Upgrade packages

[pgarrison]

Started by @pgarrison and picked up by @jcass11

Purpose

Main intent:

• Upgrade package versions

Other changes to fix mypy errors:

• change an axis object input to a tuple from a list
• use color hexcodes directly as per my comment below

Testing

• ✅ run_all_manuscript_workflows.py ran without error, checked out a random set of figures and all looked correct
• ☑️ nuc_morph_analysis/lib/preprocessing/save_datasets_for_quilt.py ran without error but ran into memory issues loading and comparing outputs
• ✅ nuc_morph_analysis/lib/preprocessing/generate_main_manifest.py ran without error and compared output to quilt, only difference is the existence of the parent IDs 71552 and 71115 which is expected)
• ✅ nuc_morph_analysis/lib/preprocessing/generate_perturbation_manifest.py ran without error and compared output to quilt, both feeding control and perturbation manifests are unchanged
• ✅ nuc_morph_analysis/lib/visualization/write_data_for_colorizer.py lots of issues with memory error when not running with --noframes but looking at the output files directly things look good (colony orientation is correct in images, timepoints in correct order, quick scan of values in jsons seem appropriate)

[jcass11]

Looks like the mypy error on usings .colors on a colormap object is an issue on matplotlib's end to fix (pydata/xarray#7802). To get around it I will just replace with the hex codes directly according to the output below. Confirmed that figures result with correct coloring.

Color 0: RGB = (27, 158, 119), Hex = #1b9e77
Color 1: RGB = (217, 95, 2), Hex = #d95f02
Color 2: RGB = (117, 112, 179), Hex = #7570b3
Color 3: RGB = (231, 41, 138), Hex = #e7298a
Color 4: RGB = (102, 166, 30), Hex = #66a61e
Color 5: RGB = (230, 171, 2), Hex = #e6ab02
Color 6: RGB = (166, 118, 29), Hex = #a6761d
Color 7: RGB = (102, 102, 102), Hex = #666666

[pgarrison]

I read the matplotlib.Colormap documentation and found that Colormap.colors is not documented (perhaps only qualitative colormaps have it), but instead Colormap.__call__ is and does the same thing. I pushed a change to continue using matplotlib.Colormap and satisfy mypy.

There is a difference though: the commit I just pushed now gives all the colors an alpha channel (fully opaque).

matplotlib.colormaps.get_cmap("Dark2").colors[7] == (0.4, 0.4, 0.4)
matplotlib.colormaps.get_cmap("Dark2")(7) == (0.4, 0.4, 0.4, 1.0)

My reading of the code makes me think this shouldn't be a problem, since we're just passing the color to matplotlib, but it may be worth further attention @jcass11.