Initial commit

.gitignore

+.mypy_cache

.gitlab-ci.yml

+image: python:3-buster
+
+before_script:
+  - pip install -r requirements-dev.txt
+
+test:
+    script:
+    - bandit *.py
+    - flake8 *.py
+    - mypy *.py
+    - pydocstyle *.py
+    - pylint *.py

README.md

+# Karaoke Cleaner
+
+Removes redundant files from a karaoke collection.
+
+A collection may include multiple versions of the same song, from multiple
+karaoke studios/labels. This script finds such songs, and moves the less-desired
+copies into a different directory.

karaokeclean.py

+#!/bin/env python3
+
+"""Karaoke collection sorting tool."""
+
+from collections import defaultdict
+import os
+import re
+import sys
+
+KARAOKE_DIR = "/media/anton/Karaoke/Karaoke"
+EXTRA_DIR = "/media/anton/Karaoke/Karaoke_Extra"
+
+FILENAME_REGEX = re.compile(r"(?P<song>.*)\[(?P<label>.*)\](?P<ext>.*)")
+
+PREFERENCES = [
+    "Karaoke",
+    "SC Karaoke",
+    "ME Karaoke",
+    "CB Karaoke",
+    "MH Karaoke",
+    "HK Karaoke",
+    "SN Karaoke",
+    "L Karaoke",
+    "Spanish Karaoke",
+    "SF Karaoke",
+    "MM Karaoke",
+    "EK Karaoke",
+    "AS Karaoke",
+    "KV Karaoke",
+    "PS Karaoke",
+    "VS Karaoke",
+    "Z Karaoke",
+    "DC Karaoke",
+]
+
+
+def karaoke_cleanup():
+    """Move redundant karaoke tracks into a separate directory."""
+    # Scan all songs
+    all_songs = defaultdict(list)
+    for filename in os.listdir(KARAOKE_DIR):
+        match = FILENAME_REGEX.search(filename)
+        if match:
+            all_songs[match.group("song")].append(filename)
+    songs_total = len(all_songs)
+    # Remove non-preferred labels whenever preferred exists.a
+    songs_done = 0
+    total_removed = 0
+    for files in all_songs.values():
+        dupes_removed = 0
+        for preferred_label in PREFERENCES:
+            preferred_string = "[{}]".format(preferred_label)
+            if any([preferred_string in filename for filename in files]):
+                # Found it; remove all others.
+                for filename in files:
+                    if preferred_string not in filename:
+                        _remove_duplicate(filename)
+                        dupes_removed += 1
+                        total_removed += 1
+                break
+        if (len(files) - dupes_removed) > 2:
+            _print_need_label_preference(files)
+        _print_progress(songs_done, songs_total, dupes_removed)
+        songs_done += 1
+    return (songs_done, songs_total, total_removed)
+
+
+def _remove_duplicate(filename):
+    old = os.path.join(KARAOKE_DIR, filename)
+    new = os.path.join(EXTRA_DIR, filename)
+    os.rename(old, new)
+
+
+def _print_progress(count_done, count_total, dupes):
+    percent = count_done / count_total * 100
+    sys.stdout.write("\033[K{:2.2f}% done. Removed {} duplicates."
+                     .format(percent, dupes))
+
+
+def _print_need_label_preference(files):
+    labels = []
+    filename = None
+    for filename in files:
+        match = FILENAME_REGEX.search(filename)
+        if match:
+            labels.append(match.group("label"))
+        else:
+            labels.append("")
+    labels = "\n".join(set(labels))
+    sys.stdout.write("\033[K")
+    print("Need label choice. Song: {}\n{}".format(filename, labels))
+    sys.exit()
+
+
+def count_songs_by_label():
+    """Return a list of Karaoke labels, sorted by number of tracks."""
+    all_labels = defaultdict(int)
+    for filename in os.listdir(KARAOKE_DIR):
+        match = FILENAME_REGEX.search(filename)
+        song_label = match.group("label") if match else ""
+        all_labels[song_label] += 1
+    return sorted(all_labels.items(), key=lambda x: x[1], reverse=True)
+
+
+if __name__ == "__main__":
+    if len(sys.argv) > 1 and sys.argv[1] == "getlabels":
+        for (label, count) in count_songs_by_label():
+            print(label, count)
+    else:
+        print("Done. Processed {} songs, removed {} extras."
+              .format(*karaoke_cleanup()[1:2]))

requirements-dev.txt

+bandit==1.6.2
+flake8==3.7.9
+mypy==0.761
+mypy-extensions==0.4.3
+pydocstyle==5.0.1
+pylint==2.4.4