ENH: Remove 15-char limit for FIF

[larsoner]

Closes #8312.

• Get it working
• Communicate with Brainstorm and FieldTrip folks about this (FIF file limits channel names to 15 Characters brainstorm-tools/brainstorm3#361, FIF file limits channel names to 15 Characters fieldtrip/fieldtrip#1596)
• Add constant to fiff-constants (gives Jukka a chance to make it not in the MNE_ namespace if MEGIN would use it) (ENH: Add mne_ch_name_mapping fiff-constants#29)
• Update PR with newer fiff-constants commit, check 'bads' handling/replacement
• Add test where a file is written with long names, then monkeypatch to make the tag empty on read, ensure it can still be written and read
• Add to MNE-MATLAB

[bloyl]

+1

I love this idea.

I've taken to storing ROI source waveforms in raw objects to facilitate additional processing. So I run into this often and this approach would eliminate a lot of note keeping i currently need to do myself.

Do you think it makes sense to allow users to specify the renaming map? I'm not sure if it would offer any benefit other then allowing users some control of the names as they would appear in non-mne-based readers (besa etc).

[larsoner]

I'm not sure if it would offer any benefit other then allowing users some control of the names as they would appear in non-mne-based readers (besa etc).

For this case where they are picky about the resulting shortened names, I would say that people should manually call rename_channels before writing out the data they want to use with software that does not support the tag.

[bloyl]

Any thing I can do to push this forward? I'd love to start using it.

[larsoner]

The next step I think is to make sure Brainstorm and FieldTrip people are on board. Happy to let you take the lead on that if you're up for it

[bloyl]

Is there an mne-matlab PR they could test against?

…

On Thu, Nov 12, 2020 at 2:18 PM Eric Larson ***@***.***> wrote: The next step I think is to make sure Brainstorm and FieldTrip people are on board. Happy to let you take the lead on that if you're up for it — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#8348 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ABKTXHLZE2F36VLR4QNRQSTSPQYI5ANCNFSM4SGJPP5A> .

[larsoner]

No, didn't want to bother making one until we got some buy-in, but if you think the other order makes more sense I can cook one up

[bloyl]

I'll open github issues and see what they think.

I think this is just a hidden change to fif file internals, so will be backward compatible and provided they use mne-matlab (for reading and writing) I don't think they will even notice.

[bloyl]

Seems like an mne-matlab PR would be sufficient to support both brainstorm and fieldtrip.

[larsoner]

Superseded by #8574 (hopefully/probably)