$ python setup.py build_ext --inplace
$ rm demo.c # this is the intermediary C file
$ python -c "import demo; print(demo.foo(2.2)); import numpy as np; x=np.arange(10).astype(np.float32); print(x); demo.scale(x,120.75); print(x)"
# ...
4.40000009537
[ 0. 1. 2. 3. 4. 5. 6. 7. 8. 9.]
[ 0. 120.75 241.5 362.25 483. 603.75 724.5 845.25
966. 1086.75]
$ python setup.py build_ext --inplace
Then, in Python:
import pyldpc_generate
H = pyldpc_generate.generate(512,1024,3.0,2,123)H is the sparse array of interest.
The following test requires MATLAB with py Python bridge and a C compiler that MEX understands.
$ python setup.py build_ext --inplace
$ rm demo.c ldpc_generate.c
$ cd mex && matlab -r "mex -largeArrayDims ldpc_generate.c; quit"
# make sure you quit matlab
$ cd ..
$ matlab -r "test_ldpc_generate; quit"
# ...
Tests passed!
If you really want to exercise the C library from a C main,
gcc -o gen.o ldpc_generate1.c gen_test.c && gen.o > gen.py
Confirm that this generates the same CSC sparse matrix as the Cython module wrapping the same library.