scatterplot bug in 0.13.0

CommonClimate commented 11 months ago

Hello and first off, major thanks to @mwaskom for this incredible package. I report this bug as a developer of Pyleoclim, which has seaborn as a dependency.

In upgrading to Python 3.11, we also upgraded to seaborn 0.13.0, and this docstring example starting giving us grief. Specifically:

Traceback (most recent call last):

  Cell In[3], line 17
    gs.map_neighbors(mgs, radius=4000)

  File ~/Documents/GitHub/Pyleoclim_util/pyleoclim/core/geoseries.py:446 in map_neighbors
    fig, ax_d = mapping.scatter_map(neighborhood, fig=fig, gs_slot=gridspec_slot, hue=hue, size=size, marker=marker, projection=projection,

  File ~/Documents/GitHub/Pyleoclim_util/pyleoclim/utils/mapping.py:1205 in scatter_map
    _, ax_d = plot_scatter(df=df, x=x, y=y, hue_var=hue, size_var=size, marker_var=marker, ax_d=ax_d, proj=None, edgecolor=edgecolor,

  File ~/Documents/GitHub/Pyleoclim_util/pyleoclim/utils/mapping.py:946 in plot_scatter
    sns.scatterplot(data=hue_data, x=x, y=y, hue=hue_var, size=size_var,transform=transform, #change to transform=scatter_kwargs['transform']

  File ~/opt/miniconda3/envs/pyleo/lib/python3.11/site-packages/seaborn/relational.py:624 in scatterplot
    p.plot(ax, kwargs)

  File ~/opt/miniconda3/envs/pyleo/lib/python3.11/site-packages/seaborn/relational.py:458 in plot
    self.add_legend_data(ax, _scatter_legend_artist, kws, attrs)

  File ~/opt/miniconda3/envs/pyleo/lib/python3.11/site-packages/seaborn/_base.py:1270 in add_legend_data
    artist = func(label=label, **{"color": ".2", **common_kws, **level_kws})

  File ~/opt/miniconda3/envs/pyleo/lib/python3.11/site-packages/seaborn/utils.py:922 in _scatter_legend_artist
    if edgecolor == "face":

ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all()

The problem goes away if I revert to seaborn 0.12.2, keeping all other packages the same. My environment.yml is copied to help troubleshoot. It seems that it the fix could be a minor change in _scatter_legend_artist but I don't have a good sense of the ramifications.

Best, J.E.G.

-- name: pyleo channels:

