Repository navigation
Correctly calculate $VIDEO_NAME macro for URL inputs #197
Description
Currently, the macro $VIDEO_NAME is not calculated correctly when using an image sequence (e.g. [1] a URL as an input video, leading to (potentially) malformed output filenames when running other commands that create output files on disk (e.g. img_%04d.jpg) orlist-scenes). This only happens when the user does not override the calculated filename.
Need to detect this case and either error out while parsing (forcing the user to provide an explicit filename), or output a warning suggesting the filename be overridden manually.
[1] Completed in v0.6
Activity
Breakthrough commentedon Jul 31, 2022
This should work correctly now for image sequences as of v0.6, but URLs still need to be tackled.
[-]Correctly calculate $VIDEO_NAME macro for image sqeuence/URL inputs[/-][+]Correctly calculate $VIDEO_NAME macro for URL inputs[/+]jennifferweslowski-design commentedon Aug 16, 2026
For URL inputs, which behavior do you want for
$VIDEO_NAMEwhen the user hasn't overridden the filename template: (a) warn and continue, (b) error out during parsing and require an explicit filename, or (c) derive the name from the URL's path component and only warn/error when there's no usable path segment (e.g.https://example.com/)?jennifferweslowski-design commentedon Aug 17, 2026
For a URL with no usable path (empty, or
$VIDEO_NAMEpolluted by query/fragment), should PySceneDetect warn and fall back to a generic name, or reject the input? I don't want to pick that API behavior without your call.Breakthrough commentedon Oct 21, 2026
Sorry for the delay in responding to this, I think either A or C sounds fine as per #582.
fix: derive $VIDEO_NAME from the path of URL inputs
jennifferweslowski-design commentedon Oct 21, 2026