Closed dlaehnemann closed 10 months ago
One further idea for this test: We could run this in parallel to all the other tests, to save overall runtime of the tests (and maybe speed up being able to merge stuff). To avoid running too much stuff while there are still blatant errors on a pull request, I would split the current Testing
into three different sections:
Dryruns
section for a quick number of --dryrun
tests (without doing the Free Disk Space (Ubuntu)
step), to catch the most blatant errors.Short Tests
section, that needs: Dryruns
to have completed successfully.Long Test
section, that also needs: Dryruns
to have completed succesfully.And I would add a mechanism to stop all other tests if any of them fails. The snakemake
test suite seems to have this, but I have not configured something like this in the past.
One downside is that both 2.
and 3.
will have to probably do the `Free Disk Space (Ubuntu) step (although we could try without them) and they will both have to separately set up all the conda environments...
This improved test case, with documentation on the data used in the
.test/config-sra/README.md
file, should be configured to run through in a reasonable amount of time. It is set up to also test the case where one sample is used in multiple groups.