Fix builds and tests

[fedarko]

See #563.

• the main build is broken due to -- it looks like -- Q2 2020.6 not being installable any more, due to seaborn version stuff?
• the standalone python 3.7 build is broken due to using a now-old version of skbio (?)

I think the general approach will be updating the versions of the Python packages used (and making sure that this doesn't break empress). Per https://devguide.python.org/versions/, Python 3.7 and below are not even supported any more, so if getting those versions running is a pain we can probably just remove support for them (and the associated old Q2 versions that fix old Python versions; the most recently released? Q2 environment is already on Python 3.9).

[fedarko]

From trying to get the Python test suite running in Q2 2024.10, it looks like there are some additional issues:

• Using nosetests causes some errors -- it looks like this is largely because nose is no longer maintained. From some testing, it looks like switching to pytest is easy to do as a drop-in replacement (this would also make it easy to profile python code coverage using the --cov flag and pytest-cov, to satisfy Compute test coverage and show in README #156)

• Using an old version of iow causes problems due to Prevent the single-node tree check from yielding "false positives" on linear trees; also fix the build improved-octo-waddle#48. Looks like pinning iow >= 1.0.8 fixes this, although presumably this pin is only required for testing (assuming that ordinary use of Empress won't involve the visualization of these "linear" trees)

• pandas.DataFrame.append() no longer exists in recent versions of pandas. It looks like ._append() can be used as a hacky workaround for the purpose of testing (https://stackoverflow.com/a/76449334), but a better long-term solution would be using .concat() or something instead.

• pandas.util.testing.assert_frame_equal() has been changed to pandas.testing.assert_frame_equal() (without the .util.)

• This test fails, because at one point tools.match_inputs() uses a set as the index of a DataFrame, and this is now no longer allowed:

empress/empress/tools.py

Lines 290 to 294 in cc3975a

	padded_metadata = pd.DataFrame(
	index=samples_without_metadata,
	columns=sample_metadata.columns,
	dtype=str
	)

(Per https://stackoverflow.com/a/73778792, we can just convert the set to a list before creating this DataFrame.)