SAM2 DIVE-DSA Integration

[BryonLewis]

resolves #256

TODO:

• Test base SAM2 tracking
• Add docker compose gpu profile
• Update to using 'dive_gpu' instead of just 'gpu'
• Default to using celery queue, but it will also use a dive_gpu celery queue if you set it in the config
• Update tracking to output each mask and each track
  • Update patch update for trackJSON to not log a new revision when pushing based on a parameter
• Update the tracking to use progress updates to update client on each frame
• Create Client Side
  • mask tracking interface inside of the mask editor. Displays the Models that are available and begin tracking
  • Tracking can start and it uses the job information to track progress
  • Add option to display results as the come in
    • Add function that takes in track from the job progress and adds it to the current list without incrementing the pendingsavecount
    • Add function that will add a new mask to the mask list and load it immediately
    • With frame tracking on, it will advance each X new frame that is added
• Add option for Additive vs replacement for the track
• UI Controls
  • Update so that user can set batch size
  • User can set Percentage update rate (1-50%)
  • Add Toggle for frame locking for updates
  • Icon for tracking should have an indicator of tracking being run
• Needs to properly stop if it goes past the last total frame (function update)