defaults dependencies:
alabaster=0.7.12=pyhd3eb1b0_0
appnope=0.1.2=py311hecd8cb5_1001
asttokens=2.0.5=pyhd3eb1b0_0
babel=2.11.0=py311hecd8cb5_0
backcall=0.2.0=pyhd3eb1b0_0
brotli-python=1.0.9=py311hcec6c5f_7
bzip2=1.0.8=h1de35cc_0
ca-certificates=2023.08.22=hecd8cb5_0
certifi=2023.11.17=py311hecd8cb5_0
cffi=1.16.0=py311h6c40b1e_0
cloudpickle=2.2.1=py311hecd8cb5_0
colorama=0.4.6=py311hecd8cb5_0
cryptography=41.0.3=py311h30e54ef_0
debugpy=1.6.7=py311hcec6c5f_0
decorator=5.1.1=pyhd3eb1b0_0
docutils=0.18.1=py311hecd8cb5_3
executing=0.8.3=pyhd3eb1b0_0
idna=3.4=py311hecd8cb5_0
imagesize=1.4.1=py311hecd8cb5_0
ipykernel=6.25.0=py311h85bffb1_0
ipython=8.15.0=py311hecd8cb5_0
jedi=0.18.1=py311hecd8cb5_1
jinja2=3.1.2=py311hecd8cb5_0
jupyter_client=8.6.0=py311hecd8cb5_0
jupyter_core=5.5.0=py311hecd8cb5_0
libcxx=14.0.6=h9765a3e_0
libffi=3.4.4=hecd8cb5_0
libsodium=1.0.18=h1de35cc_0
markupsafe=2.1.1=py311h6c40b1e_0
matplotlib-inline=0.1.6=py311hecd8cb5_0
ncurses=6.4=hcec6c5f_0
nest-asyncio=1.5.6=py311hecd8cb5_0
openssl=3.0.12=hca72f7f_0
parso=0.8.3=pyhd3eb1b0_0
pexpect=4.8.0=pyhd3eb1b0_3
pickleshare=0.7.5=pyhd3eb1b0_1003
pip=23.3=py311hecd8cb5_0
platformdirs=3.10.0=py311hecd8cb5_0
prompt-toolkit=3.0.36=py311hecd8cb5_0
psutil=5.9.0=py311h6c40b1e_0
ptyprocess=0.7.0=pyhd3eb1b0_2
pure_eval=0.2.2=pyhd3eb1b0_0
pycparser=2.21=pyhd3eb1b0_0
pygments=2.15.1=py311hecd8cb5_1
pyopenssl=23.2.0=py311hecd8cb5_0
pysocks=1.7.1=py311hecd8cb5_0
python=3.11.5=hf27a42d_0
python-dateutil=2.8.2=pyhd3eb1b0_0
pytz=2023.3.post1=py311hecd8cb5_0
pyzmq=25.1.0=py311hcec6c5f_0
readline=8.2=hca72f7f_0
requests=2.31.0=py311hecd8cb5_0
setuptools=68.0.0=py311hecd8cb5_0
six=1.16.0=pyhd3eb1b0_1
snowballstemmer=2.2.0=pyhd3eb1b0_0
sphinx=5.0.2=py311hecd8cb5_0
sphinxcontrib-applehelp=1.0.2=pyhd3eb1b0_0
sphinxcontrib-devhelp=1.0.2=pyhd3eb1b0_0
sphinxcontrib-htmlhelp=2.0.0=pyhd3eb1b0_0
sphinxcontrib-jsmath=1.0.1=pyhd3eb1b0_0
sphinxcontrib-qthelp=1.0.3=pyhd3eb1b0_0
sphinxcontrib-serializinghtml=1.1.5=pyhd3eb1b0_0
spyder-kernels=2.4.4=py311hecd8cb5_0
sqlite=3.41.2=h6c40b1e_0
stack_data=0.2.0=pyhd3eb1b0_0
tk=8.6.12=h5d9f67b_0
tornado=6.3.3=py311h6c40b1e_0
traitlets=5.7.1=py311hecd8cb5_0
wcwidth=0.2.5=pyhd3eb1b0_0
wheel=0.41.2=py311hecd8cb5_0
wurlitzer=3.0.2=py311hecd8cb5_0
xz=5.4.2=h6c40b1e_0
zeromq=4.3.4=h23ab428_0
zlib=1.2.13=h4dc903c_0
pip:
- attrs==23.1.0
- bagit==1.8.1
- beautifulsoup4==4.12.2
- bibtexparser==1.4.1
- bleach==6.1.0
- cartopy==0.22.0
- chardet==5.2.0
- charset-normalizer==3.3.2
- comm==0.2.0
- contourpy==1.2.0
- cycler==0.12.1
- defusedxml==0.7.1
- demjson3==3.0.6
- dill==0.3.7
- doi2bib==0.4.0
- fastjsonschema==2.19.0
- fonttools==4.44.3
- future==0.18.3
- ipywidgets==8.1.1
- isodate==0.6.1
- joblib==1.3.2
- jsonschema==4.20.0
- jsonschema-specifications==2023.11.1
- jupyter-sphinx==0.4.0
- jupyterlab-pygments==0.2.2
- jupyterlab-widgets==3.0.9
- kiwisolver==1.4.5
- kneed==0.8.5
- latexcodec==2.0.1
- lipd==0.2.8.8
- llvmlite==0.41.1
- matplotlib==3.8.2
- mistune==3.0.2
- multiprocess==0.70.15
- nbclient==0.9.0
- nbconvert==7.11.0
- nbformat==5.9.2
- nbsphinx==0.9.3
- nitime==0.10.2
- numba==0.58.1
- numpy==1.23.5
- numpydoc==1.6.0
- packaging==23.2
- pandas==2.1.3
- pandocfilters==1.5.0
- pathos==0.3.1
- patsy==0.5.3
- pillow==10.1.0
- ply==3.11
- pox==0.3.3
- ppft==1.7.6.7
- pybtex==0.24.0
- pyhht==0.1.0
- pylipd==1.3.6
- pyparsing==3.1.1
- pyproj==3.6.1
- pyshp==2.3.1
- pyyaml==6.0.1
- rdflib==7.0.0
- readthedocs-sphinx-search==0.3.1
- referencing==0.31.0
- rpds-py==0.13.0
- scikit-learn==1.3.2
- scipy==1.11.3
- seaborn==0.12.2
- shapely==2.0.2
- sip==6.7.12
- soupsieve==2.5
- sphinx-copybutton==0.5.2
- sphinx-rtd-theme==1.3.0
- sphinxcontrib-jquery==4.1
- statsmodels==0.14.0
- tabulate==0.9.0
- tftb==0.1.4
- threadpoolctl==3.2.0
- tinycss2==1.2.1
- tqdm==4.66.1
- tzdata==2023.3
- unidecode==1.3.7
- urllib3==2.1.0
- webencodings==0.5.1
- wget==3.2
- widgetsnbextension==4.0.9
- xlrd==2.0.1

mwaskom commented 11 months ago

Can you please craft a reproducible example that uses seaborn directly? It's going to be necessary to understand how your library is using seaborn to make any sense of what might be happening here, and you're going to be in a much better position to do that than I am.

CommonClimate commented 11 months ago

Yes, I'll be happy to. The student who wrote that part of our code is currently sick, but when they get better I will work with them to pare down the code and provide a seaborn-only MWE.

mwaskom / seaborn

scatterplot bug in 0.13.0 #3575