kylemath / mne-python Goto Github PK

Need to implement phase unwrapping from p-pod.

Seems to be located in b_norm.m, lines 57-67 (fixing phase wrap).

Seems straight-forward enough to implement.

Need to test that output is the same between matlab and python. Should be simple enough since p-pod seems to just add or subtract 360, depending on phase.

Jon can you clean up any issues you have already addressed (usually by pasting in a link to the pull request (even if closed) where those changes were added

Then once done can you also clean this issue

Right now 'read_raw_boxy' will only read data from a single boxy data file.

Since .mtg and .elp files contain channel coordinates for all conditions/blocks for a single participant, all data files associated with a participant should be imported.

Perhaps the best way is to point to a folder and find all relevant files for a given participant (similar to nirx.py). We need to make sure that file names are consistent so that each file can be easily found.

Scroll down to the test here:
mne-tools#7717

for each of them you can click though and find what is making the test fail,
the first one is easy so try that first, it will eventually get to a log that indicates a few style issues that are not correct

Here is a checklist of things to complete before submitting a formal review. An abridged version from info on this site:

https://mne.tools/stable/install/contributing.html#mne-python-coding-conventions

• Test coverage for new functionality (test output of boxy.py and compare to another source, such as p_pod)

• Document new functionality (thorough docstring explanations and tutorials explaining functionality)

• Avoid public API changes if possible (don't think we've made any such changes... Just added our own functions)

• Add deprecation warnings if removing/changing functionality (again, don't think this applies to us)

• Describes changes in the changelog, "doc/whats_new.rst" (sounds like this should be done AFTER the PR is finalised)

• Test locally before opening PR (kind of did both, tested locally and using CI tests)

• Make tests fast and thorough (currently tests compare raw, epoch, evoke to p_pod. Anything else? Need to make smaller data files for such tests)

• adhere to pep8 (currently seems fine, but can check by running "make flake")

• Use consistent variable names (should make sure variable names make sense too)

• follow NumPy style for docstrings (again, should be fine now but can be checked by running "make docstyle")

• Docstrings and documentation pages should cross-reference mentioned classes, functions, etc. (can probably use the "plot_70" tutorial as an example)

• Use single quotes where possible

• Prefer generators or comprehensives over filter() and map() (need to check into this...)

• Use explicit functional constructors (eg. list(), dict(), set())

• Avoid nested functions or class methods if possible

• Avoid *args and **kwargs in function/method signatures

• Import in this order (python built-in > standard scientific (eg. numpy, scipy) > others > MNE imports)

• Methods should modify in-place and return self

• Functions should return copies, where applicable

• Docstrings should contain informative names for return values, even if that particular name isn't used

• Visualisation capabilities should be available in method and function form

mne/preprocessing/nirs/_optical_density.py

what do these changes you made do:

I would like the tutorial to print 1/3 as many plots.
I think the best option is to not plot separately for montage a and b, but to combine them and only plot the combo

I also have an issue on my machine that the ones with subplots don't work, the first panel is plotted with a bunch of white squares for the others, than the other figures appear one by one, but blank, I assume the same isn't happening for you?

I think this is an issue with my matplotlib backend

once we get a smaller set of plots, lets focus on improving their look by fixing the range, etc.