Choropleth_mapbox and FigureWidget - 'mapbox._derived' Value Error

[macadology]

When plotting a choropleth_mapbox using plotly express and containing it within a figure widget, I am encountering a ValueError when I drag the plot around. The error does not affect the interactivity of the plot, but it does create a huge error log in the output. I thought about trying go.add_choroplethmapbox(), but realized that the arguments are different from the plotly express version.

I have included a sample code below.

Plotly [offline]
jupyter core : 4.6.3
jupyter-notebook : 6.0.3
ipywidgets : 7.5.1
traitlets : 4.3.3
plotly : 4.8.1

import plotly.express as px
import plotly.graph_objs as go

df = px.data.election()
geojson = px.data.election_geojson()

fig = px.choropleth_mapbox(df, geojson=geojson, color="Bergeron",
                           locations="district", featureidkey="properties.district",
                           center={"lat": 45.5517, "lon": -73.7073},
                           mapbox_style="carto-positron", zoom=9)
fig.update_layout(margin={"r":0,"t":0,"l":0,"b":0})
figwid = go.FigureWidget(fig)
display(figwid)

ValueError: 
Invalid property path 'mapbox._derived' for layout

[macadology]

Added a temporary hack in basedatatypes.py to skip the exception specifically for mapbox._derived and it seems to work for now.

if not BaseFigure._is_key_path_compatible(key_path_str, self.layout):
                if key_path_str == "mapbox._derived":
                    return
                raise ValueError(
                    """
Invalid property path '{key_path_str}' for layout
""".format(
                        key_path_str=key_path_str
                    )
                )

[n1rvanana]

Hello!
Have you solved the issue?
I'm a newbee in coding, would you answer, is your temporary solution should be placed in a specific place of basedatatypes.py? I've inserted it below and got the same error.

[macadology]

When editing basedatatypes.py, look for the code snippet

if not BaseFigure._is_key_path_compatible(key_path_str, self.layout):
     raise ValueError(...

Before the error is raised, create a conditional check that skips that error if key_path_str == "mapbox._derived" as I have done above. Note that this is obviously a hack that will no longer work when you install a new version of plotly.

[cjroberts]

Hello, just to note that I also have the same issue with a scatter mapbox plot. Using add_scattermapbox does not seem to help. Here is a minimal reproducer:

import plotly.graph_objects as go
import numpy as np
lons = np.random.random(1000) * 360.0
lats = np.random.random(1000) * 180.0 - 90.0
z = np.random.random(1000) * 50.0
fig = go.FigureWidget()
fig.add_scattermapbox(lon=lons, lat=lats, marker={'color': z})
fig.update_layout({'mapbox': {'style': 'stamen-terrain'}})

[sanskarvijpuria]

I have the same issue.
The workaround by @macadology seems to be temporary.

[Jmete]

I have the same issue too.
Is there anyway to suppress the warning at least without altering the .py file mentioned above? Is there some callback we can listen for that is triggering it?

[DougRzz]

Just run this to do @macadology 's edit.

basedatatypesPath = os.path.join(os.path.dirname(os.__file__),'site-packages','plotly','basedatatypes.py')

# read basedatatypes.py
with open(basedatatypesPath, 'r') as f:
    lines = f.read()

find = 'if not BaseFigure._is_key_path_compatible(key_path_str, self.layout):'

replace = """if not BaseFigure._is_key_path_compatible(key_path_str, self.layout):
                if key_path_str == "mapbox._derived":
                    return"""

# add new text
lines = lines.replace(find,replace)

# overwrite old 'basedatatypes.py'
with open(basedatatypesPath, 'w') as f:   
    f.write(lines)

[pjireland]

I'm also having the same issue. Are there any updates on when this will be addressed?

[briochh]

Just bumping this. Seeing the same behavior for @cjroberts min example.
plotly 5.1.0
ipywidgets 7.6.3
notebook 6.4.0
Is this a bug or user error? If user error, does anyone have any information on what we are missing?

[nicolaskruchten]

No one on our development team has ever been able to reproduce this behaviour starting from a clean environment, so we don't know what to recommend. If someone can provide reproducible instructions for replicating this behaviour, I can investigate further. Some folks have reportedly achieved success by un/re-installing plotly and rebooting Jupyter.

[macadology]

Yay, glad to hear back from the development team.

I tried it again from a clean environment too and I encountered the same error. See below for my setup using conda and pip. For convenience, I didn't specify a particular version but instead rely on pip to resolve the latest version.

conda create -n myenv python=3.9
conda activate myenv
pip install plotly
pip install pandas
pip install ipython
pip install ipywidgets
pip install jupyter
pip install notebook

When running the min code in a jupyter notebook environment, the error message appears whenever I interact with the plot. Below is a screenshot of the actual error. Is it possible the development team tested the code outside of the notebook environment? Let me know if I should create a mybinder notebook to reproduce the error.

...more error messages...

[nicolaskruchten]

Thanks for this. Yes, a replicable environment like Binder would be a great help :)

[macadology]

Here you go.
https://github.com/macadology/plotly_binder
https://mybinder.org/v2/gh/macadology/plotly_binder/master?filepath=%2Fmin_code.ipynb

Once the plot is generated, moving it or interacting with it will cause error messages to appear in the output cell below the plot.

[giswqs]

I am having the same issue with Python 3.8, notebook 6.1.4, plotly 5.5.0.