gronod/headphones
1
0
Fork 0
mirror of https://github.com/rembo10/headphones.git synced 2026-07-20 07:54:01 +01:00
Code Issues Packages Projects Releases Wiki Activity

Migrate as much as possible to pip and requirements.txt

Browse Source
This commit is contained in:
Jamie Magee
2015-11-25 19:52:26 +01:00
parent eaded8dfd9
commit 982594a4a5
1195 changed files with 81 additions and 196887 deletions
Download Patch File Download Diff File Expand all files Collapse all files
lib/beets/LICENSE
-21
View File
@@ -1,21 +0,0 @@
The MIT License
Copyright (c) 2010-2014 Adrian Sampson
Permission is hereby granted, free of charge, to any person obtaining a copy
of this software and associated documentation files (the "Software"), to deal
in the Software without restriction, including without limitation the rights
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
copies of the Software, and to permit persons to whom the Software is
furnished to do so, subject to the following conditions:
The above copyright notice and this permission notice shall be included in
all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
THE SOFTWARE.
lib/beets/README.rst
-94
View File
@@ -1,94 +0,0 @@
.. image:: https://travis-ci.org/sampsyo/beets.svg?branch=master
:target: https://travis-ci.org/sampsyo/beets
.. image:: http://img.shields.io/coveralls/sampsyo/beets.svg
:target: https://coveralls.io/r/sampsyo/beets
.. image:: http://img.shields.io/pypi/v/beets.svg
:target: https://pypi.python.org/pypi/beets
Beets is the media library management system for obsessive-compulsive music
geeks.
The purpose of beets is to get your music collection right once and for all.
It catalogs your collection, automatically improving its metadata as it goes.
It then provides a bouquet of tools for manipulating and accessing your music.
Here's an example of beets' brainy tag corrector doing its thing::
$ beet import ~/music/ladytron
Tagging:
Ladytron - Witching Hour
(Similarity: 98.4%)
* Last One Standing -> The Last One Standing
* Beauty -> Beauty*2
* White Light Generation -> Whitelightgenerator
* All the Way -> All the Way...
Because beets is designed as a library, it can do almost anything you can
imagine for your music collection. Via `plugins`_, beets becomes a panacea:
- Fetch or calculate all the metadata you could possibly need: `album art`_,
`lyrics`_, `genres`_, `tempos`_, `ReplayGain`_ levels, or `acoustic
fingerprints`_.
- Get metadata from `MusicBrainz`_, `Discogs`_, or `Beatport`_. Or guess
metadata using songs' filenames or their acoustic fingerprints.
- `Transcode audio`_ to any format you like.
- Check your library for `duplicate tracks and albums`_ or for `albums that
are missing tracks`_.
- Clean up crufty tags left behind by other, less-awesome tools.
- Embed and extract album art from files' metadata.
- Browse your music library graphically through a Web browser and play it in any
browser that supports `HTML5 Audio`_.
- Analyze music files' metadata from the command line.
- Listen to your library with a music player that speaks the `MPD`_ protocol
and works with a staggering variety of interfaces.
If beets doesn't do what you want yet, `writing your own plugin`_ is
shockingly simple if you know a little Python.
.. _plugins: http://beets.readthedocs.org/page/plugins/
.. _MPD: http://www.musicpd.org/
.. _MusicBrainz music collection: http://musicbrainz.org/doc/Collections/
.. _writing your own plugin:
http://beets.readthedocs.org/page/dev/plugins.html
.. _HTML5 Audio:
http://www.w3.org/TR/html-markup/audio.html
.. _albums that are missing tracks:
http://beets.readthedocs.org/page/plugins/missing.html
.. _duplicate tracks and albums:
http://beets.readthedocs.org/page/plugins/duplicates.html
.. _Transcode audio:
http://beets.readthedocs.org/page/plugins/convert.html
.. _Beatport: http://www.beatport.com/
.. _Discogs: http://www.discogs.com/
.. _acoustic fingerprints:
http://beets.readthedocs.org/page/plugins/chroma.html
.. _ReplayGain: http://beets.readthedocs.org/page/plugins/replaygain.html
.. _tempos: http://beets.readthedocs.org/page/plugins/echonest.html
.. _genres: http://beets.readthedocs.org/page/plugins/lastgenre.html
.. _album art: http://beets.readthedocs.org/page/plugins/fetchart.html
.. _lyrics: http://beets.readthedocs.org/page/plugins/lyrics.html
.. _MusicBrainz: http://musicbrainz.org/
Read More
---------
Learn more about beets at `its Web site`_. Follow `@b33ts`_ on Twitter for
news and updates.
You can install beets by typing ``pip install beets``. Then check out the
`Getting Started`_ guide.
.. _its Web site: http://beets.radbox.org/
.. _Getting Started: http://beets.readthedocs.org/page/guides/main.html
.. _@b33ts: http://twitter.com/b33ts/
Authors
-------
Beets is by `Adrian Sampson`_ with a supporting cast of thousands. For help,
please contact the `mailing list`_.
.. _mailing list: https://groups.google.com/forum/#!forum/beets-users
.. _Adrian Sampson: http://homes.cs.washington.edu/~asampson/
lib/beets/__init__.py
-28
View File
@@ -1,28 +0,0 @@
# This file is part of beets.
# Copyright 2014, Adrian Sampson.
#
# Permission is hereby granted, free of charge, to any person obtaining
# a copy of this software and associated documentation files (the
# "Software"), to deal in the Software without restriction, including
# without limitation the rights to use, copy, modify, merge, publish,
# distribute, sublicense, and/or sell copies of the Software, and to
# permit persons to whom the Software is furnished to do so, subject to
# the following conditions:
#
# The above copyright notice and this permission notice shall be
# included in all copies or substantial portions of the Software.
# This particular version has been slightly modified to work with Headphones
# https://github.com/rembo10/headphones
__version__ = '1.3.10-headphones'
__author__ = 'Adrian Sampson <adrian@radbox.org>'
import os
import beets.library
from beets.util import confit
Library = beets.library.Library
config = confit.LazyConfig(os.path.dirname(__file__), __name__)
lib/beets/autotag/__init__.py
-136
View File
@@ -1,136 +0,0 @@
# This file is part of beets.
# Copyright 2013, Adrian Sampson.
#
# Permission is hereby granted, free of charge, to any person obtaining
# a copy of this software and associated documentation files (the
# "Software"), to deal in the Software without restriction, including
# without limitation the rights to use, copy, modify, merge, publish,
# distribute, sublicense, and/or sell copies of the Software, and to
# permit persons to whom the Software is furnished to do so, subject to
# the following conditions:
#
# The above copyright notice and this permission notice shall be
# included in all copies or substantial portions of the Software.
"""Facilities for automatically determining files' correct metadata.
"""
import logging
from beets import config
# Parts of external interface.
from .hooks import AlbumInfo, TrackInfo, AlbumMatch, TrackMatch # noqa
from .match import tag_item, tag_album # noqa
from .match import Recommendation # noqa
# Global logger.
log = logging.getLogger('beets')
# Additional utilities for the main interface.
def apply_item_metadata(item, track_info):
"""Set an item's metadata from its matched TrackInfo object.
"""
item.artist = track_info.artist
item.artist_sort = track_info.artist_sort
item.artist_credit = track_info.artist_credit
item.title = track_info.title
item.mb_trackid = track_info.track_id
if track_info.artist_id:
item.mb_artistid = track_info.artist_id
# At the moment, the other metadata is left intact (including album
# and track number). Perhaps these should be emptied?
def apply_metadata(album_info, mapping):
"""Set the items' metadata to match an AlbumInfo object using a
mapping from Items to TrackInfo objects.
"""
for item, track_info in mapping.iteritems():
# Album, artist, track count.
if track_info.artist:
item.artist = track_info.artist
else:
item.artist = album_info.artist
item.albumartist = album_info.artist
item.album = album_info.album
# Artist sort and credit names.
item.artist_sort = track_info.artist_sort or album_info.artist_sort
item.artist_credit = (track_info.artist_credit or
album_info.artist_credit)
item.albumartist_sort = album_info.artist_sort
item.albumartist_credit = album_info.artist_credit
# Release date.
for prefix in '', 'original_':
if config['original_date'] and not prefix:
# Ignore specific release date.
continue
for suffix in 'year', 'month', 'day':
key = prefix + suffix
value = getattr(album_info, key) or 0
# If we don't even have a year, apply nothing.
if suffix == 'year' and not value:
break
# Otherwise, set the fetched value (or 0 for the month
# and day if not available).
item[key] = value
# If we're using original release date for both fields,
# also set item.year = info.original_year, etc.
if config['original_date']:
item[suffix] = value
# Title.
item.title = track_info.title
if config['per_disc_numbering']:
item.track = track_info.medium_index or track_info.index
item.tracktotal = track_info.medium_total or len(album_info.tracks)
else:
item.track = track_info.index
item.tracktotal = len(album_info.tracks)
# Disc and disc count.
item.disc = track_info.medium
item.disctotal = album_info.mediums
# MusicBrainz IDs.
item.mb_trackid = track_info.track_id
item.mb_albumid = album_info.album_id
if track_info.artist_id:
item.mb_artistid = track_info.artist_id
else:
item.mb_artistid = album_info.artist_id
item.mb_albumartistid = album_info.artist_id
item.mb_releasegroupid = album_info.releasegroup_id
# Compilation flag.
item.comp = album_info.va
# Miscellaneous metadata.
for field in ('albumtype',
'label',
'asin',
'catalognum',
'script',
'language',
'country',
'albumstatus',
'albumdisambig'):
value = getattr(album_info, field)
if value is not None:
item[field] = value
if track_info.disctitle is not None:
item.disctitle = track_info.disctitle
if track_info.media is not None:
item.media = track_info.media
# Headphones seal of approval
item.comments = 'tagged by headphones/beets'
lib/beets/autotag/hooks.py
-579
View File
@@ -1,579 +0,0 @@
# This file is part of beets.
# Copyright 2013, Adrian Sampson.
#
# Permission is hereby granted, free of charge, to any person obtaining
# a copy of this software and associated documentation files (the
# "Software"), to deal in the Software without restriction, including
# without limitation the rights to use, copy, modify, merge, publish,
# distribute, sublicense, and/or sell copies of the Software, and to
# permit persons to whom the Software is furnished to do so, subject to
# the following conditions:
#
# The above copyright notice and this permission notice shall be
# included in all copies or substantial portions of the Software.
"""Glue between metadata sources and the matching logic."""
import logging
from collections import namedtuple
import re
from beets import plugins
from beets import config
from beets.autotag import mb
from beets.util import levenshtein
from unidecode import unidecode
log = logging.getLogger('beets')
# Classes used to represent candidate options.
class AlbumInfo(object):
"""Describes a canonical release that may be used to match a release
in the library. Consists of these data members:
- ``album``: the release title
- ``album_id``: MusicBrainz ID; UUID fragment only
- ``artist``: name of the release's primary artist
- ``artist_id``
- ``tracks``: list of TrackInfo objects making up the release
- ``asin``: Amazon ASIN
- ``albumtype``: string describing the kind of release
- ``va``: boolean: whether the release has "various artists"
- ``year``: release year
- ``month``: release month
- ``day``: release day
- ``label``: music label responsible for the release
- ``mediums``: the number of discs in this release
- ``artist_sort``: name of the release's artist for sorting
- ``releasegroup_id``: MBID for the album's release group
- ``catalognum``: the label's catalog number for the release
- ``script``: character set used for metadata
- ``language``: human language of the metadata
- ``country``: the release country
- ``albumstatus``: MusicBrainz release status (Official, etc.)
- ``media``: delivery mechanism (Vinyl, etc.)
- ``albumdisambig``: MusicBrainz release disambiguation comment
- ``artist_credit``: Release-specific artist name
- ``data_source``: The original data source (MusicBrainz, Discogs, etc.)
- ``data_url``: The data source release URL.
The fields up through ``tracks`` are required. The others are
optional and may be None.
"""
def __init__(self, album, album_id, artist, artist_id, tracks, asin=None,
albumtype=None, va=False, year=None, month=None, day=None,
label=None, mediums=None, artist_sort=None,
releasegroup_id=None, catalognum=None, script=None,
language=None, country=None, albumstatus=None, media=None,
albumdisambig=None, artist_credit=None, original_year=None,
original_month=None, original_day=None, data_source=None,
data_url=None):
self.album = album
self.album_id = album_id
self.artist = artist
self.artist_id = artist_id
self.tracks = tracks
self.asin = asin
self.albumtype = albumtype
self.va = va
self.year = year
self.month = month
self.day = day
self.label = label
self.mediums = mediums
self.artist_sort = artist_sort
self.releasegroup_id = releasegroup_id
self.catalognum = catalognum
self.script = script
self.language = language
self.country = country
self.albumstatus = albumstatus
self.media = media
self.albumdisambig = albumdisambig
self.artist_credit = artist_credit
self.original_year = original_year
self.original_month = original_month
self.original_day = original_day
self.data_source = data_source
self.data_url = data_url
# Work around a bug in python-musicbrainz-ngs that causes some
# strings to be bytes rather than Unicode.
# https://github.com/alastair/python-musicbrainz-ngs/issues/85
def decode(self, codec='utf8'):
"""Ensure that all string attributes on this object, and the
constituent `TrackInfo` objects, are decoded to Unicode.
"""
for fld in ['album', 'artist', 'albumtype', 'label', 'artist_sort',
'catalognum', 'script', 'language', 'country',
'albumstatus', 'albumdisambig', 'artist_credit', 'media']:
value = getattr(self, fld)
if isinstance(value, str):
setattr(self, fld, value.decode(codec, 'ignore'))
if self.tracks:
for track in self.tracks:
track.decode(codec)
class TrackInfo(object):
"""Describes a canonical track present on a release. Appears as part
of an AlbumInfo's ``tracks`` list. Consists of these data members:
- ``title``: name of the track
- ``track_id``: MusicBrainz ID; UUID fragment only
- ``artist``: individual track artist name
- ``artist_id``
- ``length``: float: duration of the track in seconds
- ``index``: position on the entire release
- ``media``: delivery mechanism (Vinyl, etc.)
- ``medium``: the disc number this track appears on in the album
- ``medium_index``: the track's position on the disc
- ``medium_total``: the number of tracks on the item's disc
- ``artist_sort``: name of the track artist for sorting
- ``disctitle``: name of the individual medium (subtitle)
- ``artist_credit``: Recording-specific artist name
Only ``title`` and ``track_id`` are required. The rest of the fields
may be None. The indices ``index``, ``medium``, and ``medium_index``
are all 1-based.
"""
def __init__(self, title, track_id, artist=None, artist_id=None,
length=None, index=None, medium=None, medium_index=None,
medium_total=None, artist_sort=None, disctitle=None,
artist_credit=None, data_source=None, data_url=None,
media=None):
self.title = title
self.track_id = track_id
self.artist = artist
self.artist_id = artist_id
self.length = length
self.index = index
self.media = media
self.medium = medium
self.medium_index = medium_index
self.medium_total = medium_total
self.artist_sort = artist_sort
self.disctitle = disctitle
self.artist_credit = artist_credit
self.data_source = data_source
self.data_url = data_url
# As above, work around a bug in python-musicbrainz-ngs.
def decode(self, codec='utf8'):
"""Ensure that all string attributes on this object are decoded
to Unicode.
"""
for fld in ['title', 'artist', 'medium', 'artist_sort', 'disctitle',
'artist_credit', 'media']:
value = getattr(self, fld)
if isinstance(value, str):
setattr(self, fld, value.decode(codec, 'ignore'))
# Candidate distance scoring.
# Parameters for string distance function.
# Words that can be moved to the end of a string using a comma.
SD_END_WORDS = ['the', 'a', 'an']
# Reduced weights for certain portions of the string.
SD_PATTERNS = [
(r'^the ', 0.1),
(r'[\[\(]?(ep|single)[\]\)]?', 0.0),
(r'[\[\(]?(featuring|feat|ft)[\. :].+', 0.1),
(r'\(.*?\)', 0.3),
(r'\[.*?\]', 0.3),
(r'(, )?(pt\.|part) .+', 0.2),
]
# Replacements to use before testing distance.
SD_REPLACE = [
(r'&', 'and'),
]
def _string_dist_basic(str1, str2):
"""Basic edit distance between two strings, ignoring
non-alphanumeric characters and case. Comparisons are based on a
transliteration/lowering to ASCII characters. Normalized by string
length.
"""
str1 = unidecode(str1)
str2 = unidecode(str2)
str1 = re.sub(r'[^a-z0-9]', '', str1.lower())
str2 = re.sub(r'[^a-z0-9]', '', str2.lower())
if not str1 and not str2:
return 0.0
return levenshtein(str1, str2) / float(max(len(str1), len(str2)))
def string_dist(str1, str2):
"""Gives an "intuitive" edit distance between two strings. This is
an edit distance, normalized by the string length, with a number of
tweaks that reflect intuition about text.
"""
if str1 is None and str2 is None:
return 0.0
if str1 is None or str2 is None:
return 1.0
str1 = str1.lower()
str2 = str2.lower()
# Don't penalize strings that move certain words to the end. For
# example, "the something" should be considered equal to
# "something, the".
for word in SD_END_WORDS:
if str1.endswith(', %s' % word):
str1 = '%s %s' % (word, str1[:-len(word) - 2])
if str2.endswith(', %s' % word):
str2 = '%s %s' % (word, str2[:-len(word) - 2])
# Perform a couple of basic normalizing substitutions.
for pat, repl in SD_REPLACE:
str1 = re.sub(pat, repl, str1)
str2 = re.sub(pat, repl, str2)
# Change the weight for certain string portions matched by a set
# of regular expressions. We gradually change the strings and build
# up penalties associated with parts of the string that were
# deleted.
base_dist = _string_dist_basic(str1, str2)
penalty = 0.0
for pat, weight in SD_PATTERNS:
# Get strings that drop the pattern.
case_str1 = re.sub(pat, '', str1)
case_str2 = re.sub(pat, '', str2)
if case_str1 != str1 or case_str2 != str2:
# If the pattern was present (i.e., it is deleted in the
# the current case), recalculate the distances for the
# modified strings.
case_dist = _string_dist_basic(case_str1, case_str2)
case_delta = max(0.0, base_dist - case_dist)
if case_delta == 0.0:
continue
# Shift our baseline strings down (to avoid rematching the
# same part of the string) and add a scaled distance
# amount to the penalties.
str1 = case_str1
str2 = case_str2
base_dist = case_dist
penalty += weight * case_delta
return base_dist + penalty
class LazyClassProperty(object):
"""A decorator implementing a read-only property that is *lazy* in
the sense that the getter is only invoked once. Subsequent accesses
through *any* instance use the cached result.
"""
def __init__(self, getter):
self.getter = getter
self.computed = False
def __get__(self, obj, owner):
if not self.computed:
self.value = self.getter(owner)
self.computed = True
return self.value
class Distance(object):
"""Keeps track of multiple distance penalties. Provides a single
weighted distance for all penalties as well as a weighted distance
for each individual penalty.
"""
def __init__(self):
self._penalties = {}
@LazyClassProperty
def _weights(cls):
"""A dictionary from keys to floating-point weights.
"""
weights_view = config['match']['distance_weights']
weights = {}
for key in weights_view.keys():
weights[key] = weights_view[key].as_number()
return weights
# Access the components and their aggregates.
@property
def distance(self):
"""Return a weighted and normalized distance across all
penalties.
"""
dist_max = self.max_distance
if dist_max:
return self.raw_distance / self.max_distance
return 0.0
@property
def max_distance(self):
"""Return the maximum distance penalty (normalization factor).
"""
dist_max = 0.0
for key, penalty in self._penalties.iteritems():
dist_max += len(penalty) * self._weights[key]
return dist_max
@property
def raw_distance(self):
"""Return the raw (denormalized) distance.
"""
dist_raw = 0.0
for key, penalty in self._penalties.iteritems():
dist_raw += sum(penalty) * self._weights[key]
return dist_raw
def items(self):
"""Return a list of (key, dist) pairs, with `dist` being the
weighted distance, sorted from highest to lowest. Does not
include penalties with a zero value.
"""
list_ = []
for key in self._penalties:
dist = self[key]
if dist:
list_.append((key, dist))
# Convert distance into a negative float we can sort items in
# ascending order (for keys, when the penalty is equal) and
# still get the items with the biggest distance first.
return sorted(list_, key=lambda (key, dist): (0 - dist, key))
# Behave like a float.
def __cmp__(self, other):
return cmp(self.distance, other)
def __float__(self):
return self.distance
def __sub__(self, other):
return self.distance - other
def __rsub__(self, other):
return other - self.distance
# Behave like a dict.
def __getitem__(self, key):
"""Returns the weighted distance for a named penalty.
"""
dist = sum(self._penalties[key]) * self._weights[key]
dist_max = self.max_distance
if dist_max:
return dist / dist_max
return 0.0
def __iter__(self):
return iter(self.items())
def __len__(self):
return len(self.items())
def keys(self):
return [key for key, _ in self.items()]
def update(self, dist):
"""Adds all the distance penalties from `dist`.
"""
if not isinstance(dist, Distance):
raise ValueError(
'`dist` must be a Distance object, not {0}'.format(type(dist))
)
for key, penalties in dist._penalties.iteritems():
self._penalties.setdefault(key, []).extend(penalties)
# Adding components.
def _eq(self, value1, value2):
"""Returns True if `value1` is equal to `value2`. `value1` may
be a compiled regular expression, in which case it will be
matched against `value2`.
"""
if isinstance(value1, re._pattern_type):
return bool(value1.match(value2))
return value1 == value2
def add(self, key, dist):
"""Adds a distance penalty. `key` must correspond with a
configured weight setting. `dist` must be a float between 0.0
and 1.0, and will be added to any existing distance penalties
for the same key.
"""
if not 0.0 <= dist <= 1.0:
raise ValueError(
'`dist` must be between 0.0 and 1.0, not {0}'.format(dist)
)
self._penalties.setdefault(key, []).append(dist)
def add_equality(self, key, value, options):
"""Adds a distance penalty of 1.0 if `value` doesn't match any
of the values in `options`. If an option is a compiled regular
expression, it will be considered equal if it matches against
`value`.
"""
if not isinstance(options, (list, tuple)):
options = [options]
for opt in options:
if self._eq(opt, value):
dist = 0.0
break
else:
dist = 1.0
self.add(key, dist)
def add_expr(self, key, expr):
"""Adds a distance penalty of 1.0 if `expr` evaluates to True,
or 0.0.
"""
if expr:
self.add(key, 1.0)
else:
self.add(key, 0.0)
def add_number(self, key, number1, number2):
"""Adds a distance penalty of 1.0 for each number of difference
between `number1` and `number2`, or 0.0 when there is no
difference. Use this when there is no upper limit on the
difference between the two numbers.
"""
diff = abs(number1 - number2)
if diff:
for i in range(diff):
self.add(key, 1.0)
else:
self.add(key, 0.0)
def add_priority(self, key, value, options):
"""Adds a distance penalty that corresponds to the position at
which `value` appears in `options`. A distance penalty of 0.0
for the first option, or 1.0 if there is no matching option. If
an option is a compiled regular expression, it will be
considered equal if it matches against `value`.
"""
if not isinstance(options, (list, tuple)):
options = [options]
unit = 1.0 / (len(options) or 1)
for i, opt in enumerate(options):
if self._eq(opt, value):
dist = i * unit
break
else:
dist = 1.0
self.add(key, dist)
def add_ratio(self, key, number1, number2):
"""Adds a distance penalty for `number1` as a ratio of `number2`.
`number1` is bound at 0 and `number2`.
"""
number = float(max(min(number1, number2), 0))
if number2:
dist = number / number2
else:
dist = 0.0
self.add(key, dist)
def add_string(self, key, str1, str2):
"""Adds a distance penalty based on the edit distance between
`str1` and `str2`.
"""
dist = string_dist(str1, str2)
self.add(key, dist)
# Structures that compose all the information for a candidate match.
AlbumMatch = namedtuple('AlbumMatch', ['distance', 'info', 'mapping',
'extra_items', 'extra_tracks'])
TrackMatch = namedtuple('TrackMatch', ['distance', 'info'])
# Aggregation of sources.
def album_for_mbid(release_id):
"""Get an AlbumInfo object for a MusicBrainz release ID. Return None
if the ID is not found.
"""
try:
return mb.album_for_id(release_id)
except mb.MusicBrainzAPIError as exc:
exc.log(log)
def track_for_mbid(recording_id):
"""Get a TrackInfo object for a MusicBrainz recording ID. Return None
if the ID is not found.
"""
try:
return mb.track_for_id(recording_id)
except mb.MusicBrainzAPIError as exc:
exc.log(log)
def albums_for_id(album_id):
"""Get a list of albums for an ID."""
candidates = [album_for_mbid(album_id)]
candidates.extend(plugins.album_for_id(album_id))
return filter(None, candidates)
def tracks_for_id(track_id):
"""Get a list of tracks for an ID."""
candidates = [track_for_mbid(track_id)]
candidates.extend(plugins.track_for_id(track_id))
return filter(None, candidates)
def album_candidates(items, artist, album, va_likely):
"""Search for album matches. ``items`` is a list of Item objects
that make up the album. ``artist`` and ``album`` are the respective
names (strings), which may be derived from the item list or may be
entered by the user. ``va_likely`` is a boolean indicating whether
the album is likely to be a "various artists" release.
"""
out = []
# Base candidates if we have album and artist to match.
if artist and album:
try:
out.extend(mb.match_album(artist, album, len(items)))
except mb.MusicBrainzAPIError as exc:
exc.log(log)
# Also add VA matches from MusicBrainz where appropriate.
if va_likely and album:
try:
out.extend(mb.match_album(None, album, len(items)))
except mb.MusicBrainzAPIError as exc:
exc.log(log)
# Candidates from plugins.
out.extend(plugins.candidates(items, artist, album, va_likely))
return out
def item_candidates(item, artist, title):
"""Search for item matches. ``item`` is the Item to be matched.
``artist`` and ``title`` are strings and either reflect the item or
are specified by the user.
"""
out = []
# MusicBrainz candidates.
if artist and title:
try:
out.extend(mb.match_track(artist, title))
except mb.MusicBrainzAPIError as exc:
exc.log(log)
# Plugin candidates.
out.extend(plugins.item_candidates(item, artist, title))
return out
lib/beets/autotag/match.py
-492
View File
@@ -1,492 +0,0 @@
# This file is part of beets.
# Copyright 2013, Adrian Sampson.
#
# Permission is hereby granted, free of charge, to any person obtaining
# a copy of this software and associated documentation files (the
# "Software"), to deal in the Software without restriction, including
# without limitation the rights to use, copy, modify, merge, publish,
# distribute, sublicense, and/or sell copies of the Software, and to
# permit persons to whom the Software is furnished to do so, subject to
# the following conditions:
#
# The above copyright notice and this permission notice shall be
# included in all copies or substantial portions of the Software.
"""Matches existing metadata with canonical information to identify
releases and tracks.
"""
from __future__ import division
import datetime
import logging
import re
from munkres import Munkres
from beets import plugins
from beets import config
from beets.util import plurality
from beets.autotag import hooks
from beets.util.enumeration import OrderedEnum
# Artist signals that indicate "various artists". These are used at the
# album level to determine whether a given release is likely a VA
# release and also on the track level to to remove the penalty for
# differing artists.
VA_ARTISTS = (u'', u'various artists', u'various', u'va', u'unknown')
# Global logger.
log = logging.getLogger('beets')
# Recommendation enumeration.
class Recommendation(OrderedEnum):
"""Indicates a qualitative suggestion to the user about what should
be done with a given match.
"""
none = 0
low = 1
medium = 2
strong = 3
# Primary matching functionality.
def current_metadata(items):
"""Extract the likely current metadata for an album given a list of its
items. Return two dictionaries:
- The most common value for each field.
- Whether each field's value was unanimous (values are booleans).
"""
assert items # Must be nonempty.
likelies = {}
consensus = {}
fields = ['artist', 'album', 'albumartist', 'year', 'disctotal',
'mb_albumid', 'label', 'catalognum', 'country', 'media',
'albumdisambig']
for field in fields:
values = [item[field] for item in items if item]
likelies[field], freq = plurality(values)
consensus[field] = (freq == len(values))
# If there's an album artist consensus, use this for the artist.
if consensus['albumartist'] and likelies['albumartist']:
likelies['artist'] = likelies['albumartist']
return likelies, consensus
def assign_items(items, tracks):
"""Given a list of Items and a list of TrackInfo objects, find the
best mapping between them. Returns a mapping from Items to TrackInfo
objects, a set of extra Items, and a set of extra TrackInfo
objects. These "extra" objects occur when there is an unequal number
of objects of the two types.
"""
# Construct the cost matrix.
costs = []
for item in items:
row = []
for i, track in enumerate(tracks):
row.append(track_distance(item, track))
costs.append(row)
# Find a minimum-cost bipartite matching.
matching = Munkres().compute(costs)
# Produce the output matching.
mapping = dict((items[i], tracks[j]) for (i, j) in matching)
extra_items = list(set(items) - set(mapping.keys()))
extra_items.sort(key=lambda i: (i.disc, i.track, i.title))
extra_tracks = list(set(tracks) - set(mapping.values()))
extra_tracks.sort(key=lambda t: (t.index, t.title))
return mapping, extra_items, extra_tracks
def track_index_changed(item, track_info):
"""Returns True if the item and track info index is different. Tolerates
per disc and per release numbering.
"""
return item.track not in (track_info.medium_index, track_info.index)
def track_distance(item, track_info, incl_artist=False):
"""Determines the significance of a track metadata change. Returns a
Distance object. `incl_artist` indicates that a distance component should
be included for the track artist (i.e., for various-artist releases).
"""
dist = hooks.Distance()
# Length.
if track_info.length:
diff = abs(item.length - track_info.length) - \
config['match']['track_length_grace'].as_number()
dist.add_ratio('track_length', diff,
config['match']['track_length_max'].as_number())
# Title.
dist.add_string('track_title', item.title, track_info.title)
# Artist. Only check if there is actually an artist in the track data.
if incl_artist and track_info.artist and \
item.artist.lower() not in VA_ARTISTS:
dist.add_string('track_artist', item.artist, track_info.artist)
# Track index.
if track_info.index and item.track:
dist.add_expr('track_index', track_index_changed(item, track_info))
# Track ID.
if item.mb_trackid:
dist.add_expr('track_id', item.mb_trackid != track_info.track_id)
# Plugins.
dist.update(plugins.track_distance(item, track_info))
return dist
def distance(items, album_info, mapping):
"""Determines how "significant" an album metadata change would be.
Returns a Distance object. `album_info` is an AlbumInfo object
reflecting the album to be compared. `items` is a sequence of all
Item objects that will be matched (order is not important).
`mapping` is a dictionary mapping Items to TrackInfo objects; the
keys are a subset of `items` and the values are a subset of
`album_info.tracks`.
"""
likelies, _ = current_metadata(items)
dist = hooks.Distance()
# Artist, if not various.
if not album_info.va:
dist.add_string('artist', likelies['artist'], album_info.artist)
# Album.
dist.add_string('album', likelies['album'], album_info.album)
# Current or preferred media.
if album_info.media:
# Preferred media options.
patterns = config['match']['preferred']['media'].as_str_seq()
options = [re.compile(r'(\d+x)?(%s)' % pat, re.I) for pat in patterns]
if options:
dist.add_priority('media', album_info.media, options)
# Current media.
elif likelies['media']:
dist.add_equality('media', album_info.media, likelies['media'])
# Mediums.
if likelies['disctotal'] and album_info.mediums:
dist.add_number('mediums', likelies['disctotal'], album_info.mediums)
# Prefer earliest release.
if album_info.year and config['match']['preferred']['original_year']:
# Assume 1889 (earliest first gramophone discs) if we don't know the
# original year.
original = album_info.original_year or 1889
diff = abs(album_info.year - original)
diff_max = abs(datetime.date.today().year - original)
dist.add_ratio('year', diff, diff_max)
# Year.
elif likelies['year'] and album_info.year:
if likelies['year'] in (album_info.year, album_info.original_year):
# No penalty for matching release or original year.
dist.add('year', 0.0)
elif album_info.original_year:
# Prefer matchest closest to the release year.
diff = abs(likelies['year'] - album_info.year)
diff_max = abs(datetime.date.today().year -
album_info.original_year)
dist.add_ratio('year', diff, diff_max)
else:
# Full penalty when there is no original year.
dist.add('year', 1.0)
# Preferred countries.
patterns = config['match']['preferred']['countries'].as_str_seq()
options = [re.compile(pat, re.I) for pat in patterns]
if album_info.country and options:
dist.add_priority('country', album_info.country, options)
# Country.
elif likelies['country'] and album_info.country:
dist.add_string('country', likelies['country'], album_info.country)
# Label.
if likelies['label'] and album_info.label:
dist.add_string('label', likelies['label'], album_info.label)
# Catalog number.
if likelies['catalognum'] and album_info.catalognum:
dist.add_string('catalognum', likelies['catalognum'],
album_info.catalognum)
# Disambiguation.
if likelies['albumdisambig'] and album_info.albumdisambig:
dist.add_string('albumdisambig', likelies['albumdisambig'],
album_info.albumdisambig)
# Album ID.
if likelies['mb_albumid']:
dist.add_equality('album_id', likelies['mb_albumid'],
album_info.album_id)
# Tracks.
dist.tracks = {}
for item, track in mapping.iteritems():
dist.tracks[track] = track_distance(item, track, album_info.va)
dist.add('tracks', dist.tracks[track].distance)
# Missing tracks.
for i in range(len(album_info.tracks) - len(mapping)):
dist.add('missing_tracks', 1.0)
# Unmatched tracks.
for i in range(len(items) - len(mapping)):
dist.add('unmatched_tracks', 1.0)
# Plugins.
dist.update(plugins.album_distance(items, album_info, mapping))
return dist
def match_by_id(items):
"""If the items are tagged with a MusicBrainz album ID, returns an
AlbumInfo object for the corresponding album. Otherwise, returns
None.
"""
# Is there a consensus on the MB album ID?
albumids = [item.mb_albumid for item in items if item.mb_albumid]
if not albumids:
log.debug(u'No album IDs found.')
return None
# If all album IDs are equal, look up the album.
if bool(reduce(lambda x, y: x if x == y else (), albumids)):
albumid = albumids[0]
log.debug(u'Searching for discovered album ID: {0}'.format(albumid))
return hooks.album_for_mbid(albumid)
else:
log.debug(u'No album ID consensus.')
def _recommendation(results):
"""Given a sorted list of AlbumMatch or TrackMatch objects, return a
recommendation based on the results' distances.
If the recommendation is higher than the configured maximum for
an applied penalty, the recommendation will be downgraded to the
configured maximum for that penalty.
"""
if not results:
# No candidates: no recommendation.
return Recommendation.none
# Basic distance thresholding.
min_dist = results[0].distance
if min_dist < config['match']['strong_rec_thresh'].as_number():
# Strong recommendation level.
rec = Recommendation.strong
elif min_dist <= config['match']['medium_rec_thresh'].as_number():
# Medium recommendation level.
rec = Recommendation.medium
elif len(results) == 1:
# Only a single candidate.
rec = Recommendation.low
elif results[1].distance - min_dist >= \
config['match']['rec_gap_thresh'].as_number():
# Gap between first two candidates is large.
rec = Recommendation.low
else:
# No conclusion. Return immediately. Can't be downgraded any further.
return Recommendation.none
# Downgrade to the max rec if it is lower than the current rec for an
# applied penalty.
keys = set(min_dist.keys())
if isinstance(results[0], hooks.AlbumMatch):
for track_dist in min_dist.tracks.values():
keys.update(track_dist.keys())
max_rec_view = config['match']['max_rec']
for key in keys:
if key in max_rec_view.keys():
max_rec = max_rec_view[key].as_choice({
'strong': Recommendation.strong,
'medium': Recommendation.medium,
'low': Recommendation.low,
'none': Recommendation.none,
})
rec = min(rec, max_rec)
return rec
def _add_candidate(items, results, info):
"""Given a candidate AlbumInfo object, attempt to add the candidate
to the output dictionary of AlbumMatch objects. This involves
checking the track count, ordering the items, checking for
duplicates, and calculating the distance.
