Update all references in core, plugins, and tests to import UserError
from the new location. This centralizes exception handling and improves
code organization.
- Use @cached_property named `mb` (per snejus/semohr) instead of a
helper method, reducing the diff relative to the original code
- Move changelog entry to the Unreleased section (per snejus)
- Refactor tests to use PluginMixin from beets.test.helper (per
semohr) instead of manually managing beets.plugins._instances
The chroma plugin uses Acoustid fingerprinting, which returns MusicBrainz
release and recording IDs. It then unconditionally resolved those IDs
through a privately instantiated MusicBrainzPlugin, so MusicBrainz-sourced
candidates appeared during tagging even when the user had not enabled the
musicbrainz plugin.
Replace the direct MusicBrainzPlugin() instantiation with a lookup through
the metadata-source registry via get_metadata_source("musicbrainz"). When
that returns None, short-circuit both candidates() and item_candidates()
to return empty. This also fixes an incidental issue where the private
instance bypassed any plugin that swaps the musicbrainz plugin at runtime
(e.g. mbpseudo).
Acoustid fingerprinting itself is unaffected — acoustid_id and
acoustid_fingerprint item fields are still populated as before.
Fixes#6212
This command lets user search the database by chromaprint fingerprint
similarity.
Database item fingerprints are computed on the fly if needed.
This is useful for example to check if an unknown / untagged audio file
already exists in the database.
This PR fixes a bug (#3834) where tracks which have already been fingerprinted do not return to be used by `beet submit` (part of the Chroma plugin). This results in submission errors, as the fingerprint is omitted from the resultant payload sent to acoustID.
Since pyacoustid returns the fingerprint as bytes (and thus causes the
database to store a bytes/BLOB object), but the tag value is a string,
the acoustid_fingerprint tag always causes file change when using beet's
"write" command, even if the actual value didn't change.
Issue #2942 describes the problem.
This commit fixes that issue for newly imported/fingerprinted files. However,
you still need to change the type of all acoustid_fingerprint fields
that are already present in the database:
$ sqlite3 beets.db
SQLite version 3.26.0 2018-12-01 12:34:55
Enter ".help" for usage hints.
sqlite> UPDATE items SET acoustid_fingerprint = CAST(acoustid_fingerprint AS TEXT);
Uses
match:preferred:countries/original_year:
config options to determine whether the releases should be sorted by their respective fields.
The behaviour is the same as before when the default config is used.
Delete the remaining usages of BeetsPlugin.listen().
Since BeetsPlugin.listeners are wrapped by a loglevel-setting function,
we cannot easily check their unicity anymore.
BeetsPlugin._raw_listeners set holds the raw listeners.
Legacy plugins that did not handle enough arguments in their listenings
functions may break: dedicated code is now deleted for it would not work
with the decorated listeners.
Tests got fixed. Some modifications were done empirically: if it passes
then it's okay.
Include import of __future__ features division, absolute_imports and
print_function everywhere. Don't add unicode_literals yet for it is
harder to convert.
Goal is smoothing the transition to python 3.
