neuroconv OpenEphysRecordingInterface get_stream_names()

[JustinKillebrew]

I'm trying to convert an OpenEphys session using python. The session contains multiple streams that I'd like to add to the nwb file. When I create the OpenEphysRecordingInterface I get an error about the multiple streams and I need to specify which stream to open. The error suggests using OpenEphysRecordingInterface.get_stream_names(folder_path=...) to get the stream names, but that fails with "no attribute 'get_stream_names'".

How do I get the list of stream names and then convert and add each stream to the nwb file?

I can create the interface and then get the stream names but I have to pass in a valid stream name to the constructor but I need to know the name of a stream first.

Thanks,
JK

[oruebel]

I think the streams are just the directory names in the folder, i.e., I think you get the stream names that way via something like

from pathlib import Path

p = Path(folder_path)
streams = [item.name for item in p.iterdir() if item.is_dir()]

[JustinKillebrew]

Sorry for the delay. This only returns the top level dir, Record_Node_101. There are 4 recording dirs, each with a continuous and events directory. The continuous dirs contain the dirs: Neuropix-PXI-ProbeA-AP, Neuropix-PXI-ProbeA-LFP, NI-NAQmx-104.PXIe-6341. Those are (part of) the stream names.

Thanks,
JK

[h-mayorquin]

Hi, can you describe two things for me:

• How does the folder you are using looks like, which files you have?
• What streams you have? Is that only electrophysiology? Both bands? NIDQ? Any other analog sensor?

Best regards,
Probably related to:
catalystneuro/neuroconv#1504

[JustinKillebrew]

Of course! Here's the full message:

/home/justin/anaconda3/envs/nwb-guide/lib/python3.12/site-packages/spikeinterface/extractors/init.py:60: UserWarning: Importing classes at init has been deprecated in favor of only importing function-size wrappers and will be removed in 0.105.0. For developers that prefer working with the class versions of extractors they can be imported from spikeinterface.extractors.extractor_classes
warn(dep_msg)
Traceback (most recent call last):
File "/home/justin/jkcode/python/nwb/demo_oephys/test3.py", line 8, in
interface = OpenEphysRecordingInterface(folder_path=folder_path)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/justin/anaconda3/envs/nwb-guide/lib/python3.12/site-packages/neuroconv/datainterfaces/ecephys/openephys/openephysdatainterface.py", line 68, in new
return OpenEphysBinaryRecordingInterface(
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/justin/anaconda3/envs/nwb-guide/lib/python3.12/site-packages/neuroconv/datainterfaces/ecephys/openephys/openephysbinarydatainterface.py", line 71, in init
raise ValueError(
ValueError: More than one stream is detected! Please specify which stream you wish to load with the stream_name argument. To see what streams are available, call OpenEphysRecordingInterface.get_stream_names(folder_path=...).

[h-mayorquin]

Run this:

`from datetime import datetime
from zoneinfo import ZoneInfo
from pathlib import Path
from neuroconv.datainterfaces import OpenEphysRecordingInterface

folder_path = "/home/justin/testData/2025-03-06_14-30-33/Record_Node_101"
streams = OpenEphysRecordingInterface.get_stream_names(folder_path = folder_path)
print(streams)

what do you get?

[JustinKillebrew]

Traceback (most recent call last):
File "/home/justin/jkcode/python/nwb/demo_oephys/test4.py", line 7, in
streams = OpenEphysRecordingInterface.get_stream_names(folder_path = folder_path)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AttributeError: type object 'OpenEphysRecordingInterface' has no attribute 'get_stream_names'

[h-mayorquin]

You need to update neuroconv to the latest version.

[JustinKillebrew]

get_stream_names() is working now.

I've reinstalled neuroconv (and a couple dependencies) in a fresh environment.

$ conda create -n nwb python=3.12
$ pip install neuroconv
$ pip install spikeinterface
$ pip install lxml

Thank you.
JK

[h-mayorquin]

Do the stream names make sense to me? I think this is something that could be improved but I would like to hear from an experimentalist.

[JustinKillebrew]

I'm not sure I understand but here are the stream names as reported by get_stream_names():
'Record_Node_101#NI-DAQmx-104.PXIe-6341',
'Record_Node_101#Neuropix-PXI-100.ProbeA-AP',
'Record_Node_101#Neuropix-PXI-100.ProbeA-LFP',
'Record_Node_101#Neuropix-PXI-100.ProbeA-APSYNC',
'Record_Node_101#Neuropix-PXI-100.ProbeA-LFPSYNC'

Unfortunately I am not an experimentalist - I'm more of an engineer assigned the job of converting data to NWB. I am in contact with the group that recorded this data. The stream names are correct and make sense to me.

I'm trying to convert OpenEphys data files and include processed data/metadata from mat files so that a single nwb file will contain all the data, processed data, and trial metadata. I'm running interface.run_conversion() now on a single stream (I think) and it's taking quite awhile. Ideally the interface would convert all streams to the nwb file and not one by one since it looks like I need to iterate through each stream adding them to the nwb file.

Thanks,
JK

[h-mayorquin]

Here:
catalystneuro/neuroconv#1686

We can keep the discussion there