[MRG] add signature cat and signature split commands to combine/split signature files

[ctb]

This PR adds a cat and a split command to the sig submodule, to concatenate and split collections of signatures. See this discussion for the original motivation.

sourmash cat

% sourmash sig cat podar-ref/*.fa.sig -o out

== This is sourmash version 3.3.2.dev9+g462bc387. ==
== Please cite Brown and Irber (2016), doi:10.21105/joss.00027. ==

loaded 192 signatures total.
saved 192 signatures

sourmash split

% sourmash sig split --outdir /tmp/foo podar-ref/1.fa.sig

== This is sourmash version 3.3.2.dev9+g462bc387. ==
== Please cite Brown and Irber (2016), doi:10.21105/joss.00027. ==

Creating --outdir /tmp/foo
writing sig to /tmp/foo/c55d50b7.k=21.scaled=1000.DNA.dup=0.1.fa.sig
writing sig to /tmp/foo/c11126d0.k=31.scaled=1000.DNA.dup=0.1.fa.sig
writing sig to /tmp/foo/e2d11fb6.k=51.scaled=1000.DNA.dup=0.1.fa.sig
loaded and split 3 signatures total.

• can we include stdin in the argument list? (punted to investigate taking stdin in sourmash sig submodule and/or load_signatures #1049)
• should we also add a sig split (or something) command to write out a bunch of md5sum-named files?
• make split work on LCA and SBT databases too
• tests
• add to docs
• add --traverse, too? (punted to add --traverse-directory to sourmash sig split and sourmash sig cat #1048)

---

• Is it mergeable?
• make test Did it pass the tests?
• make coverage Is the new code covered?
• Did it change the command-line interface? Only additions are allowed
  without a major version increment. Changing file formats also requires a
  major version number increment.
• Was a spellchecker run on the source code and documentation after
  changes were made?

[codecov]

Codecov Report

Merging #1044 into master will increase coverage by 0.45%.
The diff coverage is 95.45%.

@@            Coverage Diff             @@
##           master    #1044      +/-   ##
==========================================
+ Coverage   92.09%   92.55%   +0.45%     
==========================================
  Files          72       74       +2     
  Lines        5454     5585     +131     
==========================================
+ Hits         5023     5169     +146     
+ Misses        431      416      -15     

Impacted Files	Coverage Δ
sourmash/sig/__main__.py	92.94% <90.62%> (-0.43%)	⬇️
sourmash/cli/sig/__init__.py	100.00% <100.00%> (ø)
sourmash/cli/sig/cat.py	100.00% <100.00%> (ø)
sourmash/cli/sig/split.py	100.00% <100.00%> (ø)
sourmash/cli/signature/__init__.py	100.00% <100.00%> (ø)
sourmash/sourmash_args.py	95.81% <100.00%> (+0.81%)	⬆️
sourmash/signature.py	92.12% <0.00%> (+1.38%)	⬆️
sourmash/nodegraph.py	93.75% <0.00%> (+16.07%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 6a15878...893c0c8. Read the comment docs.

[ctb]

I think this is ready for initial review @bluegenes @taylorreiter @luizirber -- the code is a bit ugly, but the functionality is really nice and already useful!

[bluegenes]

Thinking about use cases for sig cat - most of the time, I'd like to use this when I already have some sigs for a file, and want to add a new ksize, scaled or alpha. Will it be possible to pipe signatures from compute to sig cat? Or, better, could we add a compute --append or compute --add that adds (cats) signatures into an existing signature file?

[ctb]

piping is currently in a separate issue, #1049. that's not to say we couldn't address it here but it's a chunk of extra work that I might not get to soon :)

note that compute --append wouldn't be more efficient than

sourmash compute ... -o new_sigs
sourmash cat old_sigs new_sigs -o combined_sigs

although note that the implementation supports specifying the output file as one of the input files --

# overwrite old_sigs with concatenated signatures w00t w00t
sourmash cat old_sigs new_sigs -o old_sigs

which I could add as a test...

[ctb]

added in 084879c. so, once we get stdin working, you could do

sourmash compute ... -o - | sourmash sig cat old.sigs - -o old.sigs

which would meet your needs, I think?

[bluegenes]

hadn't seen your issue on this! yep, that would definitely fit. Would def be nice to avoid writing an intermediate file

[bluegenes]

looks good! excited to use :)

[ctb]

thanks!