According to the GSL documentation, the signature of is:
double gsl_stats_correlation (const double data1[],
const size_t stride1,
const double data2[],
const size_t stride2,
const size_t n)
When I try to call it from PyPy with:
from ctypes import CDLL, RTLD_GLOBAL
gslcblas = CDLL('libgslcblas.0.dylib',mode=RTLD_GLOBAL)
libgsl = CDLL('/usr/local/lib/libgsl.0.dylib')
from ctypes import c_double, c_size_t, pointer
a1 = (c_double * 5)(1, 2, 3, 4, 5)
a2 = (c_double * 5)(1, 2, 3, 6, 5)
print(libgsl.gsl_stats_correlation(a1, c_size_t(1),
a2, c_size_t(1), c_size_t(5)))
The result on my machine is 1086463496 at the moment, although it changes from run to run. That obviously is far from correct. What am I doing wrong? Note that changing the function call to:
libgsl.gsl_stats_correlation(pointer(a1), c_size_t(1),
pointer(a2), c_size_t(1), c_size_t(5)))
gives exactly the same result.
You need to set the
restype
of the function like this:Have a look at this ctypes tutorial (particularly this section) to read about specifying argument and response types properly.