This will be a comprehensive toolkit to combine data from electrophysiology recordings and Voyeur. Currently it is built to work with raw data files recorded from spikeGL, but it can be easily updated to work with the future systems (think OpenEphys).

It is designed with the following design objectives in mind:

1. Data at rest should be transparent, easy to understand, and as minimally processed as is possible. This allows users to parse data the way that makes sense for analysis at analysis run-time.
2. Data-at-rest is stored using a powerful and open format once processed (HDF5).
3. No substructure is assumed, and continuously recorded data are stored continuously (with the same time base) at rest. Information to recreate substructures (ie trials) are stored with the data.
4. Only two files should be required once data are processed: a raw data file and a processed data file. All metadata should be included within these files and propagated through the processing system.
5. Data and metadata should be merged early in the processing stream and should be persistent through subsequent processing and analysis steps and data structures.

It is designed with the following packages:

This package contains packages for:

• ephys_tools.data_handling.pre-sort parses data for sorting with the KlustaSuite of sofware (citation).
• ephys_tools.data_handling.post-sort parses data into a structure that is useful for analysis. (See below for file structure).
• ephys_tools.data_handling.data_classes module provides OO python interface for loading and manipulating data created by the above packages.
• ephys_tools.data_handling.exporters package providing modules to export HDF5 file data to other file formats (ie Matlab). Not currently implemented and may not be necessary.

This also contains a file_structure_readme.md as a reference to the file structure.

Analysis modules. TBD.