phy: interactive visualization and manual spike sorting of large-scale ephys data

phy is an open-source Python library providing a graphical user interface for visualization and manual curation of large-scale electrophysiological data. It is optimized for high-density multielectrode arrays containing hundreds to thousands of recording sites (mostly Neuropixels probes).

Phy provides two GUIs:

• Template GUI (recommended): for datasets sorted with KiloSort and Spyking Circus,
• Kwik GUI (legacy): for datasets sorted with klusta and klustakwik2.

What's new

• [5 June 2024] phy 2.0 beta 6, bug fixes, install work, fixing some deprecations
• [7 Sep 2021] Release of phy 2.0 beta 5, with some install and bug fixes
• [7 Feb 2020] Release of phy 2.0 beta 1, with many new views, new features, various improvements and bug fixes...

Links

• Documentation
• Mailing list

Hardware requirements

It is recommanded to store the data on a SSD for performance reasons.

There are no specific GPU requirements as long as relatively recent graphics and OpenGL drivers are installed on the system.

Installation instructions

Run the following commands in a terminal (currently working for Linux machines):

1. Create a new conda environment with the conda dependencies:

   conda create -n phy2 -y python=3.11 cython dask h5py joblib matplotlib numpy pillow pip pyopengl pyqt pyqtwebengine pytest python qtconsole requests responses scikit-learn scipy traitlets
   

2. Activate the new conda environment with conda activate phy2

3. Install the development version of phy: pip install git+https://github.com/cortex-lab/phy.git

4. [OPTIONAL] If you plan to use the Kwik GUI, type pip install klusta klustakwik2

5. Phy should now be installed. Open the GUI on a dataset as follows (the phy2 environment should still be activated):

cd path/to/my/spikesorting/output
phy template-gui params.py

6. If there are problems with this method we also have an environment.yml file which allows for automatic install of the necessary packages. Give that a try.

Dealing with the error ModuleNotFoundError: No module named 'PyQt5.QtWebEngineWidget

In some environments, you might get an error message related to QtWebEngineWidget. Run the command pip install PyQtWebEngine and try launching phy again. This command should not run if the error message doesn't appear, as it could break the PyQt5 installation.

Upgrading from phy 1 to phy 2

• Do not install phy 1 and phy 2 in the same Python environment.
• It is recommended to delete ~/.phy/*GUI/state.json when upgrading.

Developer instructions (and instructions for some Windows machines)

To install the development version of phy in a fresh environment, do:

git clone git@github.com:cortex-lab/phy.git
cd phy
pip install -r requirements.txt
pip install -r requirements-dev.txt
pip install -e .
cd ..
git clone git@github.com:cortex-lab/phylib.git
cd phylib
pip install -e . --upgrade

Mac Install

Since the switch to M-series chips Mac install for Phy is not being officially supported. Rarely people are able to hack together a version with old versions of python etc.

Troubleshooting

• See a list of common issues.
• Raise a GitHub issue.

Running phy from a Python script

In addition to launching phy from the terminal with the phy command, you can also launch it from a Python script or an IPython terminal. This may be useful when debugging or profiling. Here's a code example to copy-paste in a new launch.py text file within your data directory:

from phy.apps.template import template_gui
template_gui("params.py")

Credits

phy is developed and maintained by Cyrille Rossant.

• International Brain Laboratory
• Cortex Lab (UCL) (Kenneth Harris and Matteo Carandini).

Contributors to the repository are:

• Alessio Buccino
• Michael Economo
• Cedric Gestes
• Dan Goodman
• Max Hunter
• Shabnam Kadir
• Christopher Nolan
• Martin Spacek
• Nick Steinmetz