Bringing local fork up to Master from main django haystack project

Update with v2 4 1

This allows somone to do certain things outside of the base object layer.

An example:  You have to call a 3rd party for some data and it is better
to do it in bulk once and then do a lookup in memory while looping through
the individual document creation.  This could avoid tons of round trip
requests to that 3rd party.

…d log batch execution times.

…xisting __str__ method instead since it is already working for both python 2 and 3

After the multiprocessing pool is done the connection that the parent
process had established is stale.  For whatever reason, django is unable
to detect this, and tries to reuse the connection, and gets an error on
the next database query.

I am forcing all DB connections to close immediately after the proc pool
is closed, which means that it establishes a new connection before the
next query.

logging configuration set up by our app, and double messages.

with our own logger. The multiprocess logger is only useful if you are
logging to a file.  We log to stdout.  This logger also forces messages
to stderr with a specific format that we don't want.

Update to 2.5.1

…-To-Run-2

Bug/web 2808 sm reindex fails to run 2

rgg version bump

modern pip doesn't like our made-up versioning scheme

also added packaging steps

WEB-4967: Package rename and semver