year up, restricted coverage, nested extensions

Author: rusty1s

.coveragerc

@@ -1,3 +1,5 @@
+[run]
+source=torch_sparse
 [report]
 exclude_lines =
     pragma: no cover

LICENSE

@@ -1,4 +1,4 @@
-Copyright (c) 2018 Matthias Fey <[email protected]>
+Copyright (c) 2019 Matthias Fey <[email protected]>
 
 Permission is hereby granted, free of charge, to any person obtaining a copy
 of this software and associated documentation files (the "Software"), to deal

cuda/spspmm.cpp

@@ -1,4 +1,4 @@
-#include <torch/torch.h>
+#include <torch/extension.h>
 
 #define CHECK_CUDA(x) AT_ASSERTM(x.type().is_cuda(), #x " must be CUDA tensor")
 

cuda/unique.cpp

@@ -1,4 +1,4 @@
-#include <torch/torch.h>
+#include <torch/extension.h>
 
 #define CHECK_CUDA(x) AT_ASSERTM(x.type().is_cuda(), #x " must be CUDA tensor")
 

setup.py

@@ -13,24 +13,25 @@
 
 if CUDA_HOME is not None:
     if platform.system() == 'Windows':
-        extra_link_args = 'cusparse.lib'
+        extra_link_args = ['cusparse.lib']
     else:
         extra_link_args = ['-lcusparse', '-l', 'cusparse']
 
     ext_modules += [
         CUDAExtension(
-            'spspmm_cuda', ['cuda/spspmm.cpp', 'cuda/spspmm_kernel.cu'],
+            'torch_sparse.spspmm_cuda',
+            ['cuda/spspmm.cpp', 'cuda/spspmm_kernel.cu'],
             extra_link_args=extra_link_args),
-        CUDAExtension('unique_cuda',
+        CUDAExtension('torch_sparse.unique_cuda',
                       ['cuda/unique.cpp', 'cuda/unique_kernel.cu']),
     ]
     cmdclass['build_ext'] = BuildExtension
 
 setup(
     name='torch_sparse',
     version=__version__,
-    description='PyTorch Extension Library of Optimized Autograd Sparse '
-    'Matrix Operations',
+    description=('PyTorch Extension Library of Optimized Autograd Sparse '
+                 'Matrix Operations'),
     author='Matthias Fey',
     author_email='[email protected]',
     url=url,
@@ -41,4 +42,5 @@
     tests_require=tests_require,
     ext_modules=ext_modules,
     cmdclass=cmdclass,
-    packages=find_packages(), )
+    packages=find_packages(),
+)

torch_sparse/spspmm.py

@@ -4,7 +4,7 @@
 from torch_sparse import transpose
 
 if torch.cuda.is_available():
-    import spspmm_cuda
+    import torch_sparse.spspmm_cuda
 
 
 def spspmm(indexA, valueA, indexB, valueB, m, k, n):
@@ -60,7 +60,8 @@ def mm(indexA, valueA, indexB, valueB, m, k, n):
     assert valueA.dtype == valueB.dtype
 
     if indexA.is_cuda:
-        return spspmm_cuda.spspmm(indexA, valueA, indexB, valueB, m, k, n)
+        return torch_sparse.spspmm_cuda.spspmm(indexA, valueA, indexB, valueB,
+                                               m, k, n)
 
     A = to_scipy(indexA, valueA, m, k)
     B = to_scipy(indexB, valueB, k, n)

torch_sparse/utils/unique.py

@@ -2,14 +2,14 @@
 import numpy as np
 
 if torch.cuda.is_available():
-    import unique_cuda
+    import torch_sparse.unique_cuda
 
 
 def unique(src):
     src = src.contiguous().view(-1)
 
     if src.is_cuda:
-        out, perm = unique_cuda.unique(src)
+        out, perm = torch_sparse.unique_cuda.unique(src)
     else:
         out, perm = np.unique(src.numpy(), return_index=True)
         out, perm = torch.from_numpy(out), torch.from_numpy(perm)