fix pandas future warnings

Author: lorenzocerrone

fractal_tasks_core/cellvoyager/metadata.py

@@ -207,8 +207,8 @@ def read_mrf_file(mrf_path: str) -> tuple[pd.DataFrame, int]:
     meas_df = pd.read_xml(
         mrf_path, xpath="//bts:MeasurementDetail", namespaces=ns
     )
-    row_count = int(meas_df["RowCount"])
-    column_count = int(meas_df["ColumnCount"])
+    row_count = int(meas_df["RowCount"].iloc[0])
+    column_count = int(meas_df["ColumnCount"].iloc[0])
     plate_type = row_count * column_count
     return channel_df, plate_type
 
@@ -471,7 +471,7 @@ def check_group_consistency(grouped_df: pd.DataFrame, message: str = ""):
     # Check consistency in grouped df for multi-index, multi-column dataframes
     # raises an exception if there is variability
     diff_df = grouped_df.max() - grouped_df.min()
-    if not np.isclose(np.sum(np.sum(diff_df)), 0.0):
+    if not np.isclose(diff_df.to_numpy().sum(), 0.0):
         raise ValueError(
             "During metadata parsing, a consistency check failed: \n"
             f"{message}\n"

fractal_tasks_core/masked_loading.py

@@ -109,7 +109,9 @@ def _preprocess_input(
             'In _preprocess_input, "{column_name}" '
             f" missing in {ROI_table.obs.columns=}"
         )
-    label_value = int(float(ROI_table.obs[column_name][ROI_positional_index]))
+    label_value = int(
+        float(ROI_table.obs[column_name].iloc[ROI_positional_index])
+    )
 
     # Load masking-label array (lazily)
     masking_label_path = str(

fractal_tasks_core/roi/v1_overlaps.py

@@ -45,9 +45,9 @@ def get_overlapping_pair(
     num_lines = len(tmp_df.index)
     for pos_ind_1 in range(num_lines):
         for pos_ind_2 in range(pos_ind_1):
-            if is_overlapping_2D(
-                tmp_df.iloc[pos_ind_1], tmp_df.iloc[pos_ind_2], tol=tol
-            ):
+            bbox_1 = tmp_df.iloc[pos_ind_1].to_numpy()
+            bbox_2 = tmp_df.iloc[pos_ind_2].to_numpy()
+            if is_overlapping_2D(bbox_1, bbox_2, tol=tol):
                 return (pos_ind_1, pos_ind_2)
     return False
 
@@ -96,9 +96,9 @@ def get_overlapping_pairs_3D(
     overlapping_list = []
     for pos_ind_1 in range(num_lines):
         for pos_ind_2 in range(pos_ind_1):
-            overlap = is_overlapping_3D(
-                new_tmp_df.iloc[pos_ind_1], new_tmp_df.iloc[pos_ind_2], tol=tol
-            )
+            bbox_1 = new_tmp_df.iloc[pos_ind_1].to_numpy()
+            bbox_2 = new_tmp_df.iloc[pos_ind_2].to_numpy()
+            overlap = is_overlapping_3D(bbox_1, bbox_2, tol=tol)
             if overlap:
                 overlapping_list.append((pos_ind_1, pos_ind_2))
     return overlapping_list
@@ -177,7 +177,6 @@ def remove_FOV_overlaps(df: pd.DataFrame):
     # Loop over wells
     wells = sorted(list(set([ind[0] for ind in df.index])))
     for well in wells:
-
         logger.info(f"removing FOV overlaps for {well=}")
         df_well = df.loc[well].copy()
 

fractal_tasks_core/tasks/cellvoyager_to_ome_zarr_init.py

@@ -278,10 +278,10 @@ def cellvoyager_to_ome_zarr_init(
             site_metadata.set_index(["well_id", "FieldIndex"], inplace=True)
 
