clamps_large_numbers_on_end():
        mock_split_points = [-500, 100, 200, 300, 400]
>       bars, labels = _get_bus_legend_bars_and_labels(mock_split_points, [''])
E       ValueError: too many values to unpack (expected 2)

postreise/plot/tests/test_plot_shadowprice.py:129: ValueError
______________________________________________ test_get_bus_legend_bars_and_labels_min_bar_len_is_one ______________________________________________

    def test_get_bus_legend_bars_and_labels_min_bar_len_is_one():
        mock_split_points = [-.9, .1, .2, .3, .4]
>       bars, labels = _get_bus_legend_bars_and_labels(mock_split_points, [''])
E       ValueError: too many values to unpack (expected 2)

postreise/plot/tests/test_plot_shadowprice.py:137: ValueError
======================================================= 4 failed, 80 passed in 4.82 seconds =======

[merrielle]

@kasparm Ben noticed that as well so I'm addressing it here #68

[victoriahunt]

Is there a demo that shows how to use this? If not, I think would be helpful to add a quick demo.
Code for tornado plot working for me as it did above, except that in the above it has "zones" as label on the y axis, and there's no y axis (category) label currently. I think that's fine, but it was a difference. I noted that label is currently set to 'not visible'.

[victoriahunt]

the 3 tests in postreise/tests passed for me, but I also got deprecation warnings. I don't think these have to do with the tornado code at all, but I'm going to record them here for our collective memory:

[victoriahunt]

4 of the 9 test in postreise/plot/tests failed for me. I'm still learning about testing - maybe it's an error on my side?

In summary, the tornado plot code seemed to work well for me and looked good in review, but these tests are not all passing.

tests_postreise_02_03_2020.pdf

[danielolsen]

@victoriahunt I ran into the ipywidgets warning previously, and solved it by doing a pip upgrade on that package.

[merrielle]

@victoriahunt I added a demo notebook! I fixed the broken tests in this pr #68. It's already been merged so if you update PostREISE the tests in plot should all work. :)

[victoriahunt]

All tests passed. Heads up that I did get 3 warnings on tests which were all the same issue and read as follows: DeprecationWarning: Using or importing the ABCs from 'collections' instead of from 'collections.abc' is deprecated, and in 3.8 it will stop working from collections import Mapping. Thanks for adding the demo.

[victoriahunt]

Tornado plot will be useful and the demo is helpful in showing the basics of how it works. All tests are passing.