nebgbhist

Tools for building annotation histories from multiple Genbank files.

Example

$ mkdir gbhist
$ neb-rev-fetch -d gbhist NC_008512
$ neb-gbhist -d gbhist | neb-diff-apply NC_008512.git
$ git --git-dir=NC_008512.git gc --aggressive
$ neb-validate-history -p NC_008512.git -g gbhist/2009-04-29-04-04.gbk

Playing with files

If you clone the pack repository after building you can poke around at the contents on the file system. For larger genomes with lots of edits this may run afoul of directory entry limits until I rewrite the object storage. For the Carsonella (NC_008512) example I use there isn't an issue.

$ git clone NC_008512.git
$ cd NC_008512
$ ls -1
accession  dblink     keywords   objects/   seq/       version
comment    definition locus      ref/       source

Files in the root directory represent keywords defined in the Genbank file. The ref and seq subdirectories contain information about references and the sequence hashes. The objects directory lists the features in the history. The files objects/known and objects/alive are dictionaries of hashes that point toward the current version of a feature at a given location. Active features are anything that is not deleted or replaced with a new version.

I recommend that you work with nebpack.Pack objects if you want to actually do anything with the objects.