Consider using LSST computeExposureSummaryStatsTask for calexp metrics

[danjampro]

LSST already have a task that does much of the same work as our calexp metrics:

https://github.com/lsst/pipe_tasks/blob/master/python/lsst/pipe/tasks/computeExposureSummaryStats.py

However the LSST code does not contain some important metrics like zero-point error, whereas ours does. We could consider overriding this subtask in future, or just keep it as a reference.