Exception in _init_video_manager() when video fails to load on non-Windows platforms

[paulagd]

Bug/Issue Description:
I am trying to run the detection in my sample video but it fails like:

scenedetect -i sample_seconds.mp4 detect-content list-scenes -n save-images

> scenedetect -i sample_seconds.mp4 detect-content list-scenes -n save-images
[PySceneDetect] Could not parse CLI options.
Traceback (most recent call last):
  File "/home/paulagomez/.virtualenvs/research/lib/python3.6/site-packages/scenedetect/cli/c
ontext.py", line 517, in _init_video_manager
    video_files=input_list, framerate=framerate, logger=logging)
  File "/home/paulagomez/.virtualenvs/research/lib/python3.6/site-packages/scenedetect/video
_manager.py", line 373, in __init__
    video_files=video_files, framerate=framerate)
  File "/home/paulagomez/.virtualenvs/research/lib/python3.6/site-packages/scenedetect/video
_manager.py", line 241, in open_captures
    raise VideoOpenFailure(closed_caps)
scenedetect.video_manager.VideoOpenFailure: OpenCV VideoCapture object failed to return True
 when calling isOpened().

During handling of the above exception, another exception occurred:
Traceback (most recent call last):                                                  [3/1926]
  File "/home/paulagomez/.virtualenvs/research/bin/scenedetect", line 8, in <module>
    sys.exit(main())
  File "/home/paulagomez/.virtualenvs/research/lib/python3.6/site-packages/scenedetect/__mai
n__.py", line 58, in main
    cli.main(obj=cli_ctx)   # Parse CLI arguments with registered callbacks.
  File "/home/paulagomez/.virtualenvs/research/lib/python3.6/site-packages/click/core.py", l
ine 782, in main
    rv = self.invoke(ctx)
  File "/home/paulagomez/.virtualenvs/research/lib/python3.6/site-packages/click/core.py", l
ine 1268, in invoke
    Command.invoke(self, ctx)
  File "/home/paulagomez/.virtualenvs/research/lib/python3.6/site-packages/click/core.py", l
ine 1066, in invoke
    return ctx.invoke(self.callback, **ctx.params)
  File "/home/paulagomez/.virtualenvs/research/lib/python3.6/site-packages/click/core.py", l
ine 610, in invoke
    return callback(*args, **kwargs)
  File "/home/paulagomez/.virtualenvs/research/lib/python3.6/site-packages/click/decorators.
py", line 21, in new_func
    return f(get_current_context(), *args, **kwargs)
  File "/home/paulagomez/.virtualenvs/research/lib/python3.6/site-packages/scenedetect/cli/$
_init__.py", line 293, in scenedetect_cli
    frame_skip=frame_skip)
  File "/home/paulagomez/.virtualenvs/research/lib/python3.6/site-packages/scenedetect/cli/$
ontext.py", line 586, in parse_options
    input_list=input_list, framerate=framerate, downscale=downscale)
  File "/home/paulagomez/.virtualenvs/research/lib/python3.6/site-packages/scenedetect/cli/$
ontext.py", line 526, in _init_video_manager
    dll_okay, dll_name = check_opencv_ffmpeg_dll()
TypeError: 'bool' object is not iterable

Computing Environment:

• OS: Ubuntu 18.04.4 LTS
• Python Version: 3.6.9
• OpenCV Version: opencv-python 4.4.0.46

Additional Information:
I have also installed:

ffmpeg                    1.4
imageio-ffmpeg            0.4.2

Can someone tell me what am I doing wrong please?

Thank you in advance!

[Breakthrough]

Hi @paulagd;

Sorry for the delay in getting back to you. The command is correct, for some reason though it is failing to load the .mp4 using the underlying OpenCV library. This, however, results in another exception getting thrown due to a bug in check_opencv_ffmpeg_dll() (it works on Windows, but this will happen on any other platform).

I'll mark this as a bug, but note that the issue is likely caused by OpenCV/ffmpeg being unable to open your test video for some reason. You can test this yourself by trying to open the file manually via OpenCV, and checking if you are able to do it (I can provide some sample code for you to run if you aren't as familiar with OpenCV).

Thanks!

[Breakthrough]

Resolved in v0.5.5 branch, will be included in next release of PySceneDetect, should give you the proper error messaging now.

As for why the video fails to load, that will still require some more investigation. Feel free to respond to this issue, or create a new one for that.

[Breakthrough]

How did you generate your sample video? Are you sure it's a valid file? How did you install opencv-python? Could you possibly include your test video file on this issue to help with the analysis?

Thanks!

[Breakthrough]

Hi @paulagd;

I'm closing this issue due to resolving the improper error messaging. As mentioned, if you could attach the test video you are using, that would be a great help to narrow down why it's failing on your end. Feel free to respond to this issue, or create a new one, if you are still running into this issue after the next release (v0.5.5).

Thank you again for the report.