Default subplots to time difference from first sample

[hyanwong]

But also allow calendar date, if required.

This makes it much easier to compare plots from ARGs with different endpoints. We need to make sure that we run dating with e.g. constr_iterations = 500 for the dates to be sensible, though.

[jeromekelleher]

Looks good, but looks like tests are failing?

[hyanwong]

Looks good, but looks like tests are failing?

Oops, sorry. A bad edit snuck in somehow. Corrected now, and tests are passing, thanks for pointing it out.

Incidentally, one reason I didn't spot it it that I can't run tests locally: various assertion errors happen, which I presume is to do with not having the right setup files, e.g.

tests/test_cli.py::TestMatch::test_single_defaults - AssertionError: assert 4 == 5
FAILED tests/test_cli.py::TestExtend::test_include_samples - KeyError: 'num_samples_processed'
FAILED tests/test_cli.py::TestInfoMatches::test_defaults - KeyError: 'hmm_cost'
FAILED tests/test_inference.py::TestRecombinantHandling::test_get_recombinant_strains_ex1 - AssertionError: assert {np.int64(56)...example_1_1']} == {55: ['recomb...example_1_1']}
FAILED tests/test_inference.py::TestRealData::test_first_day - KeyError: 'num_samples_processed'
FAILED tests/test_inference.py::TestRealData::test_2020_01_25 - KeyError: 'exact_matches'

[jeromekelleher]

Ah yes, you need to clear the data cache in tests/data/cache and let it rebuild (next time you run the tests)