Add more plotting features

[mkavulich]

Synopsis

Adding new options for plotting. All of these sit under the plot: section of the config file.

  # vmin, vmax:
  # By default the color range will be scaled to the max/min values of the plotted data. To use a custom range,
  # set vmin and/or vmax
  vmin: null
  vmax: null

  # Settings for plotting political boundaries. To disable political boundaries, specify "boundaries:" with no options
  boundaries:
    detail: 0

  # Settings for plotting coastlines. To disable coastlines, specify "coastlines:" with no options
  coastlines:
    # Scale is the resolution of the plotted boundary dataset. Options are 110m, 50m, and 10m.
    scale: '10m'
    # Most standard Matplotlib arguments for shapes will work here, but I haven't figured out good documentation on which.
    # The ones listed here work, but likely a lot more customization can happen here.
    color: 'black'
    linewidth: 0.5
    facecolor: 'none'

Here is an example plot produced with the following settings on Hera:

data:
  filename: /scratch1/BMC/hmtb/kavulich/MPAS/plotting_scripts/debby/history.2024-08-08_12.00.00.nc
  var:
    - qv
  lev:
    - 15
plot:
  filename: 'boundary_tests/cropped_{fnme}_{var}_{lev}_counties.png'
  title: 'Plot of {varln}, for MPAS forecast, {date} {time}'
  latrange:
    - 20
    - 40
  lonrange:
    - -100
    - -60
  boundaries:
    detail: 2
  colorbar:
    orientation: vertical
    label: 'Units: {units}'
  vmin: 0
  colormap: "viridis"
  coastlines:
    color: 'black'
    scale: '50m'
  dpi: 300
  figheight: 4
  figwidth: 8

Type

• Enhancement (adds new functionality)

Impact

• This is a non-breaking change (existing functionality continues to work as expected)

Checklist

• I have added myself and any co-authors to the PR's Assignees list.
• I have reviewed the documentation and have made any updates necessitated by this change.

[willmayfield]

Great work Mike! Tested this and got plots.

A couple of notes:

1. I plotted data from diag files. For the gridfile I first tried the "static.nc" file, which I believe should contain all the necessary grid information. However, I got an error at the point when the script tried to open the gridfile. I think the relevant part of the error may be indicating that it fails trying to attach a time to this data? "ValueError: Failed to decode variable 'Time': unable to decode time units 'seconds since 0000-01-01 00:00:00' with 'the default calendar'. Try opening your dataset with decode_times=False or installing cftime if it is not installed."

I then used the "init.nc" file instead and it succeeded. I assume a "history" file would also suffice. So this isn't necessarily a problem, maybe just something to note in the comments that serve as documentation. I did not try the actual grid.nc kind of file.

2. I used the globbing ability when loading filenames to get a collection of forecast hours for plotting. I ended up needing to put '{date}-{time}' as part of the output filename so it didn't keep overwriting the same png file for each input file. Do you have any wuick ideas for how to handle plotting a series of forecast hours this way? As in, can the filename be "foo_{var}_lev{lev}_fhr{fhr}.png" with a reasonable configurability? It's ok to have the full datetime string for now, but it might be nice to simplify those filenames so they don't have to be "global_test1_t2m_lev0_2023-09-15_00:00:00.png".

Approving! Feel free to merge as-is or with any quick changes related to these comments.