The outputs of matchVisits.py (a table indexed by coadd ID, and whose columns are ID and match_distance for each visit) are used in the QA dashboard. I created this task because at the time, as far as I knew, there were not any persisted data products containing this match information, which is necessary to compute things like repeatability metrics, etc. Perhaps by now something like this exists somewhere else? If not, it seems that it should definitely be moved to/replicated in pipe_tasks.

Also another thought: moving matchVisits to pipe_tasks should probably also include an evaluation of its performance, and/or update of the matching algorithm, etc. I originally wrote the task with the expectation that the matching computation itself would be eventually replaced by something more tested/robust/used elsewhere in the pipeline.

To bookmark slack discussion with Jeffrey Carlin, the output of this task would also be useful for matchedVisitMetrics.py. This makes at least two independent use cases for matchVisits.py, suggesting it might be worth it to improve and migrate it to pipe_tasks.