I've looked for all these strings as literals in the stack, including doing matches goodSeeing, chiSquared appended to the "deep" ones. Here are the only matches which I found. There appear to be no butler calls (like get, set, or datasetExists) but these are probably unnecessary mapper definitions which could be removed. I'm not sure if there are other syntaxes which I need to worry about, other than the string itself and the string with quotes around it. I searched all .py, .h, and .cc files.
/pipe_tasks//python/lsst/pipe/tasks/mocks/simpleMapper.py:252: deepCoadd_src = SkyMapping(SourceCatalogPersistenceType),
/pipe_tasks//python/lsst/pipe/tasks/mocks/simpleMapper.py:253: deepCoadd_src_schema = SimpleMapping(SourceCatalogPersistenceType,
/pipe_tasks//python/lsst/pipe/tasks/mocks/simpleMapper.py:249: deepCoadd_icSrc = SkyMapping(SourceCatalogPersistenceType),
/pipe_tasks//python/lsst/pipe/tasks/mocks/simpleMapper.py:250: deepCoadd_icSrc_schema = SimpleMapping(SourceCatalogPersistenceType,
/obs_cfht/python/lsst/obs/cfht/megacamMapper.py:192: def bypass_stackExposureId(self, datasetType, pythonType, location, dataId):
/obs_cfht/python/lsst/obs/cfht/megacamMapper.py:196: def bypass_stackExposureId_bits(self, datasetType, pythonType, location, dataId):
/obs_subaru/python/lsst/obs/suprimecam/suprimecamMapper.py:256: def bypass_stackExposureId(self, datasetType, pythonType, location, dataId):
/obs_subaru/python/lsst/obs/suprimecam/suprimecamMapper.py:259: def bypass_stackExposureId_bits(self, datasetType, pythonType, location, dataId):
/pipe_tasks//python/lsst/pipe/tasks/mocks/simpleMapper.py:249: deepCoadd_icSrc = SkyMapping(SourceCatalogPersistenceType),
/pipe_tasks//python/lsst/pipe/tasks/mocks/simpleMapper.py:250: deepCoadd_icSrc_schema = SimpleMapping(SourceCatalogPersistenceType,
The are also a set call goodSeeingDiff which do not appear to be used.
To verify that these changes are correct (since unit test coverage of the mappers is not good), it'd be a good idea to introspect each of the mappers to get a list of all of their datasets and their template paths, dump this to a file, and then check that the list is the same after the changes have been made (note that we probably don't care about sort order in this comparison, so it probably makes sense to sort the datasets somehow before dumping them to files). I'm not entirely sure what the interface is for doing that introspection, but I think doing
DM-6858first (since that blocks this) will provide at least a hint for that.