splits dataframe into multiple numpy arrays

src/nuclai/preprocessing/split_classic_feature_matrix.py

@@ -0,0 +1,88 @@
+########################################################################################################################
+# This script splits a feature DataFrame into single sample numpy arrays.                                              #
+# Author:               Daniel Schirmacher                                                                             #
+#                       PhD Student, Cell Systems Dynamics Group, D-BSSE, ETH Zurich                                   #
+# Python:               3.12.7                                                                                         #
+########################################################################################################################
+import argparse
+import os
+
+import numpy as np
+import pandas as pd
+from tqdm import tqdm
+
+
+def arg_parse():
+    """
+    Catch user input.
+
+
+    Parameter
+    ---------
+
+    -
+
+
+    Return
+    ------
+
+    Returns a namespace from `argparse.parse_args()`.
+    """
+    desc = (
+        "Program to split a feature dataframe into single sample numpy arrays."
+    )
+    parser = argparse.ArgumentParser(description=desc)
+
+    parser.add_argument(
+        "--features",
+        type=str,
+        default=r"N:\schroeder\Data\DS\PhD\nucleus_classification\data\3d\images\segmentation\mouse\cd41\rep4\subset\7_classic_features\classic_features_3D_c0_0-68_1000-2400_3100-4500.csv",
+        help="Path to CSV file containing regionprops_table dataframe.",
+    )
+
+    parser.add_argument(
+        "--out",
+        type=str,
+        default=r"N:\schroeder\Data\DS\PhD\nucleus_classification\data\3d\images\segmentation\mouse\cd41\rep4\subset\7_classic_features\split_1000",
+        help="Path to output directory.",
+    )
+
+    parser.add_argument(
+        "--prefix",
+        type=str,
+        default="cd41_val1000",
+        help="Prefix for output file names. Will be used as '{prefix}_{mask_id}_cfeats.npy'.",
+    )
+
+    return parser.parse_args()
+
+
+def main():
+    args = arg_parse()
+
+    path_f = args.features
+    prefix = args.prefix
+    path_out = args.out
+
+    # load data
+    df = pd.read_csv(path_f)
+
+    # create output directory
+    os.makedirs(path_out, exist_ok=True)
+
+    # loop over each mask and save npy file
+    mask_ids = df.pop("mask_id")
+
+    with tqdm(total=len(mask_ids)) as pbar:
+        for i, mask_id in enumerate(mask_ids):
+            row = df.iloc[i].to_numpy().reshape(1, -1)
+            path_out_i = os.path.join(
+                path_out, f"{prefix}_{mask_id}_cfeats.npy"
+            )
+            np.save(path_out_i, row)
+
+            pbar.update(1)
+
+
+if __name__ == "__main__":
+    main()