beets configuration file Beets has an extensive configuration system that lets you customize nearly every aspect of its operation. To configure beets, you create a file called config.yaml. The …
These options control beets’ global operation.
Path to the beets library file. By default, beets will use a file called library.db alongside your configuration file.
The directory to which files will be copied/moved when adding them to the library. Defaults to a folder called Music in your home directory.
A space-separated list of plugin module names to load. See using-plugins.
A space-separated list of extra configuration files to include. Filenames are relative to the directory containing config.yaml.
Directories to search for plugins. Each Python file or directory in a plugin path represents a plugin and should define a subclass of BeetsPlugin. A plugin can then be loaded by adding the filename to the plugins configuration. The plugin path can either be a single string or a list of strings—so, if you have multiple paths, format them as a YAML list like so:
pluginpath:
- /path/one
- /path/two
A list of glob patterns specifying file and directory names to be ignored when importing. By default, this consists of .*, *~, System Volume Information, lost+found (i.e., beets ignores Unix-style hidden files, backup files, and directories that appears at the root of some Linux and Windows filesystems).
Either yes or no; whether to ignore hidden files when importing. On Windows, the “Hidden” property of files is used to detect whether or not a file is hidden. On OS X, the file’s “IsHidden” flag is used to detect whether or not a file is hidden. On both OS X and other platforms (excluding Windows), files (and directories) starting with a dot are detected as hidden files.
A set of regular expression/replacement pairs to be applied to all filenames created by beets. Typically, these replacements are used to avoid confusing problems or errors with the filesystem (for example, leading dots, which hide files on Unix, and trailing whitespace, which is illegal on Windows). To override these substitutions, specify a mapping from regular expression to replacement strings. For example, [xy]: z will make beets replace all instances of the characters x or y with the character z.
If you do change this value, be certain that you include at least enough substitutions to avoid causing errors on your operating system. Here are the default substitutions used by beets, which are sufficient to avoid unexpected behavior on all popular platforms:
replace:
'[\\/]': _
'^\.': _
'[\x00-\x1f]': _
'[<>:"\?\*\|]': _
'\.$': _
'\s+$': ''
'^\s+': ''
'^-': _
These substitutions remove forward and back slashes, leading dots, and control characters—all of which is a good idea on any OS. The fourth line removes the Windows “reserved characters” (useful even on Unix for for compatibility with Windows-influenced network filesystems like Samba). Trailing dots and trailing whitespace, which can cause problems on Windows clients, are also removed.
When replacements other than the defaults are used, it is possible that they will increase the length of the path. In the scenario where this leads to a conflict with the maximum filename length, the default replacements will be used to resolve the conflict and beets will display a warning.
Note that paths might contain special characters such as typographical quotes (“”). With the configuration above, those will not be replaced as they don’t match the typewriter quote ("). To also strip these special characters, you can either add them to the replacement list or use the asciify_paths configuration option below.
A string that replaces the path separator (for example, the forward slash / on Linux and MacOS, and the backward slash \\ on Windows) when generating filenames with beets. This option is related to replace, but is distict from it for technical reasons.
WARNING:
Default: _.
Convert all non-ASCII characters in paths to ASCII equivalents.
For example, if your path template for singletons is singletons/$title and the title of a track is “Café”, then the track will be saved as singletons/Cafe.mp3. The changes take place before applying the replace configuration and are roughly equivalent to wrapping all your path templates in the %asciify{} template function.
This uses the unidecode module which is language agnostic, so some characters may be transliterated from a different language than expected. For example, Japanese kanji will usually use their Chinese readings.
Default: no.
When importing album art, the name of the file (without extension) where the cover art image should be placed. This is a template string, so you can use any of the syntax available to /reference/pathformat. Defaults to cover (i.e., images will be named cover.jpg or cover.png and placed in the album’s directory).
Either yes or no, indicating whether the autotagger should use multiple threads. This makes things substantially faster by overlapping work: for example, it can copy files for one album in parallel with looking up data in MusicBrainz for a different album. You may want to disable this when debugging problems with the autotagger. Defaults to yes.
Format to use when listing individual items with the list-cmd command and other commands that need to print out items. Defaults to $artist - $album - $title. The -f command-line option overrides this setting.
