the Lightning App periodically updates the CollectedData.csv file by reading from the Label Studio database. Previously this was hard-coded for short intervals to ensure timely updates to the csv file, but if there are too many labels then this process will start again before the previous process completed. Now the interval is more dynamic and depends on the overall number of labeled frames, and will scale better to datasets with thousands or tens of thousands of frames.
the Lightning App periodically updates the CollectedData.csv file by reading from the Label Studio database. Previously this was hard-coded for short intervals to ensure timely updates to the csv file, but if there are too many labels then this process will start again before the previous process completed. Now the interval is more dynamic and depends on the overall number of labeled frames, and will scale better to datasets with thousands or tens of thousands of frames.