Static Image Export Fails If Kaleido is Installed in a virtualenv with Spaces in the Path

[dtoniolo]

Versions
Platform: macOS Catalina 10.15.7
Python: 3.8.5
Plotly: 4.14.3
Kaleido: 0.2.0

Description of the Problem
If kaleido is installed in a virtual environment with spaces in aitspath, an error will be generated. Here's a minimal reproducible example:

import numpy as np
import plotly.graph_objects as go

a = np.arange(10)
fig = go.Figure(go.Scatter(x=a, y=a))
fig.write_image('Foo.pdf')

Running this will produce the following error trace:

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-6-c7394f7c2527> in <module>
----> 1 fig.write_image('Foo.png')

/path with venv with spaces/lib/python3.8/site-packages/plotly/basedatatypes.py in write_image(self, *args, **kwargs)
   3804         import plotly.io as pio
   3805
-> 3806         return pio.write_image(self, *args, **kwargs)
   3807
   3808     # Static helpers

/path with venv with spaces/lib/python3.8/site-packages/plotly/io/_kaleido.py in write_image(fig, file, format, scale, width, height, validate, engine)
    243     # -------------
    244     # Do this first so we don't create a file if image conversion fails
--> 245     img_data = to_image(
    246         fig,
    247         format=format,

/path with venv with spaces/lib/python3.8/site-packages/plotly/io/_kaleido.py in to_image(fig, format, width, height, scale, validate, engine)
    129     # ---------------
    130     fig_dict = validate_coerce_fig_to_dict(fig, validate)
--> 131     img_bytes = scope.transform(
    132         fig_dict, format=format, width=width, height=height, scale=scale
    133     )

/path with venv with spaces/lib/python3.8/site-packages/kaleido/scopes/plotly.py in transform(self, figure, format, width, height, scale)
    151         # Transform in using _perform_transform rather than superclass so we can access the full
    152         # response dict, including error codes.
--> 153         response = self._perform_transform(
    154             figure, format=format, width=width, height=height, scale=scale
    155         )

/path with venv with spaces/lib/python3.8/site-packages/kaleido/scopes/base.py in _perform_transform(self, data, **kwargs)
    291         """
    292         # Ensure that kaleido subprocess is running
--> 293         self._ensure_kaleido()
    294
    295         # Perform export

/path with venv with spaces/lib/python3.8/site-packages/kaleido/scopes/base.py in _ensure_kaleido(self)
    196                             self._get_decoded_std_error()
    197                         )
--> 198                         raise ValueError(message)
    199                     else:
    200                         startup_response = json.loads(startup_response_string)

ValueError: Failed to start Kaleido subprocess. Error stream:

/path with venv with spaces/lib/python3.8/site-packages/kaleido/executable/kaleido: line 4: cd: /path: No such file or directory
/path with venv with spaces/lib/python3.8/site-packages/kaleido/executable/kaleido: line 5: ./bin/kaleido: No such file or directory

[jonmmease]

Thanks for the report

[mherndon7]

I am having the same issues with the same package versions, except I am using Windows, with Anaconda. My conda environment is located in a spot with spaces ('C:\Path\with spaces") and the error I receive is:

"C:\Path\with" is not recognized as an internal or external command, operable program or batch file.

[huni1023]

i have similar issue.

i've already installed package kaleido using "pip install -U kaleido"
but it occurs error msg.

my code
fig1.to_image(f"{os.getcwd()[:-23]}/코드개발/0630_그림저장/{subject_idx}_fig1.png", engine='kaleido')

error msg

ValueError: 
Image export using the "kaleido" engine requires the kaleido package,
which can be installed using pip:
    $ pip install -U kaleido

ver. plotly: 4.12.0
kaleido: 0.2.1

[dgoppenheimer]

@huni1023, I had the same issue running kaleido on Google Colab. I ran !pip install -U kaleido, and got Successfully installed kaleido-0.2.1. But when I ran fig.write_image("fig1.png") I got the same error as you.

I solved it by running !pip uninstall kaleido (returned: WARNING: Skipping kaleido as it is not installed.) followed by

!pip install -q condacolab
import condacolab
condacolab.install()

Kernel restart,
!conda install -c conda-forge python-kaleido
Then fig.write_image("fig1.png") worked perfectly.

ver. Plotly: 5.5.0

[tran-david]

Thanks for this useful conversation guys, @dgoppenheimer solution worked for me as well.

It seems getting Kaleido installed through conda rather than pip helped resolve my issue, too.

[drew-hill]

I have what seems to be the same problem, which I may have tagged onto a less-relevant Issue here: #131 (comment)

I am using a virtual environment through "pipenv", and a recent update in homebrew pipenv and python 3.9 (3.9.10 -> 3.9.12) seems to have surfaced this bug for me. When I move the project to a location with no spaces in the dir, it works just fine, but I cannot recreate @dgoppenheimer's fix in my pipenv virtual environment (using either pipenv or pip to uninstall/install). Downgrading versions of plotly (5.6.0 -> 5.5.0) and kaleido ( 0.2.1 --> 0.2.0) did not improve things.

