Currently the graphdb is built from an OTT version. In addition, it should also include all the additions that have been created after that OTT version took its snapshot of the additions repo.
This hasn't been a problem so far, since we generally only build a new database when we do an OTT upgrade, and additions are added to the db when they are created. However, should we at some point do a round of improvements to taxomachine that require constructing a new graphdb, and we are not at the same time doing an OTT release, this will matter.
Off on a branch, but https://github.com/mtholder/otcetera/blob/ws/scripts/copy_amendments.py takes a path to the amendments git shard, the last amendment SHA used in the build of OTT, and an output dir. It writes the amendments since that sha in subdirs of the output dir and writes an index.txt file in the output dir that gives you order of the subdirs that would mimic the order of commits.
Currently the graphdb is built from an OTT version. In addition, it should also include all the additions that have been created after that OTT version took its snapshot of the additions repo.
This hasn't been a problem so far, since we generally only build a new database when we do an OTT upgrade, and additions are added to the db when they are created. However, should we at some point do a round of improvements to taxomachine that require constructing a new graphdb, and we are not at the same time doing an OTT release, this will matter.
On Feb 28, @mtholder wrote on gitter: