add more hash manipulation utilities to sourmash CLI

[ctb]

yesterday, I spent some time digging into a sourmash use case with @shannonekj, and a few different reasonably generic utility script needs emerged.

the code for this is in a private repository so I'll try to describe things here - we were looking for differential presence of hashes in a genome between two samples (specifically, looking for hashes that correlated with male vs female genomes).

to do this, we needed the following new functionality -

• code to export hashes, together with their abundances, from a signature computed with track-abund; viz update sourmash sig export to export to CSV #1098
• code to intersect one track-abund signature with another signature, without flattening the abundances in the first signature. (note that sourmash sig intersect flattens all signatures) - this could be maybe be done by updating sourmash sig intersect
• code to select sequences in a FASTA/FASTQ file that have some number of overlapping hashes with a signature (this has been a repeatedly useful utility that I've implemented a dozen times in various contexts 😆 )
• code to estimate the abundance of sequences based on median hash abundance from a signature (i.e. estimate sequence abundance in a FASTA/FASTQ file using abundances from a track-abund sourmash signature) - this may be too niche to implement in sourmash directly, but I feel like it has come in handy.

I implemented all of this in a Jupyter notebook fairly easily, but it'd nice to have this in the sourmash CLI.

since code exists for all of this and I can make it available upon request, I'll label this as a good first issue...

[ctb]

I put several utilities in https://github.com/ctb/2020-emqc-scripts, which also includes the abundhist code from #933.

• contig-intersect.py does an intersection of hashes b/t two signatures, and produces a new signature from the abundances in one sig. It's useful for examining the abundances of assembled contigs.
• extract-contigs.py extracts contigs based on their median hash abundance in a signature.

These are pretty good candidates for a plugin :) #1353

[ctb]

some of these may now be possible with sig inflate added in #1889