Add interactive audio extraction, remuxing and encoding tool

This commit is contained in:
ajp_anton committed 2026-10-06 19:27:46 +00:00
1 parent 9c637a67aa
commit 5842d52784
5 files changed
+894

No files matched your search

+401
View File
@@ -0,0 +1,401 @@
"""Keyboard-driven settings for the audio tool."""
from __future__ import annotations
from dataclasses import dataclass, field
from pathlib import Path
from tools.audio_batch import (
COMMON_AUDIO_CODECS,
POOR_M4A_COMPATIBILITY,
AudioInput,
AudioSettings,
encoder_for,
ffmpeg_encoders,
matching_actions,
native_extension,
output_extension,
same_container,
select_fdk_ffmpeg,
universal_containers,
)
from tools.console import dark_field, draw_screen, inverse, light_red, raw_key_input
from tools.filesystem import complete_path
AAC_ESTIMATES = {1: "~64k", 2: "~80k", 3: "~104k", 4: "~128k", 5: "~176k"}
MP3_ESTIMATES = {
0: "~245k", 1: "~225k", 2: "~190k", 3: "~175k", 4: "~165k",
5: "~130k", 6: "~115k", 7: "~100k", 8: "~85k", 9: "~65k",
}
@dataclass
class AudioDialogState:
mode: str = "encode"
codec: str = "opus"
container: str = "Native"
matching_action: str = "re-encode"
aac_encoder: str = "aac"
fdk_mode: str = "VBR"
mp3_mode: str = "VBR"
bitrates: dict[str, str] = field(
default_factory=lambda: {"aac": "192k", "opus": "128k", "mp3": "192k"}
)
aac_quality: str = "5"
mp3_quality: str = "0"
compression: str = "12"
output_dir: str = ""
confirm: str = "Confirm"
@dataclass(frozen=True)
class Row:
key: str
label: str
options: tuple[str, ...] = ()
editable: bool = False
def select_stream_indexes(inputs: list[AudioInput]) -> dict[Path, int]:
multiple = [item for item in inputs if len(item.streams) > 1]
indexes = {item.path: 0 for item in inputs if item.streams}
if not multiple:
return indexes
layouts = {tuple(stream.layout for stream in item.streams) for item in multiple}
if len(layouts) != 1:
print("Warning: multi-stream inputs have different layouts; using their first audio stream.")
return indexes
streams, selected = multiple[0].streams, 0
with raw_key_input() as read_key:
while True:
options = []
for number, stream in enumerate(streams, 1):
details = ", ".join(
value for value in (
stream.codec,
f"{stream.channels} channels" if stream.channels else "",
f"{stream.sample_rate} Hz" if stream.sample_rate else "",
) if value
)
options.append(f"{number}: {details}")
rendered = " ".join(
inverse(option) if index == selected else option
for index, option in enumerate(options)
)
draw_screen([
"Select the audio stream used by multi-stream inputs",
"Left/Right: navigate Enter: confirm",
"",
rendered,
])
key = read_key()
if key == "left":
selected = max(0, selected - 1)
elif key == "right":
selected = min(len(streams) - 1, selected + 1)
elif key == "enter":
indexes.update({item.path: selected for item in multiple})
return indexes
elif key == "esc":
return indexes
def choose_settings(
ffmpeg: str,
inputs: list[AudioInput],
stream_indexes: dict[Path, int],
working_dir: Path,
) -> AudioSettings | None:
encoders = ffmpeg_encoders(ffmpeg)
codecs = [codec for codec in COMMON_AUDIO_CODECS if codec == "aac" or encoder_for(codec, encoders)]
if not codecs:
raise RuntimeError("FFmpeg has none of the supported audio encoders.")