"""
log.debug(u'Candidate: {0} - {1}'.format(info.artist, info.album))
# Discard albums with zero tracks.
if not info.tracks:
log.debug('No tracks.')
return
# Don't duplicate.
if info.album_id in results:
log.debug(u'Duplicate.')
return
# Discard matches without required tags.
for req_tag in config['match']['required'].as_str_seq():
if getattr(info, req_tag) is None:
log.debug(u'Ignored. Missing required tag: {0}'.format(req_tag))
return
# Find mapping between the items and the track info.
mapping, extra_items, extra_tracks = assign_items(items, info.tracks)
# Get the change distance.
dist = distance(items, info, mapping)
# Skip matches with ignored penalties.
penalties = [key for _, key in dist]
for penalty in config['match']['ignored'].as_str_seq():
if penalty in penalties:
log.debug(u'Ignored. Penalty: {0}'.format(penalty))
return
log.debug(u'Success. Distance: {0}'.format(dist))
results[info.album_id] = hooks.AlbumMatch(dist, info, mapping,
extra_items, extra_tracks)
def tag_album(items, search_artist=None, search_album=None,
search_id=None):
"""Return a tuple of a artist name, an album name, a list of
`AlbumMatch` candidates from the metadata backend, and a
`Recommendation`.
The artist and album are the most common values of these fields
among `items`.
The `AlbumMatch` objects are generated by searching the metadata
backends. By default, the metadata of the items is used for the
search. This can be customized by setting the parameters. The
`mapping` field of the album has the matched `items` as keys.
The recommendation is calculated from the match qualitiy of the
candidates.
"""
# Get current metadata.
likelies, consensus = current_metadata(items)
cur_artist = likelies['artist']
cur_album = likelies['album']
log.debug(u'Tagging {0} - {1}'.format(cur_artist, cur_album))
# The output result (distance, AlbumInfo) tuples (keyed by MB album
# ID).
candidates = {}
# Search by explicit ID.
if search_id is not None:
log.debug(u'Searching for album ID: {0}'.format(search_id))
search_cands = hooks.albums_for_id(search_id)
# Use existing metadata or text search.
else:
# Try search based on current ID.
id_info = match_by_id(items)
if id_info:
_add_candidate(items, candidates, id_info)
rec = _recommendation(candidates.values())
log.debug(u'Album ID match recommendation is {0}'.format(str(rec)))
if candidates and not config['import']['timid']:
# If we have a very good MBID match, return immediately.
# Otherwise, this match will compete against metadata-based
# matches.
if rec == Recommendation.strong:
log.debug(u'ID match.')
return cur_artist, cur_album, candidates.values(), rec
# Search terms.
if not (search_artist and search_album):
# No explicit search terms -- use current metadata.
search_artist, search_album = cur_artist, cur_album
log.debug(u'Search terms: {0} - {1}'.format(search_artist,
search_album))
# Is this album likely to be a "various artist" release?
va_likely = ((not consensus['artist']) or
(search_artist.lower() in VA_ARTISTS) or
any(item.comp for item in items))
log.debug(u'Album might be VA: {0}'.format(str(va_likely)))
# Get the results from the data sources.
search_cands = hooks.album_candidates(items, search_artist,
search_album, va_likely)
log.debug(u'Evaluating {0} candidates.'.format(len(search_cands)))
for info in search_cands:
_add_candidate(items, candidates, info)
# Sort and get the recommendation.
candidates = sorted(candidates.itervalues())
rec = _recommendation(candidates)
return cur_artist, cur_album, candidates, rec
def tag_item(item, search_artist=None, search_title=None,
search_id=None):
"""Attempts to find metadata for a single track. Returns a
`(candidates, recommendation)` pair where `candidates` is a list of
TrackMatch objects. `search_artist` and `search_title` may be used
to override the current metadata for the purposes of the MusicBrainz
title; likewise `search_id`.
"""
# Holds candidates found so far: keys are MBIDs; values are
# (distance, TrackInfo) pairs.
candidates = {}
# First, try matching by MusicBrainz ID.
trackid = search_id or item.mb_trackid
if trackid:
log.debug(u'Searching for track ID: {0}'.format(trackid))
for track_info in hooks.tracks_for_id(trackid):
dist = track_distance(item, track_info, incl_artist=True)
candidates[track_info.track_id] = \
hooks.TrackMatch(dist, track_info)
# If this is a good match, then don't keep searching.
rec = _recommendation(candidates.values())
if rec == Recommendation.strong and not config['import']['timid']:
log.debug(u'Track ID match.')
return candidates.values(), rec
# If we're searching by ID, don't proceed.
if search_id is not None:
if candidates:
return candidates.values(), rec
else:
return [], Recommendation.none
# Search terms.
if not (search_artist and search_title):
search_artist, search_title = item.artist, item.title
log.debug(u'Item search terms: {0} - {1}'.format(search_artist,
search_title))
# Get and evaluate candidate metadata.
for track_info in hooks.item_candidates(item, search_artist, search_title):
dist = track_distance(item, track_info, incl_artist=True)
candidates[track_info.track_id] = hooks.TrackMatch(dist, track_info)
# Sort by distance and return with recommendation.
log.debug(u'Found {0} candidates.'.format(len(candidates)))
candidates = sorted(candidates.itervalues())
rec = _recommendation(candidates)
return candidates, rec
lib/beets/autotag/mb.py
-407
View File
@@ -1,407 +0,0 @@
# This file is part of beets.
# Copyright 2013, Adrian Sampson.
#
# Permission is hereby granted, free of charge, to any person obtaining
# a copy of this software and associated documentation files (the
# "Software"), to deal in the Software without restriction, including
# without limitation the rights to use, copy, modify, merge, publish,
# distribute, sublicense, and/or sell copies of the Software, and to
# permit persons to whom the Software is furnished to do so, subject to
# the following conditions:
#
# The above copyright notice and this permission notice shall be
# included in all copies or substantial portions of the Software.
"""Searches for albums in the MusicBrainz database.
"""
import logging
import musicbrainzngs
import re
import traceback
from urlparse import urljoin
import beets.autotag.hooks
import beets
from beets import util
from beets import config
SEARCH_LIMIT = 5
VARIOUS_ARTISTS_ID = '89ad4ac3-39f7-470e-963a-56509c546377'
BASE_URL = 'http://musicbrainz.org/'
musicbrainzngs.set_useragent('beets', beets.__version__,
'http://beets.radbox.org/')
class MusicBrainzAPIError(util.HumanReadableException):
"""An error while talking to MusicBrainz. The `query` field is the
parameter to the action and may have any type.
"""
def __init__(self, reason, verb, query, tb=None):
self.query = query
super(MusicBrainzAPIError, self).__init__(reason, verb, tb)
def get_message(self):
return u'{0} in {1} with query {2}'.format(
self._reasonstr(), self.verb, repr(self.query)
)
log = logging.getLogger('beets')
RELEASE_INCLUDES = ['artists', 'media', 'recordings', 'release-groups',
'labels', 'artist-credits', 'aliases']
TRACK_INCLUDES = ['artists', 'aliases']
def track_url(trackid):
return urljoin(BASE_URL, 'recording/' + trackid)
def album_url(albumid):
return urljoin(BASE_URL, 'release/' + albumid)
def configure():
"""Set up the python-musicbrainz-ngs module according to settings
from the beets configuration. This should be called at startup.
"""
musicbrainzngs.set_hostname(config['musicbrainz']['host'].get(unicode))
musicbrainzngs.set_rate_limit(
config['musicbrainz']['ratelimit_interval'].as_number(),
config['musicbrainz']['ratelimit'].get(int),
)
def _preferred_alias(aliases):
"""Given an list of alias structures for an artist credit, select
and return the user's preferred alias alias or None if no matching
alias is found.
"""
if not aliases:
return
# Only consider aliases that have locales set.
aliases = [a for a in aliases if 'locale' in a]
# Search configured locales in order.
for locale in config['import']['languages'].as_str_seq():
# Find matching primary aliases for this locale.
matches = [a for a in aliases
if a['locale'] == locale and 'primary' in a]
# Skip to the next locale if we have no matches
if not matches:
continue
return matches[0]
def _flatten_artist_credit(credit):
"""Given a list representing an ``artist-credit`` block, flatten the
data into a triple of joined artist name strings: canonical, sort, and
credit.
"""
artist_parts = []
artist_sort_parts = []
artist_credit_parts = []
for el in credit:
if isinstance(el, basestring):
# Join phrase.
artist_parts.append(el)
artist_credit_parts.append(el)
artist_sort_parts.append(el)
else:
alias = _preferred_alias(el['artist'].get('alias-list', ()))
# An artist.
if alias:
cur_artist_name = alias['alias']
else:
cur_artist_name = el['artist']['name']
artist_parts.append(cur_artist_name)
# Artist sort name.
if alias:
artist_sort_parts.append(alias['sort-name'])
elif 'sort-name' in el['artist']:
artist_sort_parts.append(el['artist']['sort-name'])
else:
artist_sort_parts.append(cur_artist_name)
# Artist credit.
if 'name' in el:
artist_credit_parts.append(el['name'])
else:
artist_credit_parts.append(cur_artist_name)
return (
''.join(artist_parts),
''.join(artist_sort_parts),
''.join(artist_credit_parts),
)
def track_info(recording, index=None, medium=None, medium_index=None,
medium_total=None):
"""Translates a MusicBrainz recording result dictionary into a beets
``TrackInfo`` object. Three parameters are optional and are used
only for tracks that appear on releases (non-singletons): ``index``,
the overall track number; ``medium``, the disc number;
``medium_index``, the track's index on its medium; ``medium_total``,
the number of tracks on the medium. Each number is a 1-based index.
