Release Management

1. In your terminal, go to src/ontology
2. execute ./run.sh make prepare_release, which will run the entire release pipeline (updateding imports, compiling patterns, exporting release artefacts such as base, merged).
3. Make sure pattern defined phenotypes are excluded from the no-edit EQ import
   1. Run: ./run.sh robot query -f csv -i wbphenotype-merged.owl --query ../sparql/terms-more-than-one-def.sparql terms-more-than-one-def.txt
   2. This creates a file src/ontology/terms-more-than-one-def.txt which contains all the classes with duplicate logical definitions.
   3. Remember, there can be multiple definitions for two reasons:
      1. You created an new EQ in a tsv and did not delete it from the wbphenotype_eq_noedit.owl file (likely reason)
      2. An EQ for the same IRI is defined in multiple patterns (say WBP_001 is in abnormal.tsv and abnormalMorphology.tsv) <- that would be a bug and needs to be remedied.
4. Some more quality control can be done using the ROBOT report feature: ./run.sh robot report --input wbphenotype-merged.owl --output report_release.tsv I would recommend to run this every time and look at the resulting tsv (report_release.tsv). Dont worry about the error message when running the command.
5. Create new branch.
6. Create new commit, but make sure you only include release related stuff in there (top level directory .owl and .obo files, imports in various sub-directories mostly).
7. Go to GitHub and create pull request, ask for reviews
8. Merge when Travis completes and reviewers happy