dotfiles/hm/scripts/updateCompressedMusic

#!/usr/bin/env nix-shell
#! nix-shell -i python3 --pure
#! nix-shell -p python3 python3Packages.coloredlogs python3Packages.progressbar2 ffmpeg
# pylint: disable=C0103

import logging
import os
import re
import subprocess
import typing

import coloredlogs
import progressbar

coloredlogs.install(level="DEBUG", fmt="%(levelname)s %(message)s")
log = logging.getLogger()

# Constants
SOURCE_FOLDER = os.path.join(os.path.expanduser("~"), "Musiques")
OUTPUT_FOLDER = os.path.join(os.path.expanduser("~"), ".MusiqueCompressed")
CONVERSIONS = {"flac": "opus"}
FORBIDDEN_EXTENSIONS = ["jpg", "png", "pdf", "ffs_db"]
FORGIVEN_FILENAMES = [
    "cover.jpg",
    "front.jpg",
    "folder.jpg",
    "cover.png",
    "front.png",
    "folder.png",
]
IGNORED_EMPTY_FOLDER = [".stfolder"]
RESTRICT_CHARACTERS = '[\0\\/:*"<>|]'  # FAT32, NTFS
# RESTRICT_CHARACTERS = '[:/]'  # HFS, HFS+
# RESTRICT_CHARACTERS = '[\0/]'  # ext2-4, linux-based?
act = True

# TODO FEAT Make the directory structure the same as the base one and
# remove IGNORED_EMPTY_FOLDER variable

# Listing files
log.info("Listing files")
sourceFiles = dict()
for root, dirs, files in os.walk(SOURCE_FOLDER):
    for f in files:
        fullPath = os.path.join(root, f)
        path = os.path.relpath(fullPath, SOURCE_FOLDER)
        sourceFiles[path] = os.path.getctime(fullPath)

outputFiles = dict()
for root, dirs, files in os.walk(OUTPUT_FOLDER):
    for f in files:
        fullPath = os.path.join(root, f)
        path = os.path.relpath(fullPath, OUTPUT_FOLDER)
        outputFiles[path] = os.path.getctime(fullPath)

# Sorting files
remainingConversions = dict()
extraFiles = set(outputFiles.keys())


def convertPath(path: str) -> typing.Optional[str]:
    filename, extension = os.path.splitext(path)
    extension = extension[1:].lower()
    # Remove unwanted characters from filename
    filename_parts = os.path.normpath(filename).split(os.path.sep)
    filename_parts = [re.sub(RESTRICT_CHARACTERS, "_", part) for part in filename_parts]
    filename = os.path.sep.join(filename_parts)
    # If the extension isn't allowed
    if extension in FORBIDDEN_EXTENSIONS:
        basename = os.path.basename(path)
        # And the filename is not an exception
        if basename not in FORGIVEN_FILENAMES:
            # This file shouldn't be copied nor converted
            return None
    # If this needs a conversion
    elif extension in CONVERSIONS:
        extension = CONVERSIONS[extension]
        return filename + "." + extension
    # In all other case, this is a simple copy
    return path


log.info("Determining action over %d files", len(sourceFiles))
for sourceFile in sourceFiles:
    outputFile = convertPath(sourceFile)
    # If the file should not be converted, do nothing
    if not outputFile:
        continue
    # If the file already has something as an output
    elif outputFile in outputFiles:
        extraFiles.remove(outputFile)
        # If the output file is newer than the source file, do not initiate a
        # conversion
        if outputFiles[outputFile] >= sourceFiles[sourceFile]:
            continue
    # If the file needs to be converted, do it
    remainingConversions[sourceFile] = outputFile

log.debug("%d actions will need to be taken", len(remainingConversions))
log.info("Copying files that do not require a conversion")
conversions = set()
for sourceFile in remainingConversions:
    outputFile = remainingConversions[sourceFile]

    # Creating folder if it doesn't exists
    fullOutputFile = os.path.join(OUTPUT_FOLDER, outputFile)
    fullOutputDir = os.path.dirname(fullOutputFile)
    if act:
        os.makedirs(fullOutputDir, exist_ok=True)

    # Converting
    fullSourceFile = os.path.join(SOURCE_FOLDER, sourceFile)
    if sourceFile == outputFile:
        log.debug("%s → %s", fullSourceFile, fullOutputFile)
        if act and os.path.isfile(fullOutputFile):
            os.remove(fullOutputFile)
        os.link(fullSourceFile, fullOutputFile)
    else:
        conversions.add((fullSourceFile, fullOutputFile))

log.info("Removing extra files")
for extraFile in extraFiles:
    fullExtraFile = os.path.join(OUTPUT_FOLDER, extraFile)
    log.debug("× %s", fullExtraFile)
    if act:
        os.remove(fullExtraFile)

log.info("Listing files that will be converted")
for fullSourceFile, fullOutputFile in conversions:
    log.debug("%s ⇒ %s", fullSourceFile, fullOutputFile)

log.info("Converting files")
for fullSourceFile, fullOutputFile in progressbar.progressbar(conversions):
    cmd = [
        "ffmpeg",
        "-y",
        "-i",
        fullSourceFile,
        "-c:a",
        "libopus",
        "-movflags",
        "+faststart",
        "-b:a",
        "128k",
        "-vbr",
        "on",
        "-compression_level",
        "10",
        fullOutputFile,
    ]
    if act:
        subprocess.run(cmd, stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL)
    else:
        print(cmd)

# Removing empty dirs
for root, dirs, files in os.walk(OUTPUT_FOLDER):
    if not dirs and not files:
        dirBasename = os.path.basename(root)
        if act and dirBasename not in IGNORED_EMPTY_FOLDER:
            os.rmdir(root)