WIP: Deprecate EDF

[larsoner]

I added one commit on top of #5733 to show how things could pass (in theory) with stim_channel=None.

• I had to mark two BDF tests xfail to get it there. @massich if we can get rid of those xfails, then we should be able to get rid of all of the stim_channel logic in edf.py like we did everywhere else.
• One other thing is that we should probably deprecate find_edf_events, too. It shouldn't be needed anymore. Maybe we should put that in 0.17.1.
• How come we don't do self.set_annotations(read_annotations(edf_path)) in RawEDF constructor? It would save a bunch of raw.set_annotations(read_annotations(edf_path)) lines in the tests and for users. In fact I was surprised when this did not just work:

  raw = read_raw_edf(path)
  events, event_id = events_from_annotations(raw)
  

  ... because I needed the raw.set_annotations(read_annotations(path)) line in there, too.

[larsoner]

According to @cbrnr in #5738 this very well might not be the right approach.

[agramfort]

yes

raw = read_raw_edf(path) events, event_id = events_from_annotations(raw) should have worked :( you should not need to set_annotations if it was present in the file. It should just be there.

[massich]

I'll check it out

…

On Wed, Nov 21, 2018, 23:32 Alexandre Gramfort ***@***.***> wrote: > > yes raw = read_raw_edf(path) events, event_id = events_from_annotations(raw) should have worked :( you should not need to set_annotations if it was present in the file. It should just be there. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#5739 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AGt-482j5zOL-DpPPrA5b5HHYfXWQblSks5uxdR9gaJpZM4Yt46P> .