Binaries checked into repo

[jeffkaufman]

There are a few binary files checked into the repo. This is not best practice, since it slows down git and diffs don't work well. I see:

• ref/contaminants.fasta.gz
  • this is by far the biggest, and the only one that's really a problem
• readme-workflow-diagram.png
  • I don't know how the diagram is made, but if it has a source much better to check that in
  • The diagram itself could be served from data.securebio.org
• modules/local/countAlignmentDuplicates/resources/usr/bin/find_duplicates
  • also modules/local/countAlignmentDuplicates/find_duplicates/target/release/find_duplicates
  • you generally shouldn't check out
• test-data/gold-standard-results/*.tsv.gz and test-data/toy-data/*.gz
  • better to have as raw tsv instead of compressing so that deltas can be reviewed
  • if you're testing code that consumes gzipped files, add a little wrapper that gzips as part of the test
  • these aren't big