"Problem with rendering languages in right-to-left direction on Plotly charts when using fig.write_image to save as a JPG

Question

"Problem with rendering languages in right-to-left direction on Plotly charts when using fig.write_image to save as a JPG

I am experiencing an issue when trying to export a chart to jpg in languages that are right-to-left direction, such as Arabic, Hebrew, and Urdu. When using the command "plotly.offline.plot" to export to .html, there is no problem and all brackets display correctly. However, when using the command "fig.write_image" to export to a jpg image, the brackets appear messy.

The type of chart I am working with is a sunburst chart.

Here is the HTML version for reference:

https://i.stack.imgur.com/hqKu7.png

And here is the JPG version showing the issue:

https://i.stack.imgur.com/HnrLm.png

The data being used includes the following entries:

الأخ (الأقدم)

بھائی (سب سے پرانا)

אח (המבוגר ביותר)

My question is whether this is a known issue or if there is a specific property that needs to be set to address it.

python plotly export jpeg

Answer 1

Answer №1

write_image() is designed to render RTL text in an LTR context, which means that mixed script strings or strings with punctuation may not display properly.

To work around this rendering issue, one solution is to wrap the strings in bidirectional formatting characters. I have implemented a custom module for this purpose, which encapsulates each string with the characters U+202B (RIGHT-TO-LEFT EMBEDDING) and U+202C (POP DIRECTIONAL FORMATTING).

An example in Sorani Kurdish:

https://github.com/enabling-languages/python-i18n/blob/main/notebooks/plotly2.ipynb

The strings are embedded within an LTR environment, while the formatting characters create an RTL environment within the LTR context.

We opted for embedding formatting characters instead of isolate formatting characters because kaleido does not handle isolate formatting characters correctly.

For instance:

def bidi_envelope(text, dir="rtl"):
    text = f"\u202A{text}\u202C" if dir.lower() =="ltr" else f"\u202B{text}\u202C"
    return text

Answer 2

write_image() is designed to render RTL text in an LTR context, which means that mixed script strings or strings with punctuation may not display properly.

To work around this rendering issue, one solution is to wrap the strings in bidirectional formatting characters. I have implemented a custom module for this purpose, which encapsulates each string with the characters U+202B (RIGHT-TO-LEFT EMBEDDING) and U+202C (POP DIRECTIONAL FORMATTING).

An example in Sorani Kurdish:

https://github.com/enabling-languages/python-i18n/blob/main/notebooks/plotly2.ipynb

The strings are embedded within an LTR environment, while the formatting characters create an RTL environment within the LTR context.

We opted for embedding formatting characters instead of isolate formatting characters because kaleido does not handle isolate formatting characters correctly.

For instance:

def bidi_envelope(text, dir="rtl"):
    text = f"\u202A{text}\u202C" if dir.lower() =="ltr" else f"\u202B{text}\u202C"
    return text

"Problem with rendering languages in right-to-left direction on Plotly charts when using fig.write_image to save as a JPG

Answer №1

Similar questions

Leveraging the power of bash to execute a python script

Having trouble locating a webpage using BeautifulSoup based on its URL

What is the method to extract div text using Python and Selenium?

Using a user-defined object as a specified parameter type in Python functions

Tips for executing a Python function from JavaScript, receiving input from an HTML text box

Learn how to retrieve values from a .json file in real-time and then perform comparisons with user input using Python

Discovering the URL of an img tag within a CSV file column containing numerous links and then cross-referencing that link with another CSV file

Kivy: Implementing a loading animation to prevent freezing during lengthy operations

Is there a way to extract a specific substring from a pandas dataframe using a provided list for filtering?

Struggling to find a specific element on a website with Xpath in Selenium using Python

Using Python and Selenium to interact with an image button element by clicking on it

Is there an alternative method to fetching contents from Django Template loader as "get_contents()" function is not implemented in it?

The error message "No module named Image" is indicating that there is a missing module in the specified file path build/exe.win-amd64-3.6/scipy/misc

Removing file extensions using Selenium in Python

Having difficulty printing a browser.find_element_by_xpath using Selenium in Python

What is the process for retrieving a player's data from the Statistics webpage using HTML code?

The spacing in the pine tree program is not formatted correctly

Once all functions have been completed, Selenium will autonomously shut down the browser

Executing Logo Button Click with Selenium in Python

What is the best way to identify the coordinates of intricate RTL-SDR signal peaks?