It used to be named list_format_item.
Format to use when listing albums with list-cmd and other commands. Defaults to $albumartist - $album. The -f command-line option overrides this setting.
It used to be named list_format_album.
Default sort order to use when fetching items from the database. Defaults to artist+ album+ disc+ track+. Explicit sort orders override this default.
Default sort order to use when fetching albums from the database. Defaults to albumartist+ album+. Explicit sort orders override this default.
Either yes or no, indicating whether the case should be ignored when sorting lexicographic fields. When set to no, lower-case values will be placed after upper-case values (e.g., Bar Qux foo), while yes would result in the more expected Bar foo Qux. Default: yes.
Either yes or no, indicating whether matched albums should have their year, month, and day fields set to the release date of the original version of an album rather than the selected version of the release. That is, if this option is turned on, then year will always equal original_year and so on. Default: no.
Either yes or no, indicating whether matched tracks and albums should use the artist credit, rather than the artist. That is, if this option is turned on, then artist will contain the artist as credited on the release.
A boolean controlling the track numbering style on multi-disc releases. By default (per_disc_numbering: no), tracks are numbered per-release, so the first track on the second disc has track number N+1 where N is the number of tracks on the first disc. If this per_disc_numbering is enabled, then the first (non-pregap) track on each disc always has track number 1.
If you enable per_disc_numbering, you will likely want to change your Path Format Configuration also to include $disc before $track to make filenames sort correctly in album directories. For example, you might want to use a path format like this:
paths:
default: $albumartist/$album%aunique{}/$disc-$track $title
When this option is off (the default), even “pregap” hidden tracks are numbered from one, not zero, so other track numbers may appear to be bumped up by one. When it is on, the pregap track for each disc can be numbered zero.
These options are used to generate a string that is guaranteed to be unique among all albums in the library who share the same set of keys.
The defaults look like this:
aunique:
keys: albumartist album
disambiguators: albumtype year label catalognum albumdisambig releasegroupdisambig
bracket: '[]'
See aunique for more details.
The text encoding, as known to Python, to use for messages printed to the standard output. It’s also used to read messages from the standard input. By default, this is determined automatically from the locale environment variables.
When beets imports all the files in a directory, it tries to remove the directory if it’s empty. A directory is considered empty if it only contains files whose names match the glob patterns in clutter, which should be a list of strings. The default list consists of “Thumbs.DB” and “.DS_Store”.
The importer only removes recursively searched subdirectories—the top-level directory you specify on the command line is never deleted.
Set the maximum number of characters in a filename, after which names will be truncated. By default, beets tries to ask the filesystem for the correct maximum.
By default, beets writes MP3 tags using the ID3v2.4 standard, the latest version of ID3. Enable this option to instead use the older ID3v2.3 standard, which is preferred by certain older software such as Windows Media Player.
Sets the albumartist for various-artist compilations. Defaults to 'Various Artists' (the MusicBrainz standard). Affects other sources, such as /plugins/discogs, too.
The options that allow for customization of the visual appearance of the console interface.
These options are available in this section:
Either yes or no; whether to use color in console output (currently only in the import command). Turn this off if your terminal doesn’t support ANSI colors.
NOTE:
The colors that are used throughout the user interface. These are only used if the color option is set to yes. For example, you might have a section in your configuration file that looks like this:
ui:
color: yes
colors:
text_success: green
text_warning: yellow
text_error: red
text_highlight: red
text_highlight_minor: lightgray
action_default: turquoise
action: blue
Available colors: black, darkred, darkgreen, brown (darkyellow), darkblue, purple (darkmagenta), teal (darkcyan), lightgray, darkgray, red, green, yellow, blue, fuchsia (magenta), turquoise (cyan), white
The options that control the import-cmd command are indented under the import: key. For example, you might have a section in your configuration file that looks like this:
import:
write: yes
copy: yes
resume: no
These options are available in this section:
Either yes or no, controlling whether metadata (e.g., ID3) tags are written to files when using beet import. Defaults to yes. The -w and -W command-line options override this setting.
Either yes or no, indicating whether to copy files into the library directory when using beet import. Defaults to yes. Can be overridden with the -c and -C command-line options.
