Analyse tracks in a pool of worker processes

A first pass over an unanalysed library is hours of librosa, and there was no
reason for a desktop to spend them one core at a time. Analysis now runs in a
process pool sized by `general.library.analysis_workers`, defaulting to one per
core bar one (capped at 8) when the key is absent.

Two consequences worth knowing before touching this: whatever an `Analyzer`
returns has to be picklable, and an analyzer that keeps state on its own
instance - a test double counting calls - only behaves as written with
`analysis_workers=1`.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
This commit is contained in:
2026-09-19 18:04:01 +02:00
parent 75b0eed080
commit fd6283718c
7 changed files with 399 additions and 62 deletions

View File

@@ -32,7 +32,7 @@ from musicmouse.devices.mouse import MusicMouseDevice
from musicmouse.devices.null_transport import NullTransport
from musicmouse.devices.player import Player, VlcPlayer
from musicmouse.devices.serial_link import SerialLink
from musicmouse.library import MusicLibrary
from musicmouse.library import MusicLibrary, default_worker_count
from musicmouse.library.analysis import build_analyzer
from musicmouse.reactions import register_all
from musicmouse.services.base import Service
@@ -283,11 +283,16 @@ def _build_player(bus: EventBus, general: GeneralConfig, *, clock: RealClock) ->
async def build_library(config: Config) -> MusicLibrary:
library_config = config.general.library
workers = library_config.analysis_workers
return await MusicLibrary.build(
library_config.root,
library_config.cache,
frozenset(config.general.audio_extensions),
analyzer=build_analyzer(),
# Unset in the config means "use the machine": a first pass over an unanalyzed
# library is hours of DSP, and there is no reason for a desktop to do it one
# core at a time. `MusicLibrary` itself defaults to 1 - see its docstring.
analysis_workers=default_worker_count() if workers is None else workers,
figure_kinds=config.figure_kinds,
)