fix data type

bsp2/bsp2.py

@@ -5,11 +5,12 @@
 """
 
 
-from typing import Dict, Tuple, Union
+from typing import Dict, List, Tuple, Union
 
 import numpy as np
-import pandas as pd
-from scipy.sparse import csr_matrix, diags, identity, isspmatrix_csr  # type: ignore
+import pandas as pd  # type: ignore
+from scipy.sparse import (csr_matrix, diags, identity,  # type: ignore
+                          isspmatrix_csr)
 from scipy.spatial import KDTree  # type: ignore
 from scipy.stats import gmean, lognorm  # type: ignore
 from sklearn.preprocessing import minmax_scale  # type: ignore
@@ -63,15 +64,15 @@ def _binary_distance_matrix_threshold(
     )
 
 
-def _spvars(input_csr_mat: csr_matrix, axis: int) -> np.ndarray:
+def _spvars(input_csr_mat: csr_matrix, axis: int) -> List[float]:
     input_csr_mat_squared = input_csr_mat.copy()
     input_csr_mat_squared.data **= 2
     return input_csr_mat_squared.mean(axis) - np.square(input_csr_mat.mean(axis))
 
 
 def _test_scores(
     input_sp_mat: np.ndarray, input_exp_mat_raw: csr_matrix, d1: float, d2: float
-) -> np.ndarray:
+) -> List[float]:
     input_exp_mat_norm = _scale_sparse_minmax(input_exp_mat_raw).transpose()
     input_exp_mat_raw = input_exp_mat_raw.transpose()
     inverted_diag_matrix_cache: Dict[Tuple, csr_matrix] = {}
@@ -86,7 +87,7 @@ def _get_inverted_diag_matrix(sum_axis_0: np.ndarray) -> csr_matrix:
             )
         return inverted_diag_matrix_cache[cache_key]
 
-    def _var_local_means(d_val: float) -> np.ndarray:
+    def _var_local_means(d_val: float) -> list:
         patches_cells = _binary_distance_matrix_threshold(input_sp_mat, d_val)
         patches_cells_centroid = diags(
             (patches_cells.sum(axis=1) > 1).astype(float).A.ravel(),
@@ -103,15 +104,15 @@ def _var_local_means(d_val: float) -> np.ndarray:
     var_x_0_add = _spvars(input_exp_mat_raw, axis=1).A.ravel()  # type: ignore
     var_x_0_add /= max(var_x_0_add)
     t_matrix = (var_x[:, 1] / var_x[:, 0]) * var_x_0_add
-    return t_matrix
+    return t_matrix.tolist()
 
 
 def granp(
     input_sp_mat: np.ndarray,
     input_exp_mat_raw: Union[np.ndarray, pd.DataFrame, csr_matrix],
     d1: float = 1.0,
     d2: float = 3.0,
-) -> np.ndarray:
+) -> List[float]:
     # Normalize patch size
     # Using gmean for geometric mean to scale d1 and d2 accordingly
     scale_factor = (
@@ -134,7 +135,7 @@ def granp(
 
     # Calculate p-values
     t_matrix_sum_upper90 = np.quantile(t_matrix_sum, 0.90)
-    t_matrix_sum_mid = t_matrix_sum[t_matrix_sum < t_matrix_sum_upper90]
+    t_matrix_sum_mid = [val for val in t_matrix_sum if val < t_matrix_sum_upper90]
     log_t_matrix_sum_mid = np.log(t_matrix_sum_mid)
     log_norm_params = (log_t_matrix_sum_mid.mean(), log_t_matrix_sum_mid.std(ddof=1))
 

test/test_bsp2.py

@@ -118,9 +118,9 @@ def test_non_empty_matrices(self):
         result = _test_scores(input_sp_mat, input_exp_mat_raw, d1, d2)
 
         # Check if the result is a numpy.ndarray
-        self.assertIsInstance(result, np.ndarray)
+        self.assertIsInstance(result, list)
         # Check the shape of the result
-        self.assertEqual(result.shape, (2,))  # Shape depends on your function's logic
+        self.assertEqual(len(result), 2)  # Shape depends on your function's logic
 
 
 class TestGranp(unittest.TestCase):