reorganizing

stashtoolkit/Transcoded File Scripts/RMF Transcoded File Scripts/Find duplicates/findanddelete.py

@@ -0,0 +1,70 @@
+import os
+import csv
+import shutil
+from collections import defaultdict
+
+def find_duplicate_basenames(root_directory, valid_extensions):
+    # Dictionary to store basenames and their corresponding file paths
+    basenames_to_paths = defaultdict(list)
+
+    # Traverse the directory tree
+    for dirpath, dirnames, filenames in os.walk(root_directory):
+        for filename in filenames:
+            full_path = os.path.join(dirpath, filename)
+            basename, extension = os.path.splitext(filename)
+            if extension.lower() in valid_extensions:
+                basenames_to_paths[basename].append(full_path)
+
+    # Filter basenames with more than one file
+    duplicate_basenames = {basename: paths for basename, paths in basenames_to_paths.items() if len(paths) > 1}
+
+    return duplicate_basenames
+
+def get_file_size(path):
+    try:
+        return os.path.getsize(path)
+    except OSError:
+        return 0
+
+def display_parent_levels(path, levels=2):
+    # Display 'levels' parent directories above the file path
+    components = path.split(os.sep)
+    return os.sep.join(components[-levels:])
+
+def prompt_user_for_selection(paths):
+    print("Choose which file to keep:")
+    for i, path in enumerate(paths, start=1):
+        size_mb = get_file_size(path) / (1024 * 1024)
+        print(f"{i}. {display_parent_levels(path)} ({size_mb:.2f} MB)")
+
+    while True:
+        try:
+            choice = int(input("Enter the number of the file to keep (or 0 to skip): "))
+            if 0 <= choice <= len(paths):
+                return choice
+            else:
+                print("Invalid choice. Please enter a valid number.")
+        except ValueError:
+            print("Invalid input. Please enter a number.")
+
+if __name__ == "__main__":
+    server_directory = r"\\truenas\isolation\videos"  # Replace with your server directory
+    video_extensions = {".mp4", ".mkv", ".avi", ".mov"}  # Add more video extensions if needed
+    duplicate_files = find_duplicate_basenames(server_directory, video_extensions)
+
+    if duplicate_files:
+        for basename, paths in duplicate_files.items():
+            print(f"\nBasename: {basename}")
+            choice = prompt_user_for_selection(paths)
+            if choice == 0:
+                print("Skipping this duplicate set.")
+            else:
+                # Keep the selected file and delete the others
+                keep_path = paths[choice - 1]
+                for path in paths:
+                    if path != keep_path:
+                        print(f"Deleting: {display_parent_levels(path)}")
+                        os.remove(path)
+        print("\nDuplicate files processed.")
+    else:
+        print("No duplicate video files found.")