Sorts categorical values on axis that contains only numerical values in visualization.matplotlib.plot_slice #2709
Conversation
github-actions
bot
added
the
optuna.visualization
Codecov Report
@@ Coverage Diff @@
## master #2709 +/- ##
=======================================
Coverage 91.66% 91.66%
=======================================
Files 140 140
Lines 11522 11522
=======================================
Hits 10562 10562
Misses 960 960
Continue to review full report at Codecov.
|
|
Sorry, I realised that we can make the conditions simpler. How about replacing this part elif _is_numerical(trials, param):
x_values = [x for x in x_values]
scale = "numerical"
elif _is_categorical(trials, param):
x_values = [str(x) for x in x_values]
scale = "categorical"with as follows?: elif not _is_numerical(trials, param):
x_values = [str(x) for x in x_values]
scale = "categorical" |
|
Okay, let me try these conditions and if everything is good will create another comit. |
|
Hey @nzw0301! |
|
Hi @Muktan, the pull request needs two approvals from dev-members to be merged into the master branch. So please kindly give us time. |
|
Ok got it, Thank You. |
HideakiImamura
added
the
feature
Motivation
The slice plot's axis can be improved (by sorting) when a categorical distribution suggests only int or float values. #2605
Description of the changes
Now the y axis is sorted if it only contains numerical values.
Before:
After:
Also, if y contains mix of categorical and the numerical e.g. y = trial.suggest_categorical("y", [1,"2asd","3asd","asd"])
previous and current result remains same as follows:
Can be reviewed by: @nzw0301