Closed timokau closed 4 years ago
Merging #155 into master will decrease coverage by
8.41%
. The diff coverage is85.71%
.
@@ Coverage Diff @@
## master #155 +/- ##
==========================================
- Coverage 57.04% 48.62% -8.42%
==========================================
Files 113 113
Lines 6560 6566 +6
==========================================
- Hits 3742 3193 -549
- Misses 2818 3373 +555
Impacted Files | Coverage Δ | |
---|---|---|
csrank/constants.py | 100.00% <ø> (ø) |
|
csrank/discretechoice/cmpnet_discrete_choice.py | 58.82% <ø> (-41.18%) |
:arrow_down: |
csrank/objectranking/cmp_net.py | 100.00% <ø> (ø) |
|
csrank/discretechoice/fate_discrete_choice.py | 45.45% <25.00%> (-54.55%) |
:arrow_down: |
csrank/discretechoice/feta_discrete_choice.py | 19.53% <25.00%> (-31.67%) |
:arrow_down: |
csrank/learner.py | 69.11% <85.71%> (+5.32%) |
:arrow_up: |
csrank/choicefunction/cmpnet_choice.py | 92.85% <100.00%> (+0.26%) |
:arrow_up: |
csrank/choicefunction/fate_choice.py | 93.93% <100.00%> (+0.39%) |
:arrow_up: |
csrank/choicefunction/feta_choice.py | 65.00% <100.00%> (+0.76%) |
:arrow_up: |
csrank/core/cmpnet_core.py | 97.61% <100.00%> (-0.09%) |
:arrow_down: |
... and 25 more |
Rebased and ready for review.
Description
Changes are on top of #154 to avoid merge conflicts.
~#154 should be reviewed first.~
Motivation and Context
The mechanism is replaced by the already established prefixing and an explicit prefix whitelist. This is more explicit than just passing around "raw" kwargs, easier to understand and satisfies the "no attributes in init" check for scikit-learn estimators.
How Has This Been Tested?
Pre-commit and test suite.
Does this close/impact existing issues?
Related to #94, #116, #146.
Types of changes
Checklist: