test

OptiMap/pairwise.py

@@ -29,7 +29,14 @@ def section_vs_section(section1, section2, fft_molecules, fft_rev_molecules, max
         print(i)
         subject_molecules = fft_molecules[i:i+width]
         rev_subject_molecules = fft_rev_molecules[i:i+width]
-        multiple_top_corrs = get_multiple_products(subject_molecules, rev_subject_molecules, db_molecules)
+
+        multiple_corr_maxes = np.zeros((subject_molecules.shape[0], db_molecules.shape[0]), dtype=float)
+        fft_products1 = np.zeros((db_molecules.shape[0],db_molecules.shape[1]), dtype=complex)
+        fft_products2 = np.zeros((db_molecules.shape[0],db_molecules.shape[1]), dtype=complex)
+        corr_maxes = np.zeros(fft_molecules.shape[0])
+
+        multiple_top_corrs = get_multiple_products(subject_molecules, rev_subject_molecules, db_molecules, multiple_corr_maxes,
+                                                   fft_products1, fft_products2, corr_maxes)
         mol_range = np.arange(i, i + width)
         normalized_top_corrs = np.zeros(multiple_top_corrs.shape)
         numba_normalize_molecule_correlation_array(multiple_top_corrs, maxes, section2, mol_range, normalized_top_corrs)
@@ -50,21 +57,16 @@ def create_and_link_paired_matrices(period, fft_molecules, fft_rev_molecules, ma
     return merge_and_extend_pairs(pair_sets, depth=depth)
 
 
-# @nb.jit
-def get_multiple_products(fft_subject_molecules, fft_subject_rev_molecules, fft_molecules):
-    multiple_corr_maxes = np.zeros((fft_subject_molecules.shape[0], fft_molecules.shape[0]), dtype=float)
-    # fft_products = np.zeros((2,fft_molecules.shape[0],fft_molecules.shape[1]), dtype=complex)
-    fft_products1 = np.zeros((fft_molecules.shape[0],fft_molecules.shape[1]), dtype=complex)
-    fft_products2 = np.zeros((fft_molecules.shape[0],fft_molecules.shape[1]), dtype=complex)
-    # corr_products = np.zeros((2,fft_molecules.shape[0],fft_molecules.shape[1]), dtype=float)
-    corr_maxes = np.zeros(fft_molecules.shape[0])
+@nb.jit
+def get_multiple_products(fft_subject_molecules, fft_subject_rev_molecules, fft_molecules, multiple_corr_maxes, fft_products1, fft_products2, corr_maxes):
+    # multiple_corr_maxes = np.zeros((fft_subject_molecules.shape[0], fft_molecules.shape[0]), dtype=float)
+    # fft_products1 = np.zeros((fft_molecules.shape[0],fft_molecules.shape[1]), dtype=complex)
+    # fft_products2 = np.zeros((fft_molecules.shape[0],fft_molecules.shape[1]), dtype=complex)
+    # corr_maxes = np.zeros(fft_molecules.shape[0])
     for i in range(fft_subject_molecules.shape[0]):
-        print("running numba_get_products")
+        # print("running numba_get_products")
         numba_get_products(fft_subject_molecules[i], fft_subject_rev_molecules[i], fft_molecules, fft_products1, fft_products2)
-        # print("running ifft")
-        # corr_products[0] = np.fft.ifft(fft_products[0]).real
-        # corr_products[1] = np.fft.ifft(fft_products[1]).real
-        print("running get_corr_maxes and ifft")
+        # print("running get_corr_maxes and ifft")
         numba_get_corr_maxes(np.fft.ifft(fft_products1).real, np.fft.ifft(fft_products2).real, corr_maxes)
         multiple_corr_maxes[i,:] = corr_maxes
     return multiple_corr_maxes