Initial commit

bundesdata_markup_nlp/samples/create_samples.py

@@ -0,0 +1,95 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+
+import os
+import fnmatch
+import argparse
+import random
+import shutil
+
+"""
+This is just a quick script to get randomized samples from the protocols.
+"""
+
+
+def parse_arguments():
+    """Argument Parser"""
+    parser = argparse.ArgumentParser(description="Creates samples from given   \
+                                     directory with given size. Creates two    \
+                                     samples with no overlapping.")
+    parser.add_argument("-p",
+                        "--path",
+                        help="Path to data files to create sample from.",
+                        required=True,
+                        type=str,
+                        metavar="")
+    parser.add_argument("-s",
+                        "--size",
+                        help="Size of sample.",
+                        required=True,
+                        type=int,
+                        metavar="")
+    parser.add_argument("-n", "--number_of_samples",
+                        help="How many smaples should be created? should be    \
+                        created?",
+                        required=True,
+                        type=int,
+                        metavar="")
+    parser.add_argument("-t",
+                        "--file_type",
+                        help="What file types should be used as the base for   \
+                        the sample?. Accepts wildcars.",
+                        required=True,
+                        type=str)
+    args = parser.parse_args()
+    return args
+
+
+def get_files(path, file_type):
+    """Creates file list with full paths of all files in the given directory and
+    its sub directories and returns it."""
+    list_of_files = []
+    for path, subdirs, files in os.walk(path):
+        for name in files:
+            if fnmatch.fnmatch(name, file_type):
+                list_of_files.append(os.path.join(path, name))
+    return list_of_files
+
+
+def get_files_to_copy(list_of_files, sample_size):
+    """Gets random filepaths from all filepaths to create a sample out of those.
+    Filepaths that have already been use will be removed from the file list to
+    create independent sampels."""
+    counter = 0
+    sample_list = []
+    while counter < sample_size:
+        counter += 1
+        random_index = random.randint(0, len(list_of_files)-1)
+        sample_list.append(list_of_files[random_index])
+        del list_of_files[random_index]
+        pass
+    return list_of_files, sample_list
+
+
+def copy_files(path, sample_list, step_int):
+    """Copys the given files to new directories."""
+    sample_path = os.path.join(path, str(step_int))
+    print(sample_path)
+    os.mkdir(sample_path)
+    for file in sample_list:
+        shutil.copy2(file, sample_path)
+
+
+def main():
+    args = parse_arguments()
+    path = args.path
+    file_list = get_files(path, args.file_type)
+    for step in range(1, args.number_of_samples + 1):
+        file_list = get_files_to_copy(file_list, args.size)[0]
+        sample_list = get_files_to_copy(file_list, args.size)[1]
+        copy_files(path, sample_list, step)
+        file_list = get_files_to_copy(file_list, args.size)[0]
+
+
+if __name__ == '__main__':
+    main()