yjmantilla
commented
1 year ago @civier
Reading mne bids code better https://github.com/mne-tools/mne-bids/blob/388616f4a8b35653a21d32b17f6a0312f58b618a/mne_bids/write.py#L1585-L1622
it seems to me that with allow_preload in True (as it already is), should bypass the check and thus, try to convert any input format readable by MNE. I havent done the tests though.
Let me know if you have a dataset to test this, a MEG one would be a good extreme case to test
Originally asked by Oren:
The legal way is not straightforward but also is not difficult, is a matter of adding some boilerplate and test code to the mne-bids repository. The code I had to add to support cnt is here: https://github.com/mne-tools/mne-bids/pull/924/commits
The quick and dirty way is to just somehow pass the ALLOWED_INPUT_EXTENSIONS check:
([https://github.com/mne-tools/mne-bids/blob/388616f4a8b35653a21d32b17f6a0312f58b618a/mne_bids/write.py#L1586-L1604]
How to do this?
Well , from what I highlighted in the code, we could just modify raw.filenames[0] so that it passes it.
I have to test this though, as I just thought of this today (wish I would have thought of it sooner).
Do we have any formats in mind? I think we could test this quite fast if we had whatever files as targets to convert.
Moreover , just today I modified the way sovabids does the tests on datasets. Theoretically, now I can do tests on any file extension as long as we have a really small test file of that extension that I can replicate to make a simulated dataset.