search

choldgraf / ecogtools

A collection of tools for analyzing electrocorticography data
BSD 2-Clause "Simplified" License
17 stars 4 forks source link

setup #2

Open kingjr opened 8 years ago

kingjr commented 8 years ago

We need a proper setup.py and put the all functions inside a folder named after the library so that we can import them easily.

kingjr commented 8 years ago

A requirements.txt would also help installing the dependencies directly with git/pip

choldgraf commented 8 years ago

Hey - just resubmitted my paper yesterday (!) and have a bit more time to work on this stuff. Did we come to an agreement about the best way to host ecog-specific code? I wonder if we could make an ecog module in mne-sandbox, and then start splitting that off into submodules as code gets developed enough to be incorporated more broadly. Maybe it's a good time to do it now that we have ecog and seeg channel types now. WDYT?

choldgraf commented 8 years ago

(@kingjr pinging you in case you're not paying attention to this thread anymore)

kingjr commented 8 years ago

(@kingjr pinging you in case you're not paying attention to this thread anymore)

Oh but I am!

Yes, mne-sandbox is probably a good place; But I'm already behind my review commitments (PAC, and connectivity measures). I would recommend on making a series of as-small-as-possible PR, so that it's easier to make feedback and corrections.

choldgraf commented 8 years ago

For sure - I'm not going to do a mega-commit, just thinking where to start slowly porting over the code to MNE. The question is about freezing an API etc, because I think there are some things I want functions to do that might be too much for a function in MNE. E.g. I've already been chatting with a labmate about allowing the PAC function to calculate multiple windows of time for each trial, so you could plot a timecourse of how the PAC changes over time w/o re-filtering each time.