Expand test suite

[casblaauw]

Description of feature

As always, the suite of tests could be made more comprehensive. I'm currently mostly thinking about the changes I made in recent PRs and that the functionality they change should be thoroughly covered by tests:

• Add chrom checks, manual file lists, and stored target_start/end to read_bigwig/bed  #153: reading in bigwigs with bad chromosomes in your bed file or chromsizes/genome should always raise an error (and conversely, reading in proper files should return an anndata with values)
• Add chrom checks, manual file lists, and stored target_start/end to read_bigwig/bed  #153: reading in files with a list or dict in import_beds/import_bigwigs should be tested
• Add chrom checks, manual file lists, and stored target_start/end to read_bigwig/bed  #153: negative values in bigwig should raise a warning
• Allow single string model_name for plotting functions #151: plotting functions should also work with a single string as model name
• Add serialisation decorators to custom attention classes #149: all models in the zoo should load
• Small bug/convenience improvements #152: (maybe) test that positive and negative strand regions are consistently predicted
• Add extra input checks to tl.contribution_scores  #163: different cases of contribution scores (index, None, empty list)

And probably more things that currently aren't covered, but I haven't looked at the test suite in a while (which is exactly why I'm making this as a reminder)