-
Notifications
You must be signed in to change notification settings - Fork 322
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
Remove unnecessary html_theme_path
option
#573
Remove unnecessary html_theme_path
option
#573
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Seems like the PR introduced the same problem as in optimization here: qiskit-community/qiskit-optimization#475. Spell checker fails on the class names in the warning messages. I'm not sure I want to add all that class names to the dictionary.
Since #571 passed yesterday I am re-running that to see if the failure might be related to a change in a library or dependency. I tried looking at versions but the spelling checker etc seem the same versions. |
CI have run twice on this PR with the same results. |
It has to be something that changed very recently but not what was done by this PR. The sparse TorchConnector PR #571 that passed yesterday ok, fails today the same way as this PR now having just re-run it. |
Oh, I see now. Then you are right, something has been changed very recently. |
It seems to be the class names at the end of the long class paths in the deprecation messages that it didn't pull off and check before. Not sure why now. Would any alteration to the message , say putting the whole path say in singles/double quotes stop it as a alternative to adding it to spell checker - something we'd have to try I guess if we don't want to add those class names to the dict. We can choose to fix the spelling, however we do that, in a separate PR. |
I was wondering why Nature did not hit this since it has a lot of deprecations. With QuantumKernel here that is failing the class names are directly in the message (string) here versus in params of a decorator. Maybe both can go in the new name string parameter of a decorator rather than as currently done as a different option. |
At the first glance, I have not spotted the difference between two setups: when CI failed and when CI checks were successful. Looks like the library versions are identical. |
Now the failure here with the spelling in QuantumKernel, I noticed it's using a Terra decorator - that was just changed and now includes the deprecation message into the doctring which the spelling ends up seeing. So its not its presence in the source exactly I think more that it gets injected into the docstring which is checked. You can see the result in the html built: We put the full path without the :class: in the message as that will not get resolved by Sphinx - in this case it could be but that would mean parsing/altering the message which maybe is not desirable. |
Perhaps, this commit Qiskit/qiskit#9790 introduced that behavior. I will check it out later. |
Yes, Eric has been working on improving the decorators. The problem in Optimization is related but a bit different. Hopefully we can figure what the best way forward is given that it was recognized it would be very helpful for the deprecation info to be in the docs too for an end user and for it to be automated, |
Pull Request Test Coverage Report for Build 4479276456
💛 - Coveralls |
Co-authored-by: Anton Dekusar <62334182+adekusar-drl@users.noreply.github.com>
See qiskit-community/qiskit-finance#244 for the motivation.
I built the docs locally on
main
vs this branch and could find no differences between the two, as expected.