state = AudioDialogState(codec="opus" if "opus" in codecs else codecs[0])
selected = 0
typing_key = typed = error = ""
with raw_key_input() as read_key:
while True:
rows = _rows(state, inputs, stream_indexes, codecs)
_normalize(state, rows)
rows = _rows(state, inputs, stream_indexes, codecs)
_normalize(state, rows)
selected = min(selected, len(rows) - 1)
draw_screen(
_screen(
state, rows, selected, typing_key, typed, error,
working_dir, inputs, stream_indexes,
)
)
key = read_key()
if typing_key:
if key == "esc":
typing_key = typed = error = ""
elif key == "backspace":
typed = typed[:-1]
elif key == "tab" and typing_key == "output_dir":
typed = complete_path(typed, working_dir)
elif key == "enter":
try:
value = _validate_typed(typing_key, typed, working_dir)
except ValueError as exc:
error = str(exc)
else:
_set_value(state, typing_key, value)
typing_key = typed = error = ""
elif key == "space":
typed += " "
elif len(key) == 1 and key >= " ":
typed += key
continue
row = rows[selected]
if key == "up":
selected = max(0, selected - 1)
elif key == "down":
selected = min(len(rows) - 1, selected + 1)
elif key in {"left", "right"} and row.options:
_move_option(state, row, -1 if key == "left" else 1)
elif key == "enter" and row.editable:
typing_key, typed, error = row.key, _raw_value(state, row.key), ""
elif key == "enter" and row.key == "confirm":
if state.confirm == "Exit":
return None
break
elif key == "enter":
selected = min(len(rows) - 1, selected + 1)
elif key == "esc":
return None
elif row.editable and len(key) == 1 and key >= " ":
typing_key, typed, error = row.key, key, ""
output_dir = resolve_output_dir(state.output_dir, working_dir)
encoder = None if state.mode == "remux" else (
state.aac_encoder if state.codec == "aac" else encoder_for(state.codec, encoders)
)
if encoder == "aac" and encoder not in encoders:
raise RuntimeError("The FFmpeg on PATH does not provide its native AAC encoder.")
encoder_ffmpeg = select_fdk_ffmpeg(ffmpeg) if encoder == "libfdk_aac" else None
bitrate, quality, compression = _encoding_value(state)
return AudioSettings(
state.mode,
None if state.mode == "remux" else state.codec,
state.container,
state.matching_action,
encoder,
bitrate,
quality,
compression,
output_dir,
encoder_ffmpeg,
)
def parse_bitrate(value: str) -> str:
text = value.strip().lower()
if text.endswith("k"):
text = text[:-1]
if not text.isdigit() or int(text) <= 0:
raise ValueError("Bitrate must be a positive number, such as 192 or 192k.")
return f"{int(text)}k"
def resolve_output_dir(value: str, working_dir: Path) -> Path | None:
text = value.strip().strip('"')
if not text:
return None
path = Path(text).expanduser()
path = path if path.is_absolute() else working_dir / path
path = path.resolve()
if path.exists() and not path.is_dir():
raise ValueError("The output path exists but is not a directory.")
return path
def _rows(
state: AudioDialogState,
inputs: list[AudioInput],
stream_indexes: dict[Path, int],
codecs: list[str],
) -> list[Row]:
rows = [Row("mode", "Audio handling", ("encode", "remux"))]
if state.mode == "encode":
rows.append(Row("codec", "Output codec", tuple(codecs)))
if state.codec == "aac":
rows.append(Row("aac_encoder", "AAC encoder", ("aac", "libfdk_aac")))
if state.aac_encoder == "libfdk_aac":
rows.append(Row("fdk_mode", "AAC mode", ("VBR", "CBR")))
elif state.codec == "mp3":
rows.append(Row("mp3_mode", "MP3 mode", ("VBR", "CBR")))
rows.append(Row(_value_key(state), _value_label(state), editable=True))
container_codecs = _source_codecs(inputs, stream_indexes) if state.mode == "remux" else {state.codec}
rows.append(Row("container", "Output container", tuple(["Native", *universal_containers(container_codecs)])))
rows.append(Row("matching_action", "Matching inputs", tuple(_matching_options(state, inputs, stream_indexes))))
rows.append(Row("output_dir", "Output directory", editable=True))
rows.append(Row("confirm", "", ("Confirm", "Exit")))
return rows
def _normalize(state: AudioDialogState, rows: list[Row]) -> None:
for row in rows:
if row.options and _raw_value(state, row.key) not in row.options:
_set_value(state, row.key, row.options[0])
def _matching_options(
state: AudioDialogState,
inputs: list[AudioInput],
stream_indexes: dict[Path, int],
) -> list[str]:
if state.mode == "encode":
return matching_actions(
inputs,
stream_indexes,
