Closed rth closed 3 years ago
Merging #272 (1da13b6) into master (e12cf36) will increase coverage by
0.20%
. The diff coverage is100.00%
.
@@ Coverage Diff @@
## master #272 +/- ##
==========================================
+ Coverage 82.58% 82.78% +0.20%
==========================================
Files 138 137 -1
Lines 5059 4677 -382
==========================================
- Hits 4178 3872 -306
+ Misses 881 805 -76
Impacted Files | Coverage Δ | |
---|---|---|
rampwf/utils/__init__.py | 100.00% <100.00%> (ø) |
|
rampwf/utils/command_line.py | 42.42% <100.00%> (-14.98%) |
:arrow_down: |
Continue to review full report at Codecov.
Legend - Click here to learn more
Δ = absolute <relative> (impact)
,ø = not affected
,? = missing data
Powered by Codecov. Last update e12cf36...1da13b6. Read the comment docs.
Great, thanks for the confirmation!
Don't we use ramp_convert_notebook when we set up the event (to convert the notebook into html and display it on the frontend)? Just a neuron firing from distant memory.
As far as I can tell in ramp-board we are now directly using jupyter nbcovert
to convert to HTML rather than calling this script.
This is a backport of part of https://github.com/paris-saclay-cds/ramp-workflow/pull/247 (currently in advanced) to master.
Which removes following rampwf.utils.command_line endpoints,
ramp_test_submission
(in favor oframp-test
)ramp_test_notebook
ramp_convert_notebook
ramp_leaderboard
except for
ramp_test_submission
I imagine are somewhat redundant with rampwf.utils.cli endpoints or not useful? As far as I can tell they are never used in ramp kits or ramp-board?One concern is that there are deployed ramp events that suggest to use
ramp_test_submission
instead oframp-test
so we should probably not remove it but rather deprecate (or expose it as an alias toramp-test
if the API is close enough)For the other 3 endpoints, I'm not sure if we need a deprecation cycle or removing is fine. I imagine users shouldn't have been using those?
@agramfort @glemaitre Please let me know if you have any concern about this. I haven't followed the historical reasons for having both
ramp-test
andramp_test_submission
.cc @albertcthomas
TODO: