once more (trying to get the hang of this 'email' thing).
my only usage is the former, so if that's faster, great, but i think
that even the current incremental loading
is fast enough that it is rarely the bottleneck in my code--especially
with the disk-based indexes
where it's only an issue the first time. but yeah, if it's faster, cool.
assuming i understand correctly,
rather than geo_interface, it could be a function passed in via kwargs that
takes an object and returns the bounds:
def bounds_fn(feat):
return (feat.xmin, feat.ymin, feat.xmax, feat.ymax)
r = new Rtree(feat_generator, bounds_getter=bounds_fn...)
then if the feat objects are just dicts, and not objects, they can be
loaded directly without creating wrapper
objects. another way would be to have feat_generator return tuples of
(feat, bounds).
_______________________________________________
Community mailing list
[email protected]
http://lists.gispython.org/mailman/listinfo/community
