Closed jmuhlich closed 3 months ago
From the rnaseq repo it looks like we should have the input parameters in the test-dataset repo so we can use URLs in test.config and related files. Is this correct or should I just try to use the files from our assets directory?
Should we do multiple test profiles like:
input_cycle
and marker_sheet
parameters.We may want to improve the quality of the sample and marker sheets before we put them in the test data repo since I just put in garbage values to have something to run.
I'm already using URLs for the 1 row per sample per cycle samplesheets, but I still have full paths in the ones for 1 row per sample because they require an image_directory. I guess we could specify an S3 bucket 'directory'? Is there a better way to do this?
I hacked something together in the recent PR that was just merged to dev. It uses a local sample sheet that's committed to the mcmicro repo, but I reference it via its raw.githubusercontent.com
URL. I think the correct long-term solution is to commit the test samplesheets and marker sheets to the (as yet nonexistent) mcmicro branch in the test-datasets repo and reference them by https URL from there.
Since #26 also introduces a minimally functional test
profile and our CI now requires it to run successfully, I'll close this. We can improve the test profile setup with a more expansive sample sheet as time goes on.
Description of feature
The
test
profile should run our small tonsil test image data through as much of the pipeline as we have working.