Skip to content
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

Add bi2012 dependency #37

Closed
wants to merge 30 commits into from
Closed

Conversation

gcattan
Copy link
Collaborator

@gcattan gcattan commented Mar 21, 2022

This PR add dependency to brain invaders 2012, so @toncho11 can use this data for the example in #36

@toncho11
Copy link
Collaborator

@qbarthelemy can you please approve this PR, so that I can continue build on top of it?

Copy link
Member

@sylvchev sylvchev left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Nice work. I have some questions regarding the mne dependency, but thank for the work! I did not know about the https://github.com/plcrodrigues/py.BI.EEG.2012-GIPSA !

@@ -2,7 +2,7 @@ sphinx-gallery
sphinx-bootstrap_theme
numpydoc
cython
mne[data]>=0.24
mne==0.20.1
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It is not a good idea to freeze the MNE version, it should really be avoided. Is this because of bi2012 dependency?

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Unfortunately yes. It is using _fetch_file which doesn t seem available in recent versions of mne:

image

I have just opened an issue in bi2012, so we can keep trace of it.

Co-authored-by: Sylvain Chevallier <sylvain.chevallier@uvsq.fr>
@gcattan
Copy link
Collaborator Author

gcattan commented Mar 24, 2022

I think there is the same problem with fetch_dataset. bi2012 contains a duplicate of the download.py file of Moabb.
We likely need to apply the same fix as in Moabb inside bi2012, and then publish a new version of the package to pip.
But it will still be a workaround, because the problem is probably common to all of these repos, so we need to brainstorm a little bit probably.

@toncho11
Copy link
Collaborator

toncho11 commented Mar 24, 2022

Yes, I think it is better to proceed with this PR and fix properly the MNE dependency later.

Copy link
Member

@qbarthelemy qbarthelemy left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks for the contribution, but the example must be finished before merging.

gcattan and others added 3 commits March 24, 2022 18:57
Copy link
Member

@qbarthelemy qbarthelemy left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

For pyRiemann-qiskit, can you validate PRs when they are ok for you? @gcattan

gcattan and others added 3 commits March 25, 2022 18:32
Co-authored-by: Quentin Barthélemy <q.barthelemy@gmail.com>
Co-authored-by: Quentin Barthélemy <q.barthelemy@gmail.com>
Co-authored-by: Quentin Barthélemy <q.barthelemy@gmail.com>
@gcattan
Copy link
Collaborator Author

gcattan commented Mar 25, 2022

@qbarthelemy Yes, of course! Thank you all for the review!

Copy link
Member

@qbarthelemy qbarthelemy left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thx @gcattan ! The ball is in your court @sylvchev.

@sylvchev
Copy link
Member

Ok, I know that Pedro is not really available for now. Maybe the best approach could be to integrate BI in MOABB and to make it an optional dependency to pyriemann-qiskit.
I understand that this PR is important for moving to the next step, but I think it will results in a dependency hell if it is merged as is. MNE is pretty standard and it could results in important regression if someone want to try qiskit and it downgrades the MNE package.

I you left me a week, I could add the some of the BI datasets to MOABB and push a new release on pypi. What do you think @gcattan ?

@gcattan
Copy link
Collaborator Author

gcattan commented Mar 25, 2022

In general, I would agree with you, but we might end up with duplicates as it is the case between bi2013 in Moabb and in Pedro's repo. We could delete or deprecate the repos but Zenodo and the technical report are pointing to them (that said, a simple notice on the readme could do the trick).

May be a short-term compromise would be to fix mne dependency inside the bi2012 repo this week and migrate all necessary repo to Moabb after we have Pedro's input on this. How does that sound?

@toncho11
Copy link
Collaborator

Actually having the Zenodo datasets on MOABB is a very good idea. This will finally create a single interface to all the EEG data we could use and test and will allow for the creation of models on a really large dataset (from many experiments/sets) which I hear is key to success. So adding BI 2012 to MOABB sounds good to me.

Then we will not longer use Pedro's code, but that is OK as long as it works.

@gcattan
Copy link
Collaborator Author

gcattan commented Mar 25, 2022

Ok, good point. @toncho11 I just saw the response from Pedro to your message. @sylvchev Let s go with Moabb then :)

@sylvchev
Copy link
Member

I opened a PR on MOABB

@gcattan
Copy link
Collaborator Author

gcattan commented Apr 6, 2022

@sylvchev @toncho11 do we still need this PR if this is going to be integrated in moabb?

@toncho11
Copy link
Collaborator

toncho11 commented Apr 6, 2022

I guess no.
But now we depend on MOABB version > 0.4.5 (currently the one in the PR).

@sylvchev
Copy link
Member

sylvchev commented Apr 6, 2022

Agree, when new MOABB version will be available on PyPi, you will close this PR.

@gcattan
Copy link
Collaborator Author

gcattan commented Apr 7, 2022

Ok, thank you :) do you know when it will be online ?

@sylvchev
Copy link
Member

sylvchev commented Apr 7, 2022

It is live; just wait a few minutes that PyPi make the release of MOABB 0.4.6 effective.

@gcattan
Copy link
Collaborator Author

gcattan commented Apr 7, 2022

Super, thank you :)

@gcattan gcattan closed this Apr 7, 2022
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants