Repository navigation
Request for new feature - Support for .dat heka files #1826
Description
This is a feature request to add support to heka files (Dat files).
Neo recommends to use the stfio module for loading these files. The problem is that this library is c++ based and requires a complex building process.
We currently use an openly available python library to load these files. It would be fantastic to stitch our file loader into the neo framework.
I have included a jupyter file which has the 4 following sections:
- Loads a single trace from the dat file;
- Loads 50 traces from Group 0, Series 0
- Explains the full tree structure of a dat file (arrangement of ephys data according to groups, series and traces)
- Displays all the channels used while recording
The jupyter file uses this heka_reader - https://github.com/campagnola/heka_reader
I have also uploaded an example dat file in a google drive folder - https://drive.google.com/drive/folders/12qdDNXsT5PmScvdhxQpDU2vLeUPe8qai?usp=sharing
The jupyter file and heka reader file can be found on this github repo: https://github.com/vsj37/Neo_heka
Activity
apdavison commentedon Mar 17, 2026
Related issue: #1050
zm711 commentedon Apr 3, 2026
@vsj37,
sorry I've been away from Neo for a spell. But looking over this it seems like there is no license or build for that particular package. The author told one person in an issue to just fork it and take over if they wanted.
But this means that 1) we can't have a dependency since this is not easily downloadable for Neo users from pypi and 2) as there is no official license we can't really just vendor in the code to Neo either. I think encouraging @JoeZiminski to finish his reader here as discussed in #1050 is your best bet unless you can get us in touch with the author of that python script so we can decide what we can do with it.
JoeZiminski commentedon Apr 10, 2026
Thanks @zm711 @vsj37 it would be really nice to finish #1050, unfortunately I don't think I'll realistically have capacity to work on it for the foreseeable future. But if anyone else is interested in taking it on would be happy to support, or do anything necessary to make load-heka-python more accessible (e.g. pypi release) in aid of wrapping in neo.