Hi,
I'm playing around with trans.fif files, and I have found some strange
behavior. First, if you generate a quick trans.fif file with the
sample data in mne_analyze. you cannot read it in with MNE-Python
utilizing mne.read_trans . I am not sure if this problem is due to
mne_analyze creating bad fif files or an MNE-Python bug.
Here is the mne.read_trans error:
test = mne.read_trans('sample_audvis_raw-trans.fif')
Actually, I see that you're running version 0.7.1. Can you try latest
master? We might need to backport a fix, but I'm curious whether the latest
version works on your machine at least.
Eric
Dan,
there is almost a test for everything see:
https://github.com/mne-tools/mne-python/blob/master/mne/tests/test_transforms.py#L33
and buildbot is green:
https://ci.inria.fr/mne-python/job/make/
So first make sure this test pass on your machine.
If it does your file may be corrupted or we have a bug
that you should report on github.
Best,
Alex
You are using a pre-generated -trans.fif file, which is why you
couldn't replicate the problem. I am using a -trans.fif file generated
by mne_analyze. How was the -trans.fif file generated in the
mne-python sample data?
I have replicated this problem using both master and the current
stable version of mne-python. Using the current stable version of MNE
2.7.0, but I am unsure if this is an mne-python problem or an
mne_analyze problem.
As I said, I'm not sure if this is a bug in mne_analyze or mne-python.
The strange part is I know the data is in the file. If I use lower
level commands I can get access to the transformation matrix. The
question is is the format of the file wrong (missing some key fiff
parts) or is the code for read_trans wrong.
D
hi dan,
can you share the file?
the sampe trans file was generated with mne_analyze.
thanks
Alex