add script “clean-tags.py”

clean-tags.py

@@ -0,0 +1,165 @@
+#!/usr/bin/env python3
+# -*- coding: utf-8 -*-
+
+
+# Dependencies:
+# – python-mutagen
+
+
+import argparse
+import logging
+import os
+
+from mutagen.id3 import ID3
+from mutagen import MutagenError
+
+
+
+
+class Styling:
+    """Output styling constants."""
+    ENDC = '\x1b[0m'
+    BOLD = '\x1b[1m'
+
+
+
+
+class TagCleaner:
+    """
+    Class to check audio files for unwanted tags and provide routines to clean
+    them.
+    """
+    INDENTATION = "    "
+    FILE_EXTENSIONS = ['.mp3']
+    TAGS_ID3 = ['TALB', 'TDRC', 'TIT2', 'TPE1', 'TRCK']
+
+
+    def __init__(self, folder):
+        """Construct a new tag cleaner instancce."""
+        self._logger = logging.getLogger(__class__.__name__)
+        self._dry = False
+        self._folder = folder
+        self._files = []
+        self._read_files(folder)
+
+
+    def set_dry(self, dry):
+        """Set dry mode (do not save any file)."""
+        self._dry = dry
+
+
+    def clean_files(self):
+        """Clean all found files."""
+        self._logger.info("Clean files")
+        for filename in self._files:
+            self._clean_file(filename)
+
+
+    def _read_files(self, folder):
+        self._logger.info("Read files from \"{}\"".format(folder))
+        for dirname, dirnames, filenames in os.walk(folder):
+            for filename in filenames:
+                if filename.startswith("."):
+                    continue
+                (name, ext) = os.path.splitext(filename)
+                if ext.lower() not in TagCleaner.FILE_EXTENSIONS:
+                    continue
+                self._logger.debug("Found file \"%s\"", os.path.join(dirname, filename))
+                self._files.append(os.path.join(dirname, filename))
+        self._logger.info("Found %d files".format(len(self._files)))
+
+
+    def _clean_file(self, filename):
+        """Clean a file."""
+        self._logger.info("Clean file \"%s\"", filename)
+        if os.path.isfile(filename):
+            # ID3
+            self._clean_id3(filename)
+        else:
+            self._logger.info("Not a file: \"%s\"", filename)
+
+
+    def _clean_id3(self, filename):
+        """Clean ID3 tags."""
+        self._logger.info("Clean ID3")
+        try:
+            tags = ID3(filename)
+            print(Styling.BOLD + filename[len(self._folder):] + Styling.ENDC)
+            print("ID3", "v{}.{}.{}".format(*tags.version))
+            valid = True
+
+            # Check version
+            if tags.version != (2, 4, 0):
+                valid = False
+
+            # Unknown tags
+            if tags.unknown_frames:
+                valid = False
+                print("Unknown frames:")
+                for frame in tags.unknown_frames:
+                    print(frame)
+
+            # Invalid tags
+            invalid_tags = []
+            for tag in tags:
+                if len(tag) > 4:
+                    tag = tag[0:4]
+                if tag not in TagCleaner.TAGS_ID3:
+                    invalid_tags.append(tag)
+            if invalid_tags:
+                valid = False
+                print("Unwanted tags:")
+                for tag in invalid_tags:
+                    for frame in tags.getall(tag):
+                        if hasattr(frame, 'text'):
+                            for value in frame.text:
+                                print("{}{}: {}".format(TagCleaner.INDENTATION, tag, value))
+                        else:
+                            print("{}{}:".format(TagCleaner.INDENTATION, tag), frame)
+
+            # Save
+            if not valid:
+                # Delete tags
+                for tag in invalid_tags:
+                    print("Delete", tag)
+                    tags.delall(tag)
+
+                    # Save file
+                    if not self._dry:
+                        try:
+                            tags.save()
+                            print("File saved")
+                        except Exception as e:
+                            self._logger.error("Saving of file \"%s\" failed: %s", filename, e)
+                    else:
+                        print("File not saved (running in dry mode)")
+            else:
+                self._logger.info("Clean, nothing to do")
+        except MutagenError as e:
+            self._logger.error("Cleaning of ID3 tags failed: %s", e)
+
+
+
+
+if __name__ == "__main__":
+    # Setup command line
+    parser = argparse.ArgumentParser("Clean unwanted tags from audio files to keep your music library clean.")
+    parser.add_argument('-d', '--dry', dest='dry', action='store_true', default=False, help="dry run, do not modify any file, just print information")
+    parser.add_argument('-v', '--verbose', dest='verbosity', action='count', default=0, help="be verbose, show more information")
+    parser.add_argument('-l', '--logfile', dest='logfile', help="specify name of logfile")
+    parser.add_argument('folder', help="source folder to read audio files from")
+    args = parser.parse_args()
+
+    # Setup logging
+    logging.basicConfig(
+        filename=args.logfile,
+        level=logging.ERROR-(10*args.verbosity),
+        format="%(asctime)s %(levelname)s: %(message)s"
+    )
+
+    # Create tag cleaner instance
+    tag_cleaner = TagCleaner(args.folder)
+    tag_cleaner.set_dry(args.dry)
+
+    # Run action
+    tag_cleaner.clean_files()