You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
This PR NeurodataWithoutBorders/pynwb#1793 was recently merged into PyNWB. An error is now raised when a TimeSeries is created with a rate <= 0. The nwbinspector tests try to create such a TimeSeries and fail as a result.
I do not understand the need for the second test above -- how often is a TimeSeries created with a single data point? That seems like an incorrect use of a TimeSeries and could be its own separate check, regardless of the rate specified.
Have you ensured this bug was not already reported?
To the best of your ability, have you ensured this is a bug within the code that checks the NWBFile, rather than a bug in the NWBFile reader (e.g., PyNWB or MatNWB)?
The text was updated successfully, but these errors were encountered:
I do not understand the need for the second test above -- how often is a TimeSeries created with a single data point?
Most often done for people with static images generated through microscopy techniques (I've seen it once or twice in last years) - we've even done it ourselves once or twice I believe - they could use the standard Image module types but then they can't associate all the lovely metadata about the device, imaging plane including grid spacing, optic channel, etc.
I commented on the original already merged PyNWB PR for this; as it stands this is a more fundamental misalignment between the two packages so we really should reach agreement. I think PyNWB should roll back that error to consider the length 1 time-axis condition, or otherwise make some official recommendation to how to better represent data in that case
Good to hear negative rates will no longer be allowed at the PyNWB level, but we don't actually seem to have a check for that yet (will raise an issue for that applied to old files) and note that tests for it will have to use your construct hack to get around it in order to imitate usage on older files
What happened?
This PR NeurodataWithoutBorders/pynwb#1793 was recently merged into PyNWB. An error is now raised when a TimeSeries is created with a rate <= 0. The nwbinspector tests try to create such a TimeSeries and fail as a result.
https://github.com/NeurodataWithoutBorders/nwbinspector/blob/dev/tests/unit_tests/test_time_series.py#L208-L222
I think the first test should stay because old files may have a TimeSeries with a non-positive rate, but the test should be updated to create such a TimeSeries like so:
https://github.com/NeurodataWithoutBorders/pynwb/blob/0e45cd927a0734428358eab10a75672e8dd75344/tests/unit/test_base.py#L414-L428
I do not understand the need for the second test above -- how often is a TimeSeries created with a single data point? That seems like an incorrect use of a TimeSeries and could be its own separate check, regardless of the rate specified.
Operating System
M1 macOS
Python Version
3.11
Were you streaming with ROS3?
None
Package Versions
No response
Code of Conduct
The text was updated successfully, but these errors were encountered: