ENH: Change logging to warning when creating a legend with no labels #27172
Conversation
Fixes #27145 Also add test_legend_nolabels_warning()
This comment was marked as outdated.
This comment was marked as outdated.
There was a problem hiding this comment.
Thank you for opening your first PR into Matplotlib!
If you have not heard from us in a week or so, please leave a new comment below and that should bring it to our attention. Most of our reviewers are volunteers and sometimes things fall through the cracks.
You can also join us on gitter for real-time discussion.
For details on testing, writing docs, and our review process, please see the developer guide
We strive to be a welcoming and open project. Please follow our Code of Conduct.
This comment was marked as duplicate.
This comment was marked as duplicate.
With modified explanation where appropriate
|
Ok, I think I'm at a place I need help from a reviewer.
|
Also: DOC: Correct Axes docstring when no labels in legend.
|
Hey is there anything that I need to do to merge this? |
lib/matplotlib/tests/test_legend.py
Outdated
| @@ -1072,7 +1073,7 @@ def test_legend_labelcolor_rcparam_markerfacecolor_short(): | |||
|
|
|||
|
|
|||
| def test_get_set_draggable(): | |||
| legend = plt.legend() | |||
| legend = plt.legend(labels=["mock data"]) | |||
There was a problem hiding this comment.
I'm somewhat uneasy about this pattern of providing labels when there is no data.
- Passing labels only is supported but discouraged.
- While currently supported, it feels, that passing more labels than we have data should possibly warn as well. At least it's non-sensical. - But that'd be a separate PR. For here, we should just not use this pattern.
I suggest to silence the warning explicitly rather than through a sketchy code construct: Decorate the function with
@pytest.mark.filterwarnings("ignore:No artists with labels found to put in legend")
instead.
Also in the following test functions.
lib/matplotlib/tests/test_legend.py
Outdated
| plt.plot([1, 2, 3]) | ||
| with warnings.catch_warnings(): | ||
| warnings.simplefilter("ignore") | ||
| plt.legend() |
There was a problem hiding this comment.
Let's at least make sure, a legend is created and attached to the Axes:
| plt.legend() | |
| plt.legend() | |
| assert plt.gca().get_legend() is not None |
There was a problem hiding this comment.
Gotcha, makes sense! In addition, is pattern you suggested above of @pytest.mark.filterwarnings(...) preferable to warnings.catch_warnings()+simplefilter pattern?
Also assert that a legend is at least created
|
Thanks, and congratulations on your first contribution to Matplotlib! We hope to see you back. 😄 |
Fixes #27145
Also add test_legend_nolabels_warning()
PR summary
After discussion in the linked issue, it was determined that
warnings.warn()was better thanlog.warning()for alerting users that we can't draw a legend if there are no labels. This PR implements the change fromlog.warning()towarnings.warn(actually,_api.warn_external())PR checklist