When we use the LLM assistant to create a simple scene or story,
the position of segments in the timeline isn't right.
I need to fix the code that takes the rawo JSON output from the LLM and convert it to timeline coordinates:
there are gaps in the generate timeline, some which can be fixed using assumptions (eg. we always have a camera, always have a location etc so we should extend the previous segment when that happens)
sometimes it's the track which is wrong, due to a missing segment
When we use the LLM assistant to create a simple scene or story, the position of segments in the timeline isn't right.
I need to fix the code that takes the rawo JSON output from the LLM and convert it to timeline coordinates:
there are gaps in the generate timeline, some which can be fixed using assumptions (eg. we always have a camera, always have a location etc so we should extend the previous segment when that happens)
sometimes it's the track which is wrong, due to a missing segment