output_extension(state.container, state.codec),
state.codec,
)
extension = None if state.container == "Native" else f".{state.container.lower()}"
can_skip = any(
same_container(item, extension or native_extension(item.streams[stream_indexes[item.path]].codec))
for item in inputs
)
return ["skip", "remux"] if can_skip else ["remux"]
def _source_codecs(inputs: list[AudioInput], indexes: dict[Path, int]) -> set[str]:
return {item.streams[indexes[item.path]].codec for item in inputs}
def _value_key(state: AudioDialogState) -> str:
if state.codec == "flac":
return "compression"
if state.codec == "aac" and state.aac_encoder == "libfdk_aac" and state.fdk_mode == "VBR":
return "aac_quality"
if state.codec == "mp3" and state.mp3_mode == "VBR":
return "mp3_quality"
return "bitrate"
def _value_label(state: AudioDialogState) -> str:
return {
"compression": "FLAC compression",
"aac_quality": "AAC quality",
"mp3_quality": "MP3 quality",
"bitrate": "Audio bitrate",
}[_value_key(state)]
def _raw_value(state: AudioDialogState, key: str) -> str:
if key == "bitrate":
return state.bitrates[state.codec]
return str(getattr(state, key))
def _set_value(state: AudioDialogState, key: str, value: str) -> None:
if key == "bitrate":
state.bitrates[state.codec] = value
else:
setattr(state, key, value)
def _move_option(state: AudioDialogState, row: Row, movement: int) -> None:
current = row.options.index(_raw_value(state, row.key))
_set_value(state, row.key, row.options[max(0, min(len(row.options) - 1, current + movement))])
def _validate_typed(key: str, value: str, working_dir: Path) -> str:
if key == "bitrate":
return parse_bitrate(value)
if key == "aac_quality":
return _bounded_integer(value, 1, 5, "AAC quality")
if key == "mp3_quality":
return _bounded_integer(value, 0, 9, "MP3 quality")
if key == "compression":
return _bounded_integer(value, 0, 12, "FLAC compression")
if key == "output_dir":
resolve_output_dir(value, working_dir)
return value.strip().strip('"')
return value
def _bounded_integer(value: str, minimum: int, maximum: int, label: str) -> str:
text = value.strip()
if not text.isdigit() or not minimum <= int(text) <= maximum:
raise ValueError(f"{label} must be from {minimum} through {maximum}.")
return str(int(text))
def _encoding_value(state: AudioDialogState) -> tuple[str | None, int | None, int | None]:
if state.mode == "remux":
return None, None, None
key = _value_key(state)
if key == "bitrate":
return state.bitrates[state.codec], None, None
if key == "compression":
return None, None, int(state.compression)
return None, int(state.aac_quality if key == "aac_quality" else state.mp3_quality), None
def _screen(
state: AudioDialogState,
rows: list[Row],
selected: int,
typing_key: str,
typed: str,
error: str,
working_dir: Path,
inputs: list[AudioInput],
stream_indexes: dict[Path, int],
) -> list[str]:
lines = [
"Audio settings",
f"Working directory: {working_dir}",
"",
"Up/Down: navigate Left/Right: choose Enter: edit/confirm Esc: cancel",
"",
]
for index, row in enumerate(rows):
value = typed if typing_key == row.key else _display_value(state, row.key)
if typing_key == row.key:
value = dark_field(value + " ")
elif index == selected:
value = inverse(value)
lines.append(f"{row.label + ':' if row.label else '':<22}{value}")
warnings = _warnings(state, working_dir, inputs, stream_indexes)
if warnings:
lines.extend(("", *(light_red(warning) for warning in warnings)))
if error:
lines.extend(("", light_red(error)))
return lines
def _display_value(state: AudioDialogState, key: str) -> str:
value = _raw_value(state, key)
if key == "output_dir" and not value:
return "source directories"
if key == "aac_quality":
return f"{value} ({AAC_ESTIMATES[int(value)]} stereo)"
if key == "mp3_quality":
return f"{value} ({MP3_ESTIMATES[int(value)]} stereo)"
return value
def _warnings(
state: AudioDialogState,
working_dir: Path,
inputs: list[AudioInput],
stream_indexes: dict[Path, int],
) -> list[str]:
warnings = []
if state.container == "M4A":
codecs = {state.codec} if state.mode == "encode" else _source_codecs(inputs, stream_indexes)
if codecs & POOR_M4A_COMPATIBILITY:
warnings.append("M4A with this codec has limited player compatibility.")
if state.output_dir and not Path(state.output_dir).is_absolute():
warnings.append(f"Relative output directory: {(working_dir / state.output_dir).resolve()}")
return warnings