mpv-clean-priv-frames

#!/usr/bin/env python3

import json
import subprocess
import typing as t
from pathlib import Path

import click
from more_itertools import unique_everseen, ilen
from my.mpv.history_daemon import all_history, Media, inputs
from mpv_history_daemon.serialize import parse_json_file

disallowed_exact = (
    "log",
    "ytdl_description",
    "md5",
    "comment",
    "info",
    "arcue",
    "itunsmpb",
    "buycdurl",
    "cdtoc",
    "encoding",
    "encoder",
    "major_brand",
    "minor_version",
    "compatible_brands",
    "unsyncedlyrics",
    "itunes_cddb",
    "acoustid fingerprint",
    "notes",
)

disallowed_contains = (
    "location",
    "id3v2_priv",
    "gracenote",
    "organization",
    "streaming/otf/durations",
)

disallowed_startswith = ("lyrics",)


def _priv_frames(metadata: dict[str, t.Any]) -> t.Iterator[str]:
    for k in metadata.keys():
        kl = k.lower()
        for d in disallowed_exact:
            if d == kl:
                yield k
        for d in disallowed_contains:
            if d in kl:
                yield k
        for d in disallowed_startswith:
            if kl.startswith(d):
                yield k


def list_broken(mpv: Media) -> t.Iterator[str]:
    if ilen(_priv_frames(mpv.metadata)) > 0:
        yield mpv.path


def clean_file(jf: Path) -> None:
    """
    Removes any disallowed keys from the metadata object, tracked by
    https://github.com/seanbreckenridge/mpv-history-daemon
    """
    data = parse_json_file(jf)
    blobs = [data]
    if "mapping" in data:
        blobs = data["mapping"].values()
    for blob in blobs:
        for ts in list(blob):
            if "playlist" in blob[ts]:
                print(f"Removing {ts} {blob[ts]}")
                blob.pop(ts)
                continue
            if metadata := blob[ts].get("metadata"):
                if removed_keys := list(_priv_frames(metadata)):
                    for rk in removed_keys:
                        metadata.pop(rk)
                    print(f"Removed {removed_keys}, fixed {metadata}")
    # theres a possibility this is a gz file, so test for that
    # and just write back to the JSON file
    #
    # then, if it was a gz file, use gzip -9 {file} to compress it
    # back to a .json.gz file
    if jf.suffix == ".gz":
        without_gz = jf.with_suffix("")
        click.echo(f"Writing to {without_gz}", err=True)
        without_gz.write_text(json.dumps(data))
        click.echo(f"Compressing {without_gz}", err=True)
        subprocess.run(["gzip", "-f", "-9", str(without_gz)], check=True)
        assert jf.exists()
    else:
        jf.write_text(json.dumps(data))


@click.command()
@click.argument("CMD", type=click.Choice(["list", "clean"], case_sensitive=False))
def main(cmd: str) -> None:
    """
    Removes private id3 frames from mpv data files, and lets me run
    and eyeD3 (ID3) command against any listed files to prevent them
    from appearing in metadata files in the future

    These are typically big blobs of base64 which increase the size
    of the metadata files

    Should list, to check if there are any new tracked private tags:

    mpv-clean-priv-frames list

    Then remove them with a command like:

    \b
    mpv-clean-priv-frames list | sort -u | exists | tr '\\n' '\\0' | xargs -0 -I {} eyeD3 --remove-frame PRIV --user-text-frame 'GracenoteExtData:' --user-text-frame 'GracenoteFileID:' --user-text-frame 'ORGANIZATION:' --user-text-frame "GRACENOTEFILEID:" --user-text-frame "GRACENOTEEXTDATA:" "{}"

    And then run 'mpv-clean-priv-frames clean' to remove the metadata keys from my mpv data files
    """

    if cmd == "clean":
        for p in inputs():
            clean_file(p)
    else:
        for m in all_history():
            for loc in unique_everseen(list_broken(m)):
                print(loc)


if __name__ == "__main__":
    main()