Run dvc repro in main without bertmesh

nsorros commented 2 years ago

Description

dvc repro was broken in main and this fixes it. Note that dvc repro dvc.yaml still fails because of generate_validation_data_xlinear.py which requires datascience installed. I wonder whether that should be off the dvc pipeline.

Checklist

[ ] Linked to Notion or GitHub issue
[ ] Added tests
[ ] Updated README
[ ] DVC up to date

Release checklist

[ ] DVC repro up to date
[ ] Models synced in S3
[ ] Version updated
[ ] CHANGELOG.md updated
[ ] Model and package version aligned

To release:

make build
make deploy

aCampello commented 2 years ago

Testing it now, doing a dvc pull and dvc status. It should fail on generate_validation_data_xlinear.py because one of the interim files changed since then

aCampello commented 2 years ago

100% agree with separating get_grants in a different dvc file at least.. In the mean time, dvc pull and dvc status looks fine.

aCampello commented 2 years ago

Issue for breaking down pipeline https://github.com/wellcometrust/grants_tagger/issues/207

wellcometrust / grants_tagger