https://docs.djangoproject.com/en/1.8/howto/custom-management-commands/
You need to declare and do proper argument parsing.  --commit is also
more natural than a bare 'commit' on the end of a command.

Switch from calling handle() directly to call_command().
call_command() simulates using the management command so is a better
test of the command line interface.

With the 1.8 upgrade, we stopped using 'manage.py cms command commit'
and switched to --commit because we no longer directly access args[]
https://docs.djangoproject.com/en/1.8/howto/custom-management-commands/

When this command was updated, it was switched to --commit=commit, but
using store_true allows a much more natural --commit syntax.

Please note that this is a squshed commit and the work of:
Symbolist, macdiesel, nedbat, doctoryes, muzaffaryousaf and muhammad-ammar

* Fix paver violations to stablize edx-lint update
* Parens, Line2Long
* Fix missing docstrings
* Fix PEP8 issues
* Address PR feedback (thanks @nedbat!)

'force_publish' command for publishing  a course forcefully by making the published branch point to the same structure as the draft branch. add force_publish functionality to split courses.

fix quality violations

create child parent mapping to avoid potential performance hit when deleting items

(PLAT-835)
(TNL-844)

The print_out_all_courses() routine consumes a ton of memory (2G and
causes noticable mongo usage spikes).  This actually causes other
processes on production boxes to be memory starved and killed
(such as worker children on edge when this was run recently).

The behavior of this script on production is
* Print several hundred courses
* Ask if you want to delete the one you specified
* print several hundred courses minus one
On a sandbox with 5 courses, you could tell by eye that 1 is gone, but
not in production (or even in stage).

The original PLAT-619 ticket for this suggested printing a course
listing on error, but instead it always printed the course listing.
Even in the error case, hundreds of course ids is confusing and obscures
the error message saying that your course_id is invalid.

You should be getting the course id from the UI or from ./manage.py lms
dump_course_ids, not by searching a list.

Adjusted the test accordingly

Remove get_courses_keys

The old line:

    from path import path

produced pylint errors because of the baroque way that path.py defined
"path".  We tried to get them to change how they defined it, but they
deleted the name instead: https://github.com/jaraco/path.py/issues/102

(Jason then changed his mind, but this is a better way to use path.py,
it avoids the pylint error at least.)

TNL-3014

Add tests for search

Add text_search to TeamsListView

Add reindex command line tool for course teams

Add Search Pagination and update comments

Move paginate_search_results to common library

This is a fix for
https://openedx.atlassian.net/browse/PLAT-619
added custom CommandError messages for invalid syntax and
invalid course keys.

Skipped 'print' statement violations, as print() is Py3k syntax.

This is a fix for
https://openedx.atlassian.net/browse/PLAT-324
added custom CommandError messages for invalid course key.

Fixed cms course export

This makes these files importable, and puts them into the new best guess
as to where files should live.

Update edx-lint to the version that checks if tearDown uses super.

Convert a number of tearDown methods into addCleanup.

Remove some unneeded tearDown methods: no need to call patch.stopall if
none of them were started with patch.start.

TNL-1362

Testing migration from old mongo to split.

Note that now 2 courses can be imported where the only difference between them is run.

export_courses_to_output_path uses old mongo locator to create directory structure.

The way that assets are accessed (the usage of _id) seems specific to old mongo and fails with split.