Closed tamidodo closed 9 months ago
Hi Tammy, could you post the error you're getting?
As far as I'm aware, there are not supposed to be any issues with any pandas version above 1.0.5
.
So if I pip install pandas 2.0.2 it throws that error about plotly-calplot requiring pandas<2.0.0 and if I pip install plotly-calplot after pandas, it uninstalls the version of pandas I have and installs it's own version of pandas (1.5.3):
@tamidodo thanks so much, I wasn't aware of it. I'll find a fix. Meanwhile I'd suggest you to use conda, poetry or venv so that you can install plotly_calplot and use pandas < 2.0.0 in a isolated environment (without uninstalling pandas 2.0.2 from your default env).
I'm a fan of poetry as it handles lib versioning, the commands would look like
pip install poetry
poetry init
config for your project and a pyproject.toml will be created
poetry add plotly-calplot pandas
by default it uses pypi as a host, but if your company has a custom lib repo you can change it
Hi, @tamidodo
Did you try installing the libs in the inverse order? first plotly-calplot then pandas?
I've created a notebook showcasing it. Here's a gist for it.
To be clear, I have ways of getting around the installation and automatic uninstalling, I'm mostly just adding this issue to document it in case it's something that can be fixed in the package because as a user, you have to know about the issue to implement workarounds which just seemed a little janky.
+1 on updating to pandas 2 FWIW, I think this will get the job done (didn't create a PR; didn't build, didn't bump version, etc) https://github.com/AuDigent/plotly-calplot
To be clear, I have ways of getting around the installation and automatic uninstalling, I'm mostly just adding this issue to document it in case it's something that can be fixed in the package because as a user, you have to know about the issue to implement workarounds which just seemed a little janky.
How did you get around the issue? The rest of my app is dependent on Pandas 2 so I don't want to go downgrade.
Is there any way this could be updated to be compatible with newer releases of pandas?