DM-15554 was left in an inconsistent state. Commits were made in branch “15554” and “tickets/DM-15554” which were merged on tickets/DM-15554 without properly resolving the conflicts. (The merge commit contained all the conflict markers in the files.)

At today’s hacking (2018-12-17), I removed the DM-15554 merge commit, cherry-picked the two commits from branch “15554” instead and then rebased the branch to master. Unfortunately, the before the hackday state of "tickets/DM-15554" was not saved, the before rebase version is saved as u/gkovacs/hack2018-12-17_DM-15554.

I tried to keep code from the master and converted documentation from the branch, though there were some minor parallel evolution. There is an additional manual commit for minor changes. Notes:

• exampleCmdLineTask.py - old style documentation is now commented out

• _init_ documentation is removed where the class documentation is already updated.

• _all_ was set in the branch in the meantime in the following files: imageDifference.py ingest.py ingestCalibs.py ingestPgsql.py makeDiscreteSkyMap.py makeSkyMap.py matchBackgrounds.py measurePsf.py objectMasks.py repair.py scaleVariance.py setPrimaryFlags.py .

• multiBand.py — removed import numpy that was added as the new style docstring block

• objectMasks.py - readFits method docstring is somewhat converted but commented out, needs further manual consideration

• scaleZeroPoint.py - kept docstring in branch _all_ which is more extended than the master version

• setConfigFromEups.py — In the master version setAstrometryConfigFromEups() docstring was already modified in the old format, added the wildcards (asterisk).

• snapCombine.py transformMeasurement.py - Docstring brief should start with opening triple quotes line.

Package trial build testing and docs build failed with the same error at end of day, see attachment; this may be something inconsistent with the docs build on my desktop now, unrelated to the changes in this branch now.

Sorry, I could not check whether the new docstrings look OK.

Please also see CI build whether code is not got trivially broken.

I completed assembleCoadd.py and calibrate.py today. The former was mostly already done (thanks past me and DM-13950), but the large swaths of text now live in the new task topic type docs. The latter took a lot of hunting to figure out what various parameter types were.

I completed characterizeImage.py, measurePsf.py, and repair.py today, along with their corresponding task topic docs.

I'm not going to get any more done on this for now, so I'm un-assigning it from myself.

Brief current status is as follows:

• Initial numpydoc conversion has happened in all .py files, and skeleton task topic doc files exist
• assembleCoadd.py, calibrate.py, characterizeImage.py, measurePsf.py, and repair.py are all nominally complete, as well as their corresponding task topic docs
• The other 37 python files still need to be gone through manually. The bulk of the work will involve looking up the correct object type for various parameters, moving debug/example docstring prose to the task topic docs, making sure all input/return variables are included in all task/function docstrings, and removing and condensing duplicate/contradictory descriptions of tasks/functions.
• Building the docs and fixing any remaining warnings that pertain to pipe_tasks
• Building the package/running Jenkins to ensure nothing is broken

All of the remaining work is easily 10 story points or more. It may make sense to break this up into a few sub-tickets so less time is spent dealing with rebasing. At the time of this writing, the ticket branch is up to date with master.

All of the final work and ticket branches will be merged from DM-35939.