"""
info = beets.autotag.hooks.TrackInfo(
recording['title'],
recording['id'],
index=index,
medium=medium,
medium_index=medium_index,
medium_total=medium_total,
data_url=track_url(recording['id']),
)
if recording.get('artist-credit'):
# Get the artist names.
info.artist, info.artist_sort, info.artist_credit = \
_flatten_artist_credit(recording['artist-credit'])
# Get the ID and sort name of the first artist.
artist = recording['artist-credit'][0]['artist']
info.artist_id = artist['id']
if recording.get('length'):
info.length = int(recording['length']) / (1000.0)
info.decode()
return info
def _set_date_str(info, date_str, original=False):
"""Given a (possibly partial) YYYY-MM-DD string and an AlbumInfo
object, set the object's release date fields appropriately. If
`original`, then set the original_year, etc., fields.
"""
if date_str:
date_parts = date_str.split('-')
for key in ('year', 'month', 'day'):
if date_parts:
date_part = date_parts.pop(0)
try:
date_num = int(date_part)
except ValueError:
continue
if original:
key = 'original_' + key
setattr(info, key, date_num)
def album_info(release):
"""Takes a MusicBrainz release result dictionary and returns a beets
AlbumInfo object containing the interesting data about that release.
"""
# Get artist name using join phrases.
artist_name, artist_sort_name, artist_credit_name = \
_flatten_artist_credit(release['artist-credit'])
# Basic info.
track_infos = []
index = 0
for medium in release['medium-list']:
disctitle = medium.get('title')
format = medium.get('format')
for track in medium['track-list']:
# Basic information from the recording.
index += 1
ti = track_info(
track['recording'],
index,
int(medium['position']),
int(track['position']),
len(medium['track-list']),
)
ti.disctitle = disctitle
ti.media = format
# Prefer track data, where present, over recording data.
if track.get('title'):
ti.title = track['title']
if track.get('artist-credit'):
# Get the artist names.
ti.artist, ti.artist_sort, ti.artist_credit = \
_flatten_artist_credit(track['artist-credit'])
ti.artist_id = track['artist-credit'][0]['artist']['id']
if track.get('length'):
ti.length = int(track['length']) / (1000.0)
track_infos.append(ti)
info = beets.autotag.hooks.AlbumInfo(
release['title'],
release['id'],
artist_name,
release['artist-credit'][0]['artist']['id'],
track_infos,
mediums=len(release['medium-list']),
artist_sort=artist_sort_name,
artist_credit=artist_credit_name,
data_source='MusicBrainz',
data_url=album_url(release['id']),
)
info.va = info.artist_id == VARIOUS_ARTISTS_ID
info.asin = release.get('asin')
info.releasegroup_id = release['release-group']['id']
info.country = release.get('country')
info.albumstatus = release.get('status')
# Build up the disambiguation string from the release group and release.
disambig = []
if release['release-group'].get('disambiguation'):
disambig.append(release['release-group'].get('disambiguation'))
if release.get('disambiguation'):
disambig.append(release.get('disambiguation'))
info.albumdisambig = u', '.join(disambig)
# Release type not always populated.
if 'type' in release['release-group']:
reltype = release['release-group']['type']
if reltype:
info.albumtype = reltype.lower()
# Release dates.
release_date = release.get('date')
release_group_date = release['release-group'].get('first-release-date')
if not release_date:
# Fall back if release-specific date is not available.
release_date = release_group_date
_set_date_str(info, release_date, False)
_set_date_str(info, release_group_date, True)
# Label name.
if release.get('label-info-list'):
label_info = release['label-info-list'][0]
if label_info.get('label'):
label = label_info['label']['name']
if label != '[no label]':
info.label = label
info.catalognum = label_info.get('catalog-number')
# Text representation data.
if release.get('text-representation'):
rep = release['text-representation']
info.script = rep.get('script')
info.language = rep.get('language')
# Media (format).
if release['medium-list']:
first_medium = release['medium-list'][0]
info.media = first_medium.get('format')
info.decode()
return info
def match_album(artist, album, tracks=None, limit=SEARCH_LIMIT):
"""Searches for a single album ("release" in MusicBrainz parlance)
and returns an iterator over AlbumInfo objects. May raise a
MusicBrainzAPIError.
The query consists of an artist name, an album name, and,
optionally, a number of tracks on the album.
"""
# Build search criteria.
criteria = {'release': album.lower().strip()}
if artist is not None:
criteria['artist'] = artist.lower().strip()
else:
# Various Artists search.
criteria['arid'] = VARIOUS_ARTISTS_ID
if tracks is not None:
criteria['tracks'] = str(tracks)
# Abort if we have no search terms.
if not any(criteria.itervalues()):
return
try:
res = musicbrainzngs.search_releases(limit=limit, **criteria)
except musicbrainzngs.MusicBrainzError as exc:
raise MusicBrainzAPIError(exc, 'release search', criteria,
traceback.format_exc())
for release in res['release-list']:
# The search result is missing some data (namely, the tracks),
# so we just use the ID and fetch the rest of the information.
albuminfo = album_for_id(release['id'])
if albuminfo is not None:
yield albuminfo
def match_track(artist, title, limit=SEARCH_LIMIT):
"""Searches for a single track and returns an iterable of TrackInfo
objects. May raise a MusicBrainzAPIError.
"""
criteria = {
'artist': artist.lower().strip(),
'recording': title.lower().strip(),
}
if not any(criteria.itervalues()):
return
try:
res = musicbrainzngs.search_recordings(limit=limit, **criteria)
except musicbrainzngs.MusicBrainzError as exc:
raise MusicBrainzAPIError(exc, 'recording search', criteria,
traceback.format_exc())
for recording in res['recording-list']:
yield track_info(recording)
def _parse_id(s):
"""Search for a MusicBrainz ID in the given string and return it. If
no ID can be found, return None.
"""
# Find the first thing that looks like a UUID/MBID.
match = re.search('[a-f0-9]{8}(-[a-f0-9]{4}){3}-[a-f0-9]{12}', s)
if match:
return match.group()
def album_for_id(releaseid):
"""Fetches an album by its MusicBrainz ID and returns an AlbumInfo
object or None if the album is not found. May raise a
MusicBrainzAPIError.
"""
albumid = _parse_id(releaseid)
if not albumid:
log.debug(u'Invalid MBID ({0}).'.format(releaseid))
return
try:
res = musicbrainzngs.get_release_by_id(albumid,
RELEASE_INCLUDES)
except musicbrainzngs.ResponseError:
log.debug(u'Album ID match failed.')
return None
except musicbrainzngs.MusicBrainzError as exc:
raise MusicBrainzAPIError(exc, 'get release by ID', albumid,
traceback.format_exc())
return album_info(res['release'])
def track_for_id(releaseid):
"""Fetches a track by its MusicBrainz ID. Returns a TrackInfo object
or None if no track is found. May raise a MusicBrainzAPIError.
"""
trackid = _parse_id(releaseid)
if not trackid:
log.debug(u'Invalid MBID ({0}).'.format(releaseid))
return
try:
res = musicbrainzngs.get_recording_by_id(trackid, TRACK_INCLUDES)
except musicbrainzngs.ResponseError:
log.debug(u'Track ID match failed.')
return None
except musicbrainzngs.MusicBrainzError as exc:
raise MusicBrainzAPIError(exc, 'get recording by ID', trackid,
traceback.format_exc())
return track_info(res['recording'])
lib/beets/config_default.yaml
-109
View File
@@ -1,109 +0,0 @@
library: library.db
directory: ~/Music
import:
write: yes
copy: yes
move: no
link: no
delete: no
resume: ask
incremental: no
quiet_fallback: skip
none_rec_action: ask
timid: no
log:
autotag: yes
quiet: no
singletons: no
default_action: apply
languages: []
detail: no
flat: no
group_albums: no
pretend: false
clutter: ["Thumbs.DB", ".DS_Store"]
ignore: [".*", "*~", "System Volume Information"]
replace:
'[\\/]': _
'^\.': _
'[\x00-\x1f]': _
'[<>:"\?\*\|]': _
'\.$': _
'\s+$': ''
'^\s+': ''
path_sep_replace: _
asciify_paths: false
art_filename: cover
max_filename_length: 0
plugins: []
pluginpath: []
threaded: yes
color: yes
timeout: 5.0
per_disc_numbering: no
verbose: no
terminal_encoding: utf8
original_date: no
id3v23: no
ui:
terminal_width: 80
length_diff_thresh: 10.0
list_format_item: $artist - $album - $title
list_format_album: $albumartist - $album
time_format: '%Y-%m-%d %H:%M:%S'
sort_album: albumartist+ album+
sort_item: artist+ album+ disc+ track+
paths:
default: $albumartist/$album%aunique{}/$track $title
singleton: Non-Album/$artist/$title
comp: Compilations/$album%aunique{}/$track $title
statefile: state.pickle
musicbrainz:
host: musicbrainz.org
ratelimit: 1
ratelimit_interval: 1.0
match:
strong_rec_thresh: 0.04
medium_rec_thresh: 0.25
rec_gap_thresh: 0.25
max_rec:
missing_tracks: medium
unmatched_tracks: medium
distance_weights:
source: 2.0
artist: 3.0
album: 3.0
media: 1.0
mediums: 1.0
year: 1.0
country: 0.5
label: 0.5
catalognum: 0.5
albumdisambig: 0.5
album_id: 5.0
tracks: 2.0
missing_tracks: 0.9
unmatched_tracks: 0.6
track_title: 3.0
track_artist: 2.0
track_index: 1.0
track_length: 2.0
track_id: 5.0
preferred:
countries: []
media: []
original_year: no
ignored: []
required: []
track_length_grace: 10
track_length_max: 30
lib/beets/dbcore/__init__.py
-25
View File
@@ -1,25 +0,0 @@
# This file is part of beets.
# Copyright 2014, Adrian Sampson.
#
# Permission is hereby granted, free of charge, to any person obtaining
# a copy of this software and associated documentation files (the
# "Software"), to deal in the Software without restriction, including
# without limitation the rights to use, copy, modify, merge, publish,
# distribute, sublicense, and/or sell copies of the Software, and to
# permit persons to whom the Software is furnished to do so, subject to
# the following conditions:
#
# The above copyright notice and this permission notice shall be
# included in all copies or substantial portions of the Software.
"""DBCore is an abstract database package that forms the basis for beets'
Library.
