I cloned the repo and used the .devcontainer.json with VS Code to build a dev docker, on my Linux laptop.
It seemed to run well until I tried import pandas and got ModuleNotFoundError: No module named 'pandas._libs.interval'.
I tried the commands recommended here: https://stackoverflow.com/a/68935982 It didnt work.
The problem is easy to reproduce, because it also happens in Github Codespace, which, I suppose, also use .devcontainer.json.
Installation check
- [X] I have read the installation guide.
Platform
Linux-5.19.16-76051916-generic-x86_64-with-glibc2.31
Installation Method
Built from source
pandas Version
building from latest .devcontainer.json on main
Python Version
3.10.8
Installation Logs
Comment From: MarcoGorelli
hey
you need to do
python setup.py build_ext -j 4
python -m pip install -e . --no-build-isolation --no-use-pep517
https://pandas.pydata.org/docs/dev/development/contributing_environment.html
Comment From: ghost
Ha, thank you @MarcoGorelli !
Due to the structure of the documentation, I assumed that this part wasn't applicable to the Visual Studio Code paragraph.
As a Docker newbie, there is something I don't understand: why are those commands not part of the build?
Shouldn't they, if they are mandatory anyway?
Comment From: MarcoGorelli
because you need to run them regularly when the C extensions change (which is quite often), they're not just a one-off process
I agree that it's not clear though. I'll try to rewrite this page, it's been on my mind for a while, there's too many irrelevant details and the overall process isn't clear enough