While working with the inverted repeat finder I noticed that there were some really tricky off-by-one errors. One source of this turns out to be that the operon coordinates are 0-based (since they derive from Python indices), while feature coordinates are 1-based (since they come from either BLAST or piler-cr). This pull request converts features to 0-based indices when parsing the raw gene_finder output.
While updating some tests I noticed there were many identical copies of a few so I deleted them, and fixed the formatting on a few others. Really this should've been in a separate commit in retrospect.
While working with the inverted repeat finder I noticed that there were some really tricky off-by-one errors. One source of this turns out to be that the operon coordinates are 0-based (since they derive from Python indices), while feature coordinates are 1-based (since they come from either BLAST or piler-cr). This pull request converts features to 0-based indices when parsing the raw gene_finder output.
While updating some tests I noticed there were many identical copies of a few so I deleted them, and fixed the formatting on a few others. Really this should've been in a separate commit in retrospect.