Create_labeled_video_3d matching incorrect files

Attempting to create a 3d video using this code:

deeplabcut.create_labeled_video_3d(config=r'G:\Shared drives\Middleton_Lab_Share\Josh\Mouse3D-josh-2019-07-16-3d\config.yaml',
path=[r'G:\Shared drives\Middleton_Lab_Share\Josh\Mouse3D-josh-2019-07-16-3d\videos\2014-10-24_2300_T1_josh.h5'],
videofolder=r'G:\Shared drives\Middleton_Lab_Share\Josh\Mouse3D-josh-2019-07-16-3d\videos',
videotype='mp4')

generates an out of range error:

[['G:\\Shared drives\\Middleton_Lab_Share\\Josh\\Mouse3D-josh-2019-07-16-3d\\videos\\2014-10-24_2300_T1_josh.h5', 'G:\\Shared drives\\Middleton_Lab_Share\\Josh\\Mouse3D-josh-2019-07-16-3d\\videos\\2014-10-24_2300_T10_josh.h5', 'G:\\Shared drives\\Middleton_Lab_Share\\Josh\\Mouse3D-josh-2019-07-16-3d\\videos\\2014-10-24_2300_T1_Cam1.mp4', 'G:\\Shared drives\\Middleton_Lab_Share\\Josh\\Mouse3D-josh-2019-07-16-3d\\videos\\2014-10-24_2300_T1_Cam2.mp4']]
Creating 3D video from 2014-10-24_2300_T10_josh.h5 and 2014-10-24_2300_T1_Cam1.mp4 using 2014-10-24_2300_T1_josh.h5
warning: Error opening file (/build/opencv/modules/videoio/src/cap_ffmpeg_impl.hpp:908)
warning: G:\Shared drives\Middleton_Lab_Share\Josh\Mouse3D-josh-2019-07-16-3d\videos\2014-10-24_2300_T10_josh.h5 (/build/opencv/modules/videoio/src/cap_ffmpeg_impl.hpp:909)
Looking for filtered predictions...

The error comes here:

--> 129 df_cam1= pd.read_hdf(glob.glob(os.path.join(path_h5_file,str('*'+base_filename_cam1+cam1_scorer+'*filtered.h5')))[0])

Instead of getting the relevant files for T1, it grabs T10 (which also exists in that directory). If I switch to trial 2 (T2), it works as expected.

Is there a solution for this, aside from starting all my experiments with T2?

I’m using:

In [3]: deeplabcut.__version__
Out[3]: '2.1.6.2'

In windows.

Thanks.