On Tue, Jul 15, 2014 at 8:32 AM, Catherine M Moroney
<catherine.m.moro...@jpl.nasa.gov> wrote:
> The actual scope of the problem is very small, so memory/cpu time is not
> an issue. I'm just looking for the most pythonic/elegant way of doing this.
Small job? Use the simplest possible technique. Just create
"params.py" with a bunch of assignments in it:
# params.py
a = 1
b = 2
c = a + b
# every other file
import params
print("c is",params.c)
# if you need to change anything:
params.c += 5
# everyone else will see the change, because there can be
# only one instance of the module (Highlander!)
Works nicely for anything even moderately complex. Also serves as a
convenient way to separate configs from code; for instance, I do this
any time I need to have a program with database passwords, or
per-installation setup, or stuff like that. Two examples:
https://github.com/Rosuav/Yosemite/blob/master/config.py
https://github.com/Rosuav/Flask1/blob/master/1.py
In the latter case, config.py doesn't even exist in the repository, as
its main purpose is to store the database connection string - both
private (don't want that published on Github) and per-installation (my
dev and production systems use different connection strings). The
Yosemite config is actually a bit legacy now; I used to have two
distinctly different instances of it, one running on Windows and the
other on Linux, but now I have a large number of identical instances
(all on Linux and all referencing the same disk server - which,
incidentally, is the one that I've weaponized with Alice, Elsa, Anya,
a Vorpal blade, and a Portal turret). Either way, though, config.py
consists generally of simple assignments (and comments), but it's most
welcome to use all the power of Python to calculate values.
ChrisA
--
https://mail.python.org/mailman/listinfo/python-list