         # Extract pixel sizes and bit_depth
-        pixel_size_z = site_metadata["pixel_size_z"][0]
-        pixel_size_y = site_metadata["pixel_size_y"][0]
-        pixel_size_x = site_metadata["pixel_size_x"][0]
-        bit_depth = site_metadata["bit_depth"][0]
+        pixel_size_z = site_metadata["pixel_size_z"].iloc[0]
+        pixel_size_y = site_metadata["pixel_size_y"].iloc[0]
+        pixel_size_x = site_metadata["pixel_size_x"].iloc[0]
+        bit_depth = site_metadata["bit_depth"].iloc[0]
 
         if min(pixel_size_z, pixel_size_y, pixel_size_x) < 1e-9:
             raise ValueError(pixel_size_z, pixel_size_y, pixel_size_x)

fractal_tasks_core/tasks/cellvoyager_to_ome_zarr_init_multiplex.py

@@ -303,10 +303,10 @@ def cellvoyager_to_ome_zarr_init_multiplex(
             site_metadata.set_index(["well_id", "FieldIndex"], inplace=True)
 
         # Extract pixel sizes and bit_depth
-        pixel_size_z = site_metadata["pixel_size_z"][0]
-        pixel_size_y = site_metadata["pixel_size_y"][0]
-        pixel_size_x = site_metadata["pixel_size_x"][0]
-        bit_depth = site_metadata["bit_depth"][0]
+        pixel_size_z = site_metadata["pixel_size_z"].iloc[0]
+        pixel_size_y = site_metadata["pixel_size_y"].iloc[0]
+        pixel_size_x = site_metadata["pixel_size_x"].iloc[0]
+        bit_depth = site_metadata["bit_depth"].iloc[0]
 
         if min(pixel_size_z, pixel_size_y, pixel_size_x) < 1e-9:
             raise ValueError(pixel_size_z, pixel_size_y, pixel_size_x)

tests/test_unit_ROIs.py

@@ -135,7 +135,6 @@ def get_metadata_dataframe():
     "level,coarsening_xy,full_res_pxl_sizes_zyx", list_params
 )
 def test_ROI_indices_3D(level, coarsening_xy, full_res_pxl_sizes_zyx):
-
     metadata_dataframe = get_metadata_dataframe()
     adata = prepare_FOV_ROI_table(metadata_dataframe)
     assert list(adata.obs_names) == FOV_NAMES
@@ -193,7 +192,6 @@ def test_ROI_indices_3D(level, coarsening_xy, full_res_pxl_sizes_zyx):
     "level,coarsening_xy,full_res_pxl_sizes_zyx", list_params
 )
 def test_ROI_indices_2D(level, coarsening_xy, full_res_pxl_sizes_zyx):
-
     metadata_dataframe = get_metadata_dataframe()
     adata = prepare_FOV_ROI_table(metadata_dataframe)
     adata = convert_ROIs_from_3D_to_2D(adata, PIXEL_SIZE_Z)
@@ -238,7 +236,6 @@ def test_prepare_well_ROI_table(testdata_path: Path):
 
 
 def test_overlaps_in_indices():
-
     list_indices = [
         [0, 1, 100, 200, 1000, 2000],
         [0, 1, 200, 300, 1000, 2000],
@@ -478,7 +475,6 @@ def test_load_region_fail():
 
 
 def test_is_ROI_table_valid(tmp_path):
-
     # Write valid table to a zarr group
     columns = EXPECTED_COLUMNS.copy()
     adata = ad.AnnData(np.ones((1, len(columns))))
@@ -576,9 +572,9 @@ def test_search_first_ROI(testdata_path: Path):
         )
     )
     full_res_pxl_sizes_zyx = [
-        big_df["pixel_size_z"][0],
-        big_df["pixel_size_y"][0],
-        big_df["pixel_size_x"][0],
+        big_df["pixel_size_z"].iloc[0],
+        big_df["pixel_size_y"].iloc[0],
+        big_df["pixel_size_x"].iloc[0],
     ]
 
     well_ids = big_df.well_id.unique()