[MRG] Speed up copy_local_genomes.py with symbolic links

[mr-eyes]

genome_grist.copy_local_genomes takes forever to copy and compress FASTA files. This PR resolves #138 by adding the command line option --sym to use symbolic links. --sym defaults to False, might be changed later after discussing the PR.

Also upgrading private genomes config file to the latest version in docs (samples instead of sample).

[mr-eyes]

@ctb symlinks are working. I suspect the issue you observed was symlinking relative paths, not absolute paths, maybe?

The problem now after this PR is that the Snakefile always expects gzipped files to process {{ident}}_genomic.fna.gz, which might not exist if symlinking uncompressed Fasta files. Does genome-grist really need the Fasta files to be compressed?

[ctb]

cool, thanks!

the sticking point for the compression are the mpileup_wc and build_new_consensus_wc rules, which need to make the files available to bcftools for alignment/pileup purposes. All of the custom code we've written doesn't care about compression or format (because it uses screed). So that could be fixed, but I couldn't think of a simple solution.

Anyhoo, rather than fix the code to figure out if the file is gzipped and then do the right thing, I just decided that compressing files is good and we should be doing it in the first place :)

[mr-eyes]

Perfect! The PR is now ready for the review :) Symlinking will only work now if the Fasta files are already compressed.

[mr-eyes]

Observed behavior: Using symlinks will work, but the original filename will show in prefetch results, not the new one in the symlink.

[ctb]

if it's just in the prefetch CSV filename field, that's probably fine! those paths aren't used for anything.

could you separate out the doc change into its own PR, or do you mind if I do? I think that can/should be merged quickly but I might not get to the symlink PR for a bit.