WIP: preprocessing module #40
Conversation
|
Some files are missing I think |
|
doh.. silly me, the files are here now |
| @@ -116,10 +38,10 @@ def compute_proj_ecg(in_fif_fname, tmin, tmax, n_grad, n_mag, n_eeg, l_freq, | |||
| default=2) | |||
| parser.add_option("--l-freq", dest="l_freq", | |||
| help="Filter low cut-off frequency in Hz", | |||
| default=None) # XXX | |||
| default=5) | |||
There was a problem hiding this comment.
true, changed it to 1 Hz
|
for maxfiler I would go for mne_maxfilter.py with resonable default parameters |
|
@agramfort have a look, I added mne_maxfilter and EOG |
| mne.preprocessing.compute_proj_ecg(raw_in, tmin, tmax, | ||
| n_grad, n_mag, n_eeg, l_freq, h_freq, average, preload, | ||
| filter_length, n_jobs, ch_name, reject, bads, | ||
| avg_ref, include_existing, ecg_proj_fname, ecg_event_fname) |
There was a problem hiding this comment.
I think raw_in should be an instance of mne.fiff.Raw and the function should return a list of Projection
the save would be done here with mne.write_proj
There was a problem hiding this comment.
The reason why I didn't use mne.fiff.Raw as input is that the function modifies the data, if someone uses compute_proj_ecg directly from python this could be a problem.
I can save the events and proj here, no problem.
There was a problem hiding this comment.
The reason why I didn't use mne.fiff.Raw as input is that the function modifies the data, if someone uses compute_proj_ecg directly from python this could be a problem.
ok but you may want to do something on the raw before using
compute_proj_ecg like filtering, marking bad channels etc. I don't
think modifying the data is a problem if it's clear in the doc string
and we could have a copy parameter as in numpy
I can save the events and proj here, no problem.
good
|
we should have a test_ssp.py that tests the compute_proj_ecg and compute_proj_eog functions the artifacts module should be integrated into the preprocessing module |
|
tests are important to keep a good code coverage :) that's it for tonight ;) |
|
@agramfort: have a look, I think this covers everything. |
|
looks great ! thanks I'll try the script on a couple of datasets tomorrow and I'll merge |
…CG proj in mne_compute_proj_eog.py + adding EXG name to SSPs
Preproc refact cosmit
@agramfort: first rework for the preprocessing module. Let me know if you think things should be done differently.