-
Notifications
You must be signed in to change notification settings - Fork 12
Share mwe for the annotation error #47
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
Uhm, thanks for the example, but it appears to work 😅 maybe something is needed after the call? |
Maybe it's some kind of version issue? I am using sphinx 4.2 and have tried both docutils 0.16 and 0.17.1 and it does not work for me locally. I remove the build directory and run
if I remove that line and change nothing it works, I add it again it breaks |
I'll have a look around the code later today to see the exact line and its surroundings. Maybe we can make an educated guess and at least catch the error gracefully if we can't find the source. Could you paste your Python version and pip freeze in a clean environment here? |
Here is the broken env:
It is a bit bloated because these are the requirements for both sphinx-codeautolink[dev] and the doc only dependencies for arviz (what we install for our rtd build -> matplotlib, bokeh, myst... to get extra directives and arviz dependencies to get arviz to autoexecute the code examples via those third party directives). I have created a minimal env with sphinx-codeautolink[dev] only I too was able to build the docs including the |
Oh I see, thanks for digging in! At least it's good to know that the simple case works for you too. |
Issue seems to be with bokeh somehow. Here is what I did: Create fresh environment with python 3.9. Install codeautolink[dev]. Run make html in docs folder and sucessfully generate the docs from this branch. Install bokeh and remove the build folder. Run make html and reproduce the annotations error |
So a seemingly unrelated package is changing the way numpy behaves and breaking this extension. Wow 😅 okay, I'll investigate later myself. Thanks a bunch for the debugging! |
I managed to reproduce it without Bokeh! Though the underlying cause is still a bit unclear, |
Related to #44. I tracked the error down to the presence of
np.random.<distribution>
type function calls.tox
checks passed