"""
from .db import Model, Database
from .query import Query, FieldQuery, MatchQuery, AndQuery, OrQuery
from .types import Type
from .queryparse import query_from_strings
from .queryparse import sort_from_strings
from .queryparse import parse_sorted_query
# flake8: noqa
lib/beets/dbcore/db.py
-820
View File
@@ -1,820 +0,0 @@
# This file is part of beets.
# Copyright 2014, Adrian Sampson.
#
# Permission is hereby granted, free of charge, to any person obtaining
# a copy of this software and associated documentation files (the
# "Software"), to deal in the Software without restriction, including
# without limitation the rights to use, copy, modify, merge, publish,
# distribute, sublicense, and/or sell copies of the Software, and to
# permit persons to whom the Software is furnished to do so, subject to
# the following conditions:
#
# The above copyright notice and this permission notice shall be
# included in all copies or substantial portions of the Software.
"""The central Model and Database constructs for DBCore.
"""
import time
import os
from collections import defaultdict
import threading
import sqlite3
import contextlib
import collections
import beets
from beets.util.functemplate import Template
from beets.dbcore import types
from .query import MatchQuery, NullSort, TrueQuery
class FormattedMapping(collections.Mapping):
"""A `dict`-like formatted view of a model.
The accessor `mapping[key]` returns the formated version of
`model[key]` as a unicode string.
If `for_path` is true, all path separators in the formatted values
are replaced.
"""
def __init__(self, model, for_path=False):
self.for_path = for_path
self.model = model
self.model_keys = model.keys(True)
def __getitem__(self, key):
if key in self.model_keys:
return self._get_formatted(self.model, key)
else:
raise KeyError(key)
def __iter__(self):
return iter(self.model_keys)
def __len__(self):
return len(self.model_keys)
def get(self, key, default=None):
if default is None:
default = self.model._type(key).format(None)
return super(FormattedMapping, self).get(key, default)
def _get_formatted(self, model, key):
value = model._type(key).format(model.get(key))
if isinstance(value, bytes):
value = value.decode('utf8', 'ignore')
if self.for_path:
sep_repl = beets.config['path_sep_replace'].get(unicode)
for sep in (os.path.sep, os.path.altsep):
if sep:
value = value.replace(sep, sep_repl)
return value
# Abstract base for model classes.
class Model(object):
"""An abstract object representing an object in the database. Model
objects act like dictionaries (i.e., the allow subscript access like
``obj['field']``). The same field set is available via attribute
access as a shortcut (i.e., ``obj.field``). Three kinds of attributes are
available:
* **Fixed attributes** come from a predetermined list of field
names. These fields correspond to SQLite table columns and are
thus fast to read, write, and query.
* **Flexible attributes** are free-form and do not need to be listed
ahead of time.
* **Computed attributes** are read-only fields computed by a getter
function provided by a plugin.
Access to all three field types is uniform: ``obj.field`` works the
same regardless of whether ``field`` is fixed, flexible, or
computed.
Model objects can optionally be associated with a `Library` object,
in which case they can be loaded and stored from the database. Dirty
flags are used to track which fields need to be stored.
"""
# Abstract components (to be provided by subclasses).
_table = None
"""The main SQLite table name.
"""
_flex_table = None
"""The flex field SQLite table name.
"""
_fields = {}
"""A mapping indicating available "fixed" fields on this type. The
keys are field names and the values are `Type` objects.
"""
_search_fields = ()
"""The fields that should be queried by default by unqualified query
terms.
"""
_types = {}
"""Optional Types for non-fixed (i.e., flexible and computed) fields.
"""
_sorts = {}
"""Optional named sort criteria. The keys are strings and the values
are subclasses of `Sort`.
"""
_always_dirty = False
"""By default, fields only become "dirty" when their value actually
changes. Enabling this flag marks fields as dirty even when the new
value is the same as the old value (e.g., `o.f = o.f`).
"""
@classmethod
def _getters(cls):
"""Return a mapping from field names to getter functions.
"""
# We could cache this if it becomes a performance problem to
# gather the getter mapping every time.
raise NotImplementedError()
def _template_funcs(self):
"""Return a mapping from function names to text-transformer
functions.
"""
# As above: we could consider caching this result.
raise NotImplementedError()
# Basic operation.
def __init__(self, db=None, **values):
"""Create a new object with an optional Database association and
initial field values.
"""
self._db = db
self._dirty = set()
self._values_fixed = {}
self._values_flex = {}
# Initial contents.
self.update(values)
self.clear_dirty()
@classmethod
def _awaken(cls, db=None, fixed_values={}, flex_values={}):
"""Create an object with values drawn from the database.
This is a performance optimization: the checks involved with
ordinary construction are bypassed.
"""
obj = cls(db)
for key, value in fixed_values.iteritems():
obj._values_fixed[key] = cls._type(key).from_sql(value)
for key, value in flex_values.iteritems():
obj._values_flex[key] = cls._type(key).from_sql(value)
return obj
def __repr__(self):
return '{0}({1})'.format(
type(self).__name__,
', '.join('{0}={1!r}'.format(k, v) for k, v in dict(self).items()),
)
def clear_dirty(self):
"""Mark all fields as *clean* (i.e., not needing to be stored to
the database).
"""
self._dirty = set()
def _check_db(self, need_id=True):
"""Ensure that this object is associated with a database row: it
has a reference to a database (`_db`) and an id. A ValueError
exception is raised otherwise.
"""
if not self._db:
raise ValueError('{0} has no database'.format(type(self).__name__))
if need_id and not self.id:
raise ValueError('{0} has no id'.format(type(self).__name__))
# Essential field accessors.
@classmethod
def _type(self, key):
"""Get the type of a field, a `Type` instance.
If the field has no explicit type, it is given the base `Type`,
which does no conversion.
"""
return self._fields.get(key) or self._types.get(key) or types.DEFAULT
def __getitem__(self, key):
"""Get the value for a field. Raise a KeyError if the field is
not available.
"""
getters = self._getters()
if key in getters: # Computed.
return getters[key](self)
elif key in self._fields: # Fixed.
return self._values_fixed.get(key)
elif key in self._values_flex: # Flexible.
return self._values_flex[key]
else:
raise KeyError(key)
def __setitem__(self, key, value):
"""Assign the value for a field.
"""
# Choose where to place the value.
if key in self._fields:
source = self._values_fixed
else:
source = self._values_flex
# If the field has a type, filter the value.
value = self._type(key).normalize(value)
# Assign value and possibly mark as dirty.
old_value = source.get(key)
source[key] = value
if self._always_dirty or old_value != value:
self._dirty.add(key)
def __delitem__(self, key):
"""Remove a flexible attribute from the model.
"""
if key in self._values_flex: # Flexible.
del self._values_flex[key]
self._dirty.add(key) # Mark for dropping on store.
elif key in self._getters(): # Computed.
raise KeyError('computed field {0} cannot be deleted'.format(key))
elif key in self._fields: # Fixed.
raise KeyError('fixed field {0} cannot be deleted'.format(key))
else:
raise KeyError('no such field {0}'.format(key))
def keys(self, computed=False):
"""Get a list of available field names for this object. The
`computed` parameter controls whether computed (plugin-provided)
fields are included in the key list.
"""
base_keys = list(self._fields) + self._values_flex.keys()
if computed:
return base_keys + self._getters().keys()
else:
return base_keys
# Act like a dictionary.
def update(self, values):
"""Assign all values in the given dict.
"""
for key, value in values.items():
self[key] = value
def items(self):
"""Iterate over (key, value) pairs that this object contains.
Computed fields are not included.
"""
for key in self:
yield key, self[key]
def get(self, key, default=None):
"""Get the value for a given key or `default` if it does not
exist.
"""
if key in self:
return self[key]
else:
return default
def __contains__(self, key):
"""Determine whether `key` is an attribute on this object.
"""
return key in self.keys(True)
def __iter__(self):
"""Iterate over the available field names (excluding computed
fields).
