Reorganize and clean up helper functions for kernel testing

src/kbmod/search/bindings.cpp

@@ -15,9 +15,10 @@ namespace py = pybind11;
 #include "image_stack.cpp"
 #include "stack_search.cpp"
 #include "stamp_creator.cpp"
-#include "filtering.cpp"
+#include "kernel_testing_helpers.cpp"
 #include "psi_phi_array.cpp"
 
+
 PYBIND11_MODULE(search, m) {
     m.attr("KB_NO_DATA") = pybind11::float_(search::NO_DATA);
     m.attr("HAS_GPU") = pybind11::bool_(search::HAVE_GPU);
@@ -45,7 +46,6 @@ PYBIND11_MODULE(search, m) {
     m.def("create_median_image", &search::create_median_image);
     m.def("create_summed_image", &search::create_summed_image);
     m.def("create_mean_image", &search::create_mean_image);
-    // Functions from filtering.cpp
+    // Functions from kernel_testing_helpers.cpp
     m.def("sigmag_filtered_indices", &search::sigmaGFilteredIndices);
-    m.def("calculate_likelihood_psi_phi", &search::calculateLikelihoodFromPsiPhi);
 }

src/kbmod/search/kernel_testing_helpers.cpp

@@ -0,0 +1,43 @@
+/* Helper functions for testing functions in the .cu files from Python. */
+
+#include <vector>
+
+namespace search {
+#ifdef HAVE_CUDA
+/* The filter_kenerls.cu functions. */
+extern "C" void SigmaGFilteredIndicesCU(float *values, int num_values, float sgl0, float sgl1, float sg_coeff,
+                                        float width, int *idx_array, int *min_keep_idx, int *max_keep_idx);
+#endif
+
+/* Used for testing SigmaGFilteredIndicesCU for python
+ *
+ * Return the list of indices from the values array such that those elements
+ * pass the sigmaG filtering defined by percentiles [sgl0, sgl1] with coefficient
+ * sigma_g_coeff and a multiplicative factor of width.
+ *
+ * The vector values is passed by value to create a local copy which will be modified by
+ * SigmaGFilteredIndicesCU.
+ */
+std::vector<int> sigmaGFilteredIndices(std::vector<float> values, float sgl0, float sgl1, float sigma_g_coeff,
+                                       float width) {
+    int num_values = values.size();
+    std::vector<int> idx_array(num_values, 0);
+    int min_keep_idx = 0;
+    int max_keep_idx = num_values - 1;
+
+#ifdef HAVE_CUDA
+    SigmaGFilteredIndicesCU(values.data(), num_values, sgl0, sgl1, sigma_g_coeff, width, idx_array.data(),
+                            &min_keep_idx, &max_keep_idx);
+#else
+    throw std::runtime_error("Non-GPU SigmaGFilteredIndicesCU is not implemented.");
+#endif
+
+    // Copy the result into a vector and return it.
+    std::vector<int> result;
+    for (int i = min_keep_idx; i <= max_keep_idx; ++i) {
+        result.push_back(idx_array[i]);
+    }
+    return result;
+}
+
+} /* namespace search */

tests/test_filtering.py

@@ -98,29 +98,6 @@ def test_sigmag_filtered_indices_three_outliers(self):
             valid = i != 13 and i != 14 and i != 27
             self.assertEqual(i in inds, valid)
 
-    def test_calculate_likelihood_psiphi(self):
-        # make sure that the calculate_likelihood_psi_phi works.
-        psi_values = [1.0 for _ in range(20)]
-        phi_values = [1.0 for _ in range(20)]
-
-        lh = calculate_likelihood_psi_phi(psi_values, phi_values)
-
-        self.assertEqual(lh, 4.47213595499958)
-
-    def test_calculate_likelihood_psiphi_zero_or_negative_phi(self):
-        # make sure that the calculate_likelihood_psi_phi works
-        # properly when phi values are less than or equal to zero.
-        psi_values = [1.0 for _ in range(20)]
-        phi_values = [-1.0 for _ in range(20)]
-
-        # test negatives
-        lh = calculate_likelihood_psi_phi(psi_values, phi_values)
-        self.assertEqual(lh, 0.0)
-
-        # test zero
-        lh = calculate_likelihood_psi_phi([1.0], [0.0])
-        self.assertEqual(lh, 0.0)
-
 
 if __name__ == "__main__":
     unittest.main()