Various comments/suggestions

[psathyrella]

• I think the phrasing here says you don't have to run the separate phylip install command below, but afaict you still need to run it https://matsengrp.github.io/gctree/install.html#anaconda-install
• i don't see a sequence named '17' here?
• --idmapfile says it's for isotype mapping, but I found it extremely useful even without isotype mapping (didn't figure any other way to get back to my sequence names) so maybe worth changing the help msg?
• i don't understand the first part of the --idlabel help (--idlabel flag for labeling the sequence ids of the nodes in the output tree images, also write associated fasta alignment if ``True`` (default: False), is it saying that otherwise the nodes are not labelled? In either case it was the only way I could figure out to get the inferred ancestral sequences without parsing phylip output horrendousness, so if that's true maybe worth emphasizing that?
• I don't understand the details of this traceback message, but it seems to get triggered a lot, e.g. if you try rerunning without perfectly nuking everything so phylip doesn't get confused. So maybe it could be more informative that that's what's happening since it'll probably take people a while to figure out that's what's actually wrong?

          Traceback (most recent call last):
            File "/home/dralph/miniconda3/envs/gctree/bin/gctree", line 10, in <module>
              sys.exit(main())
            File "/home/dralph/miniconda3/envs/gctree/lib/python3.9/site-packages/gctree/cli.py", line 735, in main
              args.func(args)
            File "/home/dralph/miniconda3/envs/gctree/lib/python3.9/site-packages/gctree/cli.py", line 169, in infer
              *pp.parse_outfile(args.infiles[0], args.infiles[1], args.root)
            File "/home/dralph/miniconda3/envs/gctree/lib/python3.9/site-packages/gctree/phylip_parse.py", line 142, in parse_outfile
              sequence_counts = {sequences[name]: count for name, count in counts.items()}
            File "/home/dralph/miniconda3/envs/gctree/lib/python3.9/site-packages/gctree/phylip_parse.py", line 142, in <dictcomp>
              sequence_counts = {sequences[name]: count for name, count in counts.items()}
          KeyError: 'naive'

[willdumm]

Thanks for this, Duncan! The first point I just fixed this morning. I'll work on the others sometime soon.

[willdumm]

• fixed this morning. For some reason I had thought I'd made PHYLIP a dependency, but now I think it's best that it isn't, since gctree does not strictly depend on PHYLIP to run (it just needs its outputs) and a user might want to use PHYLIP compiled on their system, for example if the bioconda recipe doesn't work on their machine.
• sorry, we removed the ability to use abundances as fasta names, but missed this reference when we were updating the docs
• Both deduplicate and gctree infer accept the argument --idmapfile. For deduplicate, this argument provides a filename to write the idmap to, and for gctree infer, the argument specifies a filename from which to read this data. I think the help strings are correct, but I've tried to clarify the one for gctree infer to emphasize that it is for reading in an existing file.
• Hopefully it's better now. You're right, this is the easiest way to get a map of ancestral nodes to sequences. You could also load the pickled tree and read the sequences directly from the nodes with the gctree/ete3 api.
• I added a couple of error messages!

This is all done in #85.

[psathyrella]

wooooooo, thank you!