The option is ignored if move is enabled (i.e., beets can move or copy files but it doesn’t make sense to do both).
Either yes or no, indicating whether to move files into the library directory when using beet import. Defaults to no.
The effect is similar to the copy option but you end up with only one copy of the imported file. (“Moving” works even across filesystems; if necessary, beets will copy and then delete when a simple rename is impossible.) Moving files can be risky—it’s a good idea to keep a backup in case beets doesn’t do what you expect with your files.
This option overrides copy, so enabling it will always move (and not copy) files. The -c switch to the beet import command, however, still takes precedence.
Either yes or no, indicating whether to use symbolic links instead of moving or copying files. (It conflicts with the move, copy and hardlink options.) Defaults to no.
This option only works on platforms that support symbolic links: i.e., Unixes. It will fail on Windows.
It’s likely that you’ll also want to set write to no if you use this option to preserve the metadata on the linked files.
Either yes or no, indicating whether to use hard links instead of moving, copying, or symlinking files. (It conflicts with the move, copy, and link options.) Defaults to no.
As with symbolic links (see link, above), this will not work on Windows and you will want to set write to no. Otherwise, metadata on the original file will be modified.
Either yes, no, or auto, indicating whether to use copy-on-write file clones (a.k.a. “reflinks”) instead of copying or moving files. The auto option uses reflinks when possible and falls back to plain copying when necessary. Defaults to no.
This kind of clone is only available on certain filesystems: for example, btrfs and APFS. For more details on filesystem support, see the pyreflink documentation. Note that you need to install pyreflink, either through python -m pip install beets[reflink] or python -m pip install reflink.
The option is ignored if move is enabled (i.e., beets can move or copy files but it doesn’t make sense to do both).
Either yes, no, or ask. Controls whether interrupted imports should be resumed. “Yes” means that imports are always resumed when possible; “no” means resuming is disabled entirely; “ask” (the default) means that the user should be prompted when resuming is possible. The -p and -P flags correspond to the “yes” and “no” settings and override this option.
Either yes or no, controlling whether imported directories are recorded and whether these recorded directories are skipped. This corresponds to the -i flag to beet import.
Either yes or no, controlling whether skipped directories are recorded in the incremental list. When set to yes, skipped directories won’t be recorded, and beets will try to import them again later. When set to no, skipped directories will be recorded, and skipped later. Defaults to no.
Either yes or no (default), controlling whether existing metadata is discarded when a match is applied. This corresponds to the --from_scratch flag to beet import.
Either yes or no (default), controlling whether to ask for a manual decision from the user when the importer is unsure how to proceed. This corresponds to the --quiet flag to beet import.
Either skip (default) or asis, specifying what should happen in quiet mode (see the -q flag to import, above) when there is no strong recommendation.
Either ask (default), asis or skip. Specifies what should happen during an interactive import session when there is no recommendation. Useful when you are only interested in processing medium and strong recommendations interactively.
Either yes or no, controlling whether the importer runs in timid mode, in which it asks for confirmation on every autotagging match, even the ones that seem very close. Defaults to no. The -t command-line flag controls the same setting.
Specifies a filename where the importer’s log should be kept. By default, no log is written. This can be overridden with the -l flag to import.
One of apply, skip, asis, or none, indicating which option should be the default when selecting an action for a given match. This is the action that will be taken when you type return without an option letter. The default is apply.
A list of locale names to search for preferred aliases. For example, setting this to en uses the transliterated artist name “Pyotr Ilyich Tchaikovsky” instead of the Cyrillic script for the composer’s name when tagging from MusicBrainz. You can use a space-separated list of language abbreviations, like en jp es, to specify a preference order. Defaults to an empty list, meaning that no language is preferred.
Whether the importer UI should show detailed information about each match it finds. When enabled, this mode prints out the title of every track, regardless of whether it matches the original metadata. (The default behavior only shows changes.) Default: no.
By default, the beets importer groups tracks into albums based on the directories they reside in. This option instead uses files’ metadata to partition albums. Enable this option if you have directories that contain tracks from many albums mixed together.
The --group-albums or -g option to the import-cmd command is equivalent, and the G interactive option invokes the same workflow.
Default: no.
By default, the beets importer always attempts to autotag new music. If most of your collection consists of obscure music, you may be interested in disabling autotagging by setting this option to no. (You can re-enable it with the -a flag to the import-cmd command.)
Default: yes.
Either skip, keep, remove, merge or ask. Controls how duplicates are treated in import task. “skip” means that new item(album or track) will be skipped; “keep” means keep both old and new items; “remove” means remove old item; “merge” means merge into one album; “ask” means the user should be prompted for the action each time. The default is ask.
Ring the terminal bell to get your attention when the importer needs your input.
Default: no.
A dictionary indicating fields to set to values for newly imported music. Here’s an example:
set_fields:
genre: 'To Listen'
collection: 'Unordered'
Other field/value pairs supplied via the --set option on the command-line override any settings here for fields with the same name.
Fields are set on both the album and each individual track of the album. Fields are persisted to the media files of each track.
Default: {} (empty).
You can instruct beets to use your own MusicBrainz database instead of the main server. Use the host, https and ratelimit options under a musicbrainz: header, like so:
musicbrainz:
host: localhost:5000
https: no
ratelimit: 100
The host key, of course, controls the Web server hostname (and port, optionally) that will be contacted by beets (default: musicbrainz.org). The https key makes the client use HTTPS instead of HTTP. This setting applies only to custom servers. The official MusicBrainz server always uses HTTPS. (Default: no.) The server must have search indices enabled (see Building search indexes).
The ratelimit option, an integer, controls the number of Web service requests per second (default: 1). Do not change the rate limit setting if you’re using the main MusicBrainz server—on this public server, you’re limited to one request per second.
The number of matches returned when sending search queries to the MusicBrainz server.
Default: 5.
By default, beets will use only the artist, album, and track count to query MusicBrainz. Additional tags to be queried can be supplied with the extra_tags setting. For example:
musicbrainz:
extra_tags: [year, catalognum, country, media, label]
This setting should improve the autotagger results if the metadata with the given tags match the metadata returned by MusicBrainz.
Note that the only tags supported by this setting are the ones listed in the above example.
Default: []
Use MusicBrainz genre tags to populate (and replace if it’s already set) the genre tag. This will make it a list of all the genres tagged for the release and the release-group on MusicBrainz, separated by “; ” and sorted by the total number of votes. Default: no
You can configure some aspects of the logic beets uses when automatically matching MusicBrainz results under the match: section. To control how tolerant the autotagger is of differences, use the strong_rec_thresh option, which reflects the distance threshold below which beets will make a “strong recommendation” that the metadata be used. Strong recommendations are accepted automatically (except in “timid” mode), so you can use this to make beets ask your opinion more or less often.
The threshold is a distance value between 0.0 and 1.0, so you can think of it as the opposite of a similarity value. For example, if you want to automatically accept any matches above 90% similarity, use:
match:
strong_rec_thresh: 0.10
The default strong recommendation threshold is 0.04.
The medium_rec_thresh and rec_gap_thresh options work similarly. When a match is below the medium recommendation threshold or the distance between it and the next-best match is above the gap threshold, the importer will suggest that match but not automatically confirm it. Otherwise, you’ll see a list of options to choose from.
As mentioned above, autotagger matches have recommendations that control how the UI behaves for a certain quality of match. The recommendation for a certain match is based on the overall distance calculation. But you can also control the recommendation when a specific distance penalty is applied by defining maximum recommendations for each field:
To define maxima, use keys under max_rec: in the match section. The defaults are “medium” for missing and unmatched tracks and “strong” (i.e., no maximum) for everything else:
match:
max_rec:
missing_tracks: medium
unmatched_tracks: medium
If a recommendation is higher than the configured maximum and the indicated penalty is applied, the recommendation is downgraded. The setting for each field can be one of none, low, medium or strong. When the maximum recommendation is strong, no “downgrading” occurs. The available penalty names here are:
In addition to comparing the tagged metadata with the match metadata for similarity, you can also specify an ordered list of preferred countries and media types.
A distance penalty will be applied if the country or media type from the match metadata doesn’t match. The specified values are preferred in descending order (i.e., the first item will be most preferred). Each item may be a regular expression, and will be matched case insensitively. The number of media will be stripped when matching preferred media (e.g. “2x” in “2xCD”).