"""
return iter(self.keys())
# Convenient attribute access.
def __getattr__(self, key):
if key.startswith('_'):
raise AttributeError('model has no attribute {0!r}'.format(key))
else:
try:
return self[key]
except KeyError:
raise AttributeError('no such field {0!r}'.format(key))
def __setattr__(self, key, value):
if key.startswith('_'):
super(Model, self).__setattr__(key, value)
else:
self[key] = value
def __delattr__(self, key):
if key.startswith('_'):
super(Model, self).__delattr__(key)
else:
del self[key]
# Database interaction (CRUD methods).
def store(self):
"""Save the object's metadata into the library database.
"""
self._check_db()
# Build assignments for query.
assignments = []
subvars = []
for key in self._fields:
if key != 'id' and key in self._dirty:
self._dirty.remove(key)
assignments.append(key + '=?')
value = self._type(key).to_sql(self[key])
subvars.append(value)
assignments = ','.join(assignments)
with self._db.transaction() as tx:
# Main table update.
if assignments:
query = 'UPDATE {0} SET {1} WHERE id=?'.format(
self._table, assignments
)
subvars.append(self.id)
tx.mutate(query, subvars)
# Modified/added flexible attributes.
for key, value in self._values_flex.items():
if key in self._dirty:
self._dirty.remove(key)
tx.mutate(
'INSERT INTO {0} '
'(entity_id, key, value) '
'VALUES (?, ?, ?);'.format(self._flex_table),
(self.id, key, value),
)
# Deleted flexible attributes.
for key in self._dirty:
tx.mutate(
'DELETE FROM {0} '
'WHERE entity_id=? AND key=?'.format(self._flex_table),
(self.id, key)
)
self.clear_dirty()
def load(self):
"""Refresh the object's metadata from the library database.
"""
self._check_db()
stored_obj = self._db._get(type(self), self.id)
assert stored_obj is not None, "object {0} not in DB".format(self.id)
self._values_fixed = {}
self._values_flex = {}
self.update(dict(stored_obj))
self.clear_dirty()
def remove(self):
"""Remove the object's associated rows from the database.
"""
self._check_db()
with self._db.transaction() as tx:
tx.mutate(
'DELETE FROM {0} WHERE id=?'.format(self._table),
(self.id,)
)
tx.mutate(
'DELETE FROM {0} WHERE entity_id=?'.format(self._flex_table),
(self.id,)
)
def add(self, db=None):
"""Add the object to the library database. This object must be
associated with a database; you can provide one via the `db`
parameter or use the currently associated database.
The object's `id` and `added` fields are set along with any
current field values.
"""
if db:
self._db = db
self._check_db(False)
with self._db.transaction() as tx:
new_id = tx.mutate(
'INSERT INTO {0} DEFAULT VALUES'.format(self._table)
)
self.id = new_id
self.added = time.time()
# Mark every non-null field as dirty and store.
for key in self:
if self[key] is not None:
self._dirty.add(key)
self.store()
# Formatting and templating.
_formatter = FormattedMapping
def formatted(self, for_path=False):
"""Get a mapping containing all values on this object formatted
as human-readable unicode strings.
"""
return self._formatter(self, for_path)
def evaluate_template(self, template, for_path=False):
"""Evaluate a template (a string or a `Template` object) using
the object's fields. If `for_path` is true, then no new path
separators will be added to the template.
"""
# Perform substitution.
if isinstance(template, basestring):
template = Template(template)
return template.substitute(self.formatted(for_path),
self._template_funcs())
# Parsing.
@classmethod
def _parse(cls, key, string):
"""Parse a string as a value for the given key.
"""
if not isinstance(string, basestring):
raise TypeError("_parse() argument must be a string")
return cls._type(key).parse(string)
# Database controller and supporting interfaces.
class Results(object):
"""An item query result set. Iterating over the collection lazily
constructs LibModel objects that reflect database rows.
"""
def __init__(self, model_class, rows, db, query=None, sort=None):
"""Create a result set that will construct objects of type
`model_class`.
`model_class` is a subclass of `LibModel` that will be
constructed. `rows` is a query result: a list of mappings. The
new objects will be associated with the database `db`.
If `query` is provided, it is used as a predicate to filter the
results for a "slow query" that cannot be evaluated by the
database directly. If `sort` is provided, it is used to sort the
full list of results before returning. This means it is a "slow
sort" and all objects must be built before returning the first
one.
"""
self.model_class = model_class
self.rows = rows
self.db = db
self.query = query
self.sort = sort
# We keep a queue of rows we haven't yet consumed for
# materialization. We preserve the original total number of
# rows.
self._rows = rows
self._row_count = len(rows)
# The materialized objects corresponding to rows that have been
# consumed.
self._objects = []
def _get_objects(self):
"""Construct and generate Model objects for they query. The
objects are returned in the order emitted from the database; no
slow sort is applied.
For performance, this generator caches materialized objects to
avoid constructing them more than once. This way, iterating over
a `Results` object a second time should be much faster than the
first.
"""
index = 0 # Position in the materialized objects.
while index < len(self._objects) or self._rows:
# Are there previously-materialized objects to produce?
if index < len(self._objects):
yield self._objects[index]
index += 1
# Otherwise, we consume another row, materialize its object
# and produce it.
else:
while self._rows:
row = self._rows.pop(0)
obj = self._make_model(row)
# If there is a slow-query predicate, ensurer that the
# object passes it.
if not self.query or self.query.match(obj):
self._objects.append(obj)
index += 1
yield obj
break
def __iter__(self):
"""Construct and generate Model objects for all matching
objects, in sorted order.
"""
if self.sort:
# Slow sort. Must build the full list first.
objects = self.sort.sort(list(self._get_objects()))
return iter(objects)
else:
# Objects are pre-sorted (i.e., by the database).
return self._get_objects()
def _make_model(self, row):
# Get the flexible attributes for the object.
with self.db.transaction() as tx:
flex_rows = tx.query(
'SELECT * FROM {0} WHERE entity_id=?'.format(
self.model_class._flex_table
),
(row['id'],)
)
cols = dict(row)
values = dict((k, v) for (k, v) in cols.items()
if not k[:4] == 'flex')
flex_values = dict((row['key'], row['value']) for row in flex_rows)
# Construct the Python object
obj = self.model_class._awaken(self.db, values, flex_values)
return obj
def __len__(self):
"""Get the number of matching objects.
"""
if not self._rows:
# Fully materialized. Just count the objects.
return len(self._objects)
elif self.query:
# A slow query. Fall back to testing every object.
count = 0
for obj in self:
count += 1
return count
else:
# A fast query. Just count the rows.
return self._row_count
def __nonzero__(self):
"""Does this result contain any objects?
"""
return bool(len(self))
def __getitem__(self, n):
"""Get the nth item in this result set. This is inefficient: all
items up to n are materialized and thrown away.
"""
if not self._rows and not self.sort:
# Fully materialized and already in order. Just look up the
# object.
return self._objects[n]
it = iter(self)
try:
for i in range(n):
it.next()
return it.next()
except StopIteration:
raise IndexError('result index {0} out of range'.format(n))
def get(self):
"""Return the first matching object, or None if no objects
match.
"""
it = iter(self)
try:
return it.next()
except StopIteration:
return None
class Transaction(object):
"""A context manager for safe, concurrent access to the database.
All SQL commands should be executed through a transaction.
"""
def __init__(self, db):
self.db = db
def __enter__(self):
"""Begin a transaction. This transaction may be created while
another is active in a different thread.
"""
with self.db._tx_stack() as stack:
first = not stack
stack.append(self)
if first:
# Beginning a "root" transaction, which corresponds to an
# SQLite transaction.
self.db._db_lock.acquire()
return self
def __exit__(self, exc_type, exc_value, traceback):
"""Complete a transaction. This must be the most recently
entered but not yet exited transaction. If it is the last active
transaction, the database updates are committed.
"""
with self.db._tx_stack() as stack:
assert stack.pop() is self
empty = not stack
if empty:
# Ending a "root" transaction. End the SQLite transaction.
self.db._connection().commit()
self.db._db_lock.release()
def query(self, statement, subvals=()):
"""Execute an SQL statement with substitution values and return
a list of rows from the database.
"""
cursor = self.db._connection().execute(statement, subvals)
return cursor.fetchall()
def mutate(self, statement, subvals=()):
"""Execute an SQL statement with substitution values and return
the row ID of the last affected row.
"""
cursor = self.db._connection().execute(statement, subvals)
return cursor.lastrowid
def script(self, statements):
"""Execute a string containing multiple SQL statements."""
self.db._connection().executescript(statements)
class Database(object):
"""A container for Model objects that wraps an SQLite database as
the backend.
"""
_models = ()
"""The Model subclasses representing tables in this database.
"""
def __init__(self, path):
self.path = path
self._connections = {}
self._tx_stacks = defaultdict(list)
# A lock to protect the _connections and _tx_stacks maps, which
# both map thread IDs to private resources.
self._shared_map_lock = threading.Lock()
# A lock to protect access to the database itself. SQLite does
# allow multiple threads to access the database at the same
# time, but many users were experiencing crashes related to this
# capability: where SQLite was compiled without HAVE_USLEEP, its
# backoff algorithm in the case of contention was causing
# whole-second sleeps (!) that would trigger its internal
# timeout. Using this