Determine minimum viable test data set
TimothyStiles opened this issue · 3 comments
Our minimum viable test set should be small enough to allow local development but complete (and probably large enough) to do some actually meaningful work. Thought is to perhaps keep just a large collection of b. sub strains or some model organism and their affiliated genes/proteins? rhea, chembl, and reactome.org should all be small enough that we can just keep their relevant data in the test set.
Right now the scraper is not sophisticated enough to just take b. Sub genomes. It simply just downloads all of genbank / uniprots data dumps.
We should have something in the range of 100-300MB. It will otherwise clog up everyone's LFS quotas when they fork it. Additionally, if we can avoid LFS, we should that way you won't have any issues if someone forgets to check-in LFS objects.
git lfs is nightmare <- Krishna
minimum viable dataset <- most interesting reaction Isaac
isaac 👇
protein -> metabolic pathway -> map against genome
checking out @rkrishnasanka's minimal viable dataset on his fork.