The "submission_scored" trigger should be triggered only for the active dataset

[wil93]

Right now, if I clone a dataset and I run some submissions (on the cloned dataset) the "social" scores will be updated anyway. A possible solution is to run the trigger only for the active dataset. This has the drawback of forcing you to re-run the scoring phase after activating a dataset. A better solution is to run the trigger also if the active dataset has changed.