cryptoholder-la / eegvis

visualization functions for EEG in matplotlib and jupyter notebooks

eegvis

• work in progress to visualize EEG files in python
• currently lots of experimentation - not ready for general consumption
• goal is to provide portion of software infrastructure for EEG analysis

Scope of Project

• functions to work with matplotlib to visualize EEG
• functions to interactively plot EEGs in the jupyter notebook
• web (bokeh) applications to browse and annotate EEGs

Status

• currently this is still experimental, but it has a useful eeg browser for the jupyter notebook

• basic filtering added, with notch and low-pass/high-pass filter

• version 0.2.1 works with bokeh 2.0.2 !

install

jupyter nbextension enable --py widgetsnbextension --sys-prefix

To do

• matplotlib EEG plotting (firstdraft)

• basic bokeh EEG plotting (firstdraft)

• basic montage EEG plotting, (firstdraft)

• simple browsing EEG with bokeh in jupyter (firstdraft)

• first filtering dropdowns added to nb browser tool

• allow kwargs to set plot width and height

• [/] notch and HF filters dropdowns - problem with ringing on current firwin filters

• [/] need scale/calibration bars - [x] vertical scale bars can be added for any stackplot

• catch when current displayed data is not big enough to filter

• add common avg reference montage (CAR)

• remove cruft from plotting in various widgets

• [/] bokeh application to browse + annotate EEG, - still experimenting

• montage parser/loader (priority long-term)

• keyboard responses, howto?

• add ability to control scale of each electrode waveform individually

• rewrite and package

• publish

• possible re-write/extend Bokeh for canvas widget

• update to bokeh 1.0.x (now at 2.3)

• tests for mpl_helpers

• clearcut examples of using plotting tools with:

  • numpy arrays
  • edf EEGs
  • eeghdf EEGs
  • mne eeg and/or meg (should this use, be in eegml-mne-utils)

• [/] trial implement panel based version of browser widget

• explore low-level implementation of browser widget using bokeh for more control

• explore new annotation bokeh controls

• [/] switch to new packaging for ppa using pyproject.toml and move setup.py to setup-dev.py

• add tools to mark events in plots and in bokeh

Other notes

• Note, to respond to keyboard commands in bokeh, probably need an extension: see: https://groups.google.com/a/continuum.io/forum/#!searchin/bokeh/keypress/bokeh/XCLqg1nyIgE/CU7lJGcuBgAJ

Developer Install

currently this is developer only package To install from the head of development

pip install -U git+https://github.com/eegml/eegvis.git

I have started using f-strings in my code, so I no longer support python 3.6 and below

To work on the code yourself:

    git clone https://github.com/eegml/eegvis.git
    cd eegvis
    pip install -e .

jupyter notebook use

conda install ipywidgets conda install widgetsnbextension

jupyter lab use

jupyter labextension install jupyterlab_bokeh jupyter labextension install @jupyter-widgets/jupyterlab-manager