Closed nsheff closed 6 months ago
Do you have an ETA on ubiquerg 0.8.0? Sometime in the next week?
can be anytime. you can release it if it helps you. I don't have anything immediately planned to add to it.
Attention: Patch coverage is 95.12195%
with 2 lines
in your changes are missing coverage. Please review.
Project coverage is 55.12%. Comparing base (
4daf0c6
) to head (a091ecb
). Report is 6 commits behind head on dev.:exclamation: Current head a091ecb differs from pull request most recent head 3e00e41. Consider uploading reports for the commit 3e00e41 to get more accurate results
Files | Patch % | Lines |
---|---|---|
pipestat/parsed_schema.py | 94.11% | 1 Missing :warning: |
pipestat/pipestat.py | 92.85% | 1 Missing :warning: |
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
mk_abs_via_cfg
contains the additional step of using os.makedirs
to create the full path if it does not already exist. This is important for the file backend where the user could specify a results file that is nested into subfolders that don't exist yet. I added a commit that moved this functionality into a small separate func so that mkabs
can be used everywhere and then we can create the subdirectories to the file path only when absolutely needed.
I've also refactored read_yaml_data with yacman.load_yaml.
I will go ahead and merge these changes to dev, adding the action item to update the documentation to the original issue in #160
This allows the schemas to be remote. I did this by using
yacman.load_yaml
, instead of an internalread_yaml_data
which did not handle URLs. I also replaced an internalmk_abs_via_cfg
function call withubiquerg.mkabs
--mk_abs_via_cfg
use bymkabs
, and then remove that function?read_yaml_data
with theyacman.load_yaml
? This would require a bit of refactoring because the internalread_yaml_data
was not just reading yaml, but was used for a few other things as well. In general I don't like that entanglement, and it would reduce maintenance to just useyacman.load_yaml
.One nice thing about this is that it will make it so much easier to get started using pipestat. Our tutorials can just gloss over the schema complexity and point to some general-purposes, training schemas, so then the user doesn't have to create one locally to just get started quickly.
TODO: