I am working on a lambda cloud jupyter notebook. When trying to load an excel file as a pandas dataframe, I’m getting the error, “ImportError: Missing optional dependency ‘openpyxl’. Use pip or conda to install openpyxl.”
I have tried to solve it by installing openpyxl within the notebook, importing it in my code, etc, but I have yet to find a solution that works.
One thing that I think might be contributing to the issue is the fact that, when installing openpyxl and pandas, they are being installed to the location “/home/ubuntu/.lambda/lib/python3.10/site-packages.” However, my working environment is in “/usr/bin/python3.” I’m not sure if this is relevant, though.
Has anyone had a similar issue? Any ideas on how to potentially fix it?
As the person at lambda who should have tested this better, I’d like to apologize.
The goal was to have Jupyter installed inside a virtual environment, while still having notebooks and interactive terminals started through Jupyter using the default namespace. On the first attempt, we fell short on that second part.
If you launch an instance now, you should find that !pip install in notebooks works again, and that running “pip list -v | grep home” on a fresh instance returns no results, meaning that the Jupyter install should not conflict with any of the dependencies of your machine learning projects. Also, no amount of "pip install"s in the base environment should prevent the Jupyter service from running as expected.
Sorry again, and @grady37 , if you would like to test this new configuration and our new ARM64 GH200 instances, I’m happy to provide you with 1 hour’s worth of cloud credits to use on GH200 instances.
Just reply with the email address associated with your lambda cloud account, or file a support ticket and mention “I’m @grady37 on the deeptak forums and Jordan said he would add an hour’s worth of GH200 credits to my account” . (so that you’re not publicly posting your email address here)