You can also tell the autotagger to prefer matches that have a release year closest to the original year for an album.
Here’s an example:
match:
preferred:
countries: ['US', 'GB|UK']
media: ['CD', 'Digital Media|File']
original_year: yes
By default, none of these options are enabled.
You can completely avoid matches that have certain penalties applied by adding the penalty name to the ignored setting:
match:
ignored: missing_tracks unmatched_tracks
The available penalties are the same as those for the max_rec setting.
For example, setting ignored: missing_tracks will skip any album matches where your audio files are missing some of the tracks. The importer will not attempt to display these matches. It does not ignore the fact that the album is missing tracks, which would allow these matches to apply more easily. To do that, you’ll want to adjust the penalty for missing tracks.
You can avoid matches that lack certain required information. Add the tags you want to enforce to the required setting:
match:
required: year label catalognum country
No tags are required by default.
A list of media (i.e., formats) in metadata databases to ignore when matching music. You can use this to ignore all media that usually contain video instead of audio, for example:
match:
ignored_media: ['Data CD', 'DVD', 'DVD-Video', 'Blu-ray', 'HD-DVD',
'VCD', 'SVCD', 'UMD', 'VHS']
No formats are ignored by default.
By default, audio files contained in data tracks within a release are included in the album’s tracklist. If you want them to be included, set it no.
Default: yes.
By default, video tracks within a release will be ignored. If you want them to be included (for example if you would like to track the audio-only versions of the video tracks), set it to no.
Default: yes.
You can also configure the directory hierarchy beets uses to store music. These settings appear under the paths: key. Each string is a template string that can refer to metadata fields like $artist or $title. The filename extension is added automatically. At the moment, you can specify three special paths: default for most releases, comp for “various artist” releases with no dominant artist, and singleton for non-album tracks. The defaults look like this:
paths:
default: $albumartist/$album%aunique{}/$track $title
singleton: Non-Album/$artist/$title
comp: Compilations/$album%aunique{}/$track $title
Note the use of $albumartist instead of $artist; this ensures that albums will be well-organized. For more about these format strings, see pathformat. The aunique{} function ensures that identically-named albums are placed in different directories; see aunique for details.
In addition to default, comp, and singleton, you can condition path queries based on beets queries (see /reference/query). This means that a config file like this:
paths:
albumtype:soundtrack: Soundtracks/$album/$track $title
will place soundtrack albums in a separate directory. The queries are tested in the order they appear in the configuration file, meaning that if an item matches multiple queries, beets will use the path format for the first matching query.
Note that the special singleton and comp path format conditions are, in fact, just shorthand for the explicit queries singleton:true and comp:true. In contrast, default is special and has no query equivalent: the default format is only used if no queries match.
The beets configuration file is usually located in a standard location that depends on your OS, but there are a couple of ways you can tell beets where to look.
First, you can set the BEETSDIR environment variable to a directory containing a config.yaml file. This replaces your configuration in the default location. This also affects where auxiliary files, like the library database, are stored by default (that’s where relative paths are resolved to). This environment variable is useful if you need to manage multiple beets libraries with separate configurations.
Alternatively, you can use the --config command-line option to indicate a YAML file containing options that will then be merged with your existing options (from BEETSDIR or the default locations). This is useful if you want to keep your configuration mostly the same but modify a few options as a batch. For example, you might have different strategies for importing files, each with a different set of importer options.
In the absence of a BEETSDIR variable, beets searches a few places for your configuration, depending on the platform:
Beets uses the first directory in your platform’s list that contains config.yaml. If no config file exists, the last path in the list is used.
Here’s an example file:
directory: /var/mp3
import:
copy: yes
write: yes
log: beetslog.txt
art_filename: albumart
plugins: bpd
pluginpath: ~/beets/myplugins
ui:
color: yes
paths:
default: $genre/$albumartist/$album/$track $title
singleton: Singletons/$artist - $title
comp: $genre/$album/$track $title
albumtype:soundtrack: Soundtracks/$album/$track $title
https://beets.readthedocs.org/
Adrian Sampson
2016, Adrian Sampson