Use DEDUPLICATION_TAG to determine whether a citation node is in a docstring

[jnothman]

Fixes #177

This should handle the case of a module docstring when relabelling munged reference labels.

This probably deserves a non-regression test.

[jnothman]

@prisae says this works for him, which is a good start!

[prisae]

Thanks @jnothman, would indeed have been better to state it here.

[prisae]

Hi @jnothman. Do you know if there is any new release planned in the near future (so these changes would be available, for instance, in readthedocs.org)?

[jnothman]

Sorry, I've not given this project much attention (along with several others I have not had time to give to). This would need to be reviewed by other contributors and merged before release.

[rgommers]

I don't have plans for a release in the near future, sorry, too many other balls to juggle at the moment. Unless something breaks or gets fixed that ups the priority.

[prisae]

Thanks @jnothman and @rgommers . No problem, just thought I'd ask. Might do it again in a couple of months... 😉

[jnothman]

This has been road tested as noted above.

[prisae]

Yes, I can use it locally without issues. However, if it is released it would be possible to use it, e.g., on readthedocs.io etc. (There are probably workarounds to install directly from GitHub instead of pypi on readthedocs and friends.)

[larsoner]

@jnothman is it easy enough to add a regression test? If not perhaps we can live with the user tests here

[larsoner]

There are probably workarounds to install directly from GitHub instead of pypi on readthedocs and friends

pip will accept things like:

$ pip install https://api.github.com/repos/numpy/numpydoc/zipball/master

And such entries can live in requirements.txt files

[prisae]

Thanks @larsoner, will give that a try!

[rgommers]

Looks like we should merge this for the next release, even without a test.

@jnothman agree? can you rebase?

[jnothman]

Happy to do the rebase. I haven't reassessed how easily it can be tested.

[larsoner]

Works fine for me on MNE and SciPy docs, but neither of those showed the problem.

@prisae can you test this PR to see if it fixes your problem?

[prisae]

I did, and it works fine for me (#177 (comment)). Locally I am using it since @jnothman published it, without issues.

[larsoner]

Okay let's merge for now then since it seems to work, not break anything, and @rgommers is happy (and so am I). Thanks @jnothman

[prisae]

Great, happy to have that in, thanks everyone!