I ended up making a shim between the two.
from numpy import *
from sympy import *
x = array([(1.1, 2.2), (3.3, 4.4), (5.5, 6.6)], dtype=[('s0', '<f8'),
('s1', '<f8')])
eq = 2.0 * Symbol('s0') - sqrt(Symbol('s1'))
from collections import Mapping
class FilterNdarray(Mapping):
def __init__(self, source, t):
self.source = source
self.t = t
def __getitem__(self, key):
return self.source[key.name][self.t]
def __len__(self):
return len(self.source)
def __iter__(self):
for key in self.source:
yield key
def set_time(self, t):
self.t = t
return self
y = FilterNdarray(x, 0)
[eq.evalf(subs=y.set_time(t)) for t in range(x.size)]
It's still horribly slow. I found someone who wrote an integrator, and they
have an evalf that compiles it down to bytecode somehow. I'm going to
pursue that, and it bears fruit, I'll post it as a sympy -> numpy evalf.
--
You received this message because you are subscribed to the Google Groups
"sympy" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at http://groups.google.com/group/sympy.
For more options, visit https://groups.google.com/groups/opt_out.
