Hey!
I was wondering if there’s “best practice” to update Redash’s python version.
Use case:
I’ve added python as a Redash datasource (and everything works fine),
However, I need to work with Scipy version 1.9, which is supported just from python 3.8. However, Redash’s python version is set to 3.7.
I can custom build Redash with base python 3.8, which seems like the easiest solution.
However, in general, I would rather to stick with the official docker images.
Is there such image that I’ve missed?
Alternative, any other idea other then building it on my own?
I have added FROM python:3.8.5-slim and everything works fine
1 Like
I have no idea how that worked for you.
As for example, one of the required modules is pymssql which calls platform.linux_distribution() as part of its setup.
package.linux_distribution was deprecated starting with Python 3.5(?). and removed altogether for Python 3.8.
I’ll just rebuild it my self an issue a PR for the fixed docker file.
Thanks for your response.
1 Like
Yes, I recall, I’ve been updating versions of some libraries as well. In case of pymysql, it was updated to 2.1.5.
You can check locally for versions compatibility, just run all requirements.txt files in a virtual environment and “play” with them a little bit.
It is what I did and found a golden mean
1 Like