Has anyone else run into this and/or found a fix?

kaleido==0.2.1
pipenv==2022.3.28
plotly==5.6.0
python==3.9.12
Mac OS Monterey (12.3), Intel chip

Output of "python3 -m pip show kaleido":
Name: kaleido
Version: 0.2.1
Summary: Static image export for web-based visualization libraries with zero dependencies
Home-page: UNKNOWN
Author: Jon Mease
Author-email: jon@plotly.com
License: MIT
Location: <path to .venv>/lib/python3.9/site-packages
Requires:
Required-by:

[ZaxShen]

You can install the Kaleido globally and use the package without any virtual env.

…

On Wed, Mar 30, 2022 at 14:55 Drew Hill ***@***.***> wrote: I have what seems to be the same problem, which I may have tagged onto a less-relevant Issue here: #131 (comment) <#131 (comment)> I am using a virtual environment through "pipenv", and a recent update in homebrew pipenv and python 3.9 (3.9.10 -> 3.9.12) seems to have surfaced this bug for me. When I move the project to a location with no spaces in the dir, it works just fine, but I cannot recreate @dgoppenheimer <https://github.com/dgoppenheimer>'s fix in my pipenv virtual environment (using either pipenv or pip to uninstall/install). Downgrading versions of plotly (5.6.0 -> 5.5.0) and kaleido ( 0.2.1 --> 0.2.0) did not improve things. Has anyone else run into this and/or found a fix? — Reply to this email directly, view it on GitHub <#78 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AF3F6DBPAYVFSRGULF263HLVCP3FFANCNFSM4YTNC5IQ> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

[drew-hill]

Thanks @ZacksAmber. Unfortunately, I am working on a project that currently requires a virtual environment.

On the other hand, you did inspire me to confirm that this bug is indeed only in the virtual environment space -- running in my user environment works just fine.

[ZaxShen]

You’re welcome. Actually, I was trying to run Kaleido inside an virtual env many hours and failed. If you do need the special plot formats that can only be generated by Kaleido such as SVG, you can keep read the following solutions; If you can accept traditional image formats such as PNG, you don’t really need Kaleido — just output the image by Plotly directly. Here is my website http://zacks.cloud and http://www.zacks.cloud/projects/disaster_response_pipelines/ For section SKILL LEVELS on http://zacks.cloud 1. I pre-generated the SVG by a global Python interpreter with Kaleido since those images are rarely changed — I don’t need a real time plot here. For section OVERVIEW OF TRAINING DATASET on http://www.zacks.cloud/projects/disaster_response_pipelines/ 1. I used Python Plotly low level API (import plotly.graph_objects as go) to output plot into JSON, which was encapsulated into a docket image. Unfortunately plotly.express here cannot work since it cannot output JSON. 2. Then built an API to return the JSON generated by step 1. 3. Finally, the Flask app (my website that is in another docker image) calls the API from step 2. Then the Plotly JavaScript render the interactive plot based on the JSON. In conclusion: - If you need a web app, consider the second solution. In this case, you don’t need Kaleido since you only use the interactive plots instead of a static images; - If you do need static images, consider the first solution and you can make a improvement. Build a Docker image with minimum required Python packages such as Plotly, Kaleido, and API packages into a global Python env. This docker image running as an API that returns SVGs. Then your main project gets the SVG from that API. Hope my experience can help you.

…

On Wed, Mar 30, 2022 at 23:52 Drew Hill ***@***.***> wrote: Thanks @ZacksAmber <https://github.com/ZacksAmber>. Unfortunately, I am working on a project that currently requires a virtual environment. On the other hand, you did inspire me to confirm that this bug is indeed only in the virtual environment space -- running in my user environment works just fine. — Reply to this email directly, view it on GitHub <#78 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AF3F6DGT2USJH2T6PVL73ATVCR2E3ANCNFSM4YTNC5IQ> . You are receiving this because you were mentioned.Message ID: ***@***.***>

[portfedh]

I just wanted to add myself to this problem. I'm having the same Issue.

"name": "ValueError",
"message": "Failed to start Kaleido subprocess. Error stream:<my_path>/venv_mac/lib/python3.9/site-packages/kaleido/executable/kaleido: line 4: cd: <my_path>: No such file or directory\n<my_path>/venv_mac/lib/python3.9/site-packages/kaleido/executable/kaleido: line 5: ./bin/kaleido: No such file or directory\n"

Im going to stop using a virtual enviroment for now.

[Tammura]

Hi all, any news on this thread? Actually I got the same error after more than 2 years from @dtoniolo report!

[milesgpt]

happening to me too still.

[MadsLang]

I also ran into this issue, and for me the solution was in the title of this issue. I was using a virtual environment with poetry inside a folder with spaces (like "My folder/.venv"). The error was fixed by simply removing spaces in the folder path to something like "My_folder/.venv". Maybe it is obvious, but for me this solution is a lot better than installing plotly and kaleido in your global environment.

Running on Ubuntu with:
Poetry (version 1.8.3)
Python 3.10.12
plotly==5.22.0
kaleido==0.2.1