Merge branch 'develop'

Author: vmarkovtsev

python.cc

@@ -152,7 +152,7 @@ static PyObject *py_minhash_cuda_retrieve_vars(PyObject *self, PyObject *args) {
     return NULL;
   }
   auto params = mhcuda_get_parameters(gen);
-  npy_intp dims[] = {params.dim, params.samples, 0};
+  npy_intp dims[] = {params.samples, params.dim, 0};
   auto rs_obj = reinterpret_cast<PyArrayObject *>(PyArray_EMPTY(
       2, dims, NPY_FLOAT32, false));
   auto ln_cs_obj = reinterpret_cast<PyArrayObject *>(PyArray_EMPTY(

setup.py

@@ -46,7 +46,7 @@ def is_pure(self):
 setup(
     name="libMHCUDA",
     description="Accelerated Weighted MinHash-ing on GPU",
-    version="1.1.4",
+    version="1.1.5",
     license="MIT",
     author="Vadim Markovtsev",
     author_email="[email protected]",

test.py

@@ -73,6 +73,9 @@ def test_random_vars(self):
         gen = libMHCUDA.minhash_cuda_init(1000, 128, devices=1, verbosity=2)
         rs, ln_cs, betas = libMHCUDA.minhash_cuda_retrieve_vars(gen)
         libMHCUDA.minhash_cuda_fini(gen)
+        self.assertEqual(rs.shape, (128, 1000))
+        self.assertEqual(ln_cs.shape, (128, 1000))
+        self.assertEqual(betas.shape, (128, 1000))
         cs = numpy.exp(ln_cs)
         a, loc, scale = gamma.fit(rs)
         self.assertTrue(1.97 < a < 2.03)
@@ -120,6 +123,34 @@ def test_slice(self):
         libMHCUDA.minhash_cuda_fini(gen)
         self.assertTrue((hashes[3200:4800] == hashes2).all())
 
+    def test_backwards(self):
+        v1 = [1, 0, 0, 0, 3, 4, 5, 0, 0, 0, 0, 6, 7, 8, 0, 0, 0, 0, 0, 0, 9, 10, 4]
+        v2 = [2, 0, 0, 0, 4, 3, 8, 0, 0, 0, 0, 4, 7, 10, 0, 0, 0, 0, 0, 0, 9, 0, 0]
+        gen = libMHCUDA.minhash_cuda_init(len(v1), 128, devices=1, verbosity=2)
+        rs, ln_cs, betas = libMHCUDA.minhash_cuda_retrieve_vars(gen)
+        bgen = WeightedMinHashGenerator.__new__(WeightedMinHashGenerator)
+        bgen.dim = len(v1)
+        bgen.rs = rs
+        bgen.ln_cs = ln_cs
+        bgen.betas = betas
+        bgen.sample_size = 128
+        bgen.seed = None
+        m = csr_matrix(numpy.array([v1, v2], dtype=numpy.float32))
+        hashes = libMHCUDA.minhash_cuda_calc(gen, m)
+        libMHCUDA.minhash_cuda_fini(gen)
+        self.assertEqual(hashes.shape, (2, 128, 2))
+        true_hashes = numpy.array([bgen.minhash(v1).hashvalues,
+                                   bgen.minhash(v2).hashvalues], dtype=numpy.uint32)
+        self.assertEqual(true_hashes.shape, (2, 128, 2))
+        try:
+            self.assertTrue((hashes == true_hashes).all())
+        except AssertionError as e:
+            print("---- TRUE ----")
+            print(true_hashes)
+            print("---- FALSE ----")
+            print(hashes)
+            raise e from None
+
 
 if __name__ == "__main__":
     unittest.main()