multiprocessing module best practices?

[Christian Calderon]

Is there a good blog or tutorial on when it is most appropriate to use the
Process vs Pool classes, and when you should subclass either of them? I
have written a few programs using these tools, and each time I have to sort
of make up my own design patterns as I go. I want to know what the
community thinks is the most pythonic use of the multiprocessing module (I
could just subclass Process every single time, why shouldn't I?)

I'm also curious about when I should be importing modules in the main
module to use in a Process, and when I should import the module inside the
run method/target function. Certain modules just won't work unless I import
it in the Process; I hate having a ton of import statements inside my
functions, but I also hate having to guess-and-check just to figure out
which modules I can import outside the function. And it is just kinda ugly
to have import statements at multiple parts of the same piece of code, IMO.

Here is an example of something I wrote that feels kinda funky (my excuse
is that the code-base I was working with was incredibly funky to begin
with, and I was trying to make a small corner of it nicer/actually work.):
    https://github.com/zack-bitcoin/augur-core/blob/master/database.py
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/python-list/attachments/20150223/bbcc5156/attachment.html>