compression scheme (have request to not bother tar-ring one thing by itself)
hard coded version in new path (have request for stable path)
md5sum assumes it is still in EBI's file system (requested relative path from same dir)
Miraculously, the columns did not change although some contents did.
Most notably the "project full names" we used are all NULL so I have switched to the
plain shorter "project name" some new internal constants to be translated. (see: translation tables)
While in there; made an effort to make various bnodes explicit and have regular digest ids.
fix bitrot tests.
Notice:
Geno labels are now published on OLS with different label than these legacy versions.
More concerning is at least one GENO term has changed ("0000000" is now "0000719").
will update terms and labels according to OLS, (and fix the inevitable bitrot)
Fair amount of churn.
Miraculously, the columns did not change although some contents did. Most notably the "project full names" we used are all NULL so I have switched to the plain shorter "project name" some new internal constants to be translated. (see: translation tables)
While in there; made an effort to make various bnodes explicit and have regular digest ids.
fix bitrot tests.
Notice: Geno labels are now published on OLS with different label than these legacy versions. More concerning is at least one GENO term has changed ("0000000" is now "0000719"). will update terms and labels according to OLS, (and fix the inevitable bitrot)