separation function is a memory hog

[Enno Middelberg]

The function that calculates the separation of bodies (either in RA and Dec or in Az/El) consumes a lot of memory when called repeatedly. The following code grabs 1.5GB of RAM in a few seconds:

import ephem as e
centre=e.readdb("centre,f|M|F7,10:00:00, 50:00:30.0,2.02,2000")
target=e.readdb("test,f|M|F7,10:00:00, 50:00:00.0,2.02,2000")
centre.compute()
target.compute()

for x in range(1e7):
   sep=e.separation([centre.ra, centre.dec], [target.ra, target.dec])

If this isn't a bug then I'd like to know how I can call the separation function many times without it grabbing so much RAM.

Regards,

Enno

[Brandon Rhodes (brandon-rhodes)]

Enno, thank you very much for reporting this! And sorry for any trouble that this bug caused you; I'll buy you a beer if we ever wind up at a conference together. :-)

There was indeed a memory leak in separation() in the specific case that it is given coordinate sequences instead of bodies or observers. I have pushed a fix to the PyEphem Bitbucket repository, and will be releasing a new PyEphem version this weekend with the fix. Here is the changeset in the repository in case you want to try the new version before I do the release:

http://bitbucket.org/brandon/pyephem/changeset/95bf603370e7

Let me know if you have any further problems!