Add initial evaluation notebook to metacells module #868
Conversation
|
I can't seem to solve the issue we are seeing in CI. There is some strange interaction between matplotlib and reticulate that seems to only be coming up in conda environments, and I don't quite know what to do about it. I get various different errors depending on whether I am running with I think I am going to have to abandon the idea of doing this in an R/Python hybrid environment and instead use either a plain Jupyter notebook, or perhaps try a python-only quarto notebook. I'm leaving this up as a draft, but I will work on a different approach next. |
|
@jashapiro - I do not plan to review this while it is a draft. If you would like a review at this stage and/or when you're ready for a review, please mention me. Thank you! |
That is fine. It is quite likely that I will end up simply closing this PR. Unfortunately I don't seem to be able to get the rmarkdown/reticulate working on linux without running within RStudio. No idea what's going on, but it doesn't seem worth the effort and I will go back to just jupyter |
Here I am adding a small notebook with some initial plots to evaluate the SEACells output. The main purpose here is not so much to give a full accounting of performance, but to get the infrastructure in place for further updates.
The notebook is written in a mix of Python and R so that I can use tools from both to make plots, depending which seems more convenient. So far it seems to be working pretty well locally, but I do have some concerns about the test environment and how well this will work for others, which is part of why I am submitting it in this state.
Future updates will likely include the randomization noted in the final paragraphs, as well as additional plots of compactness and separation, as discussed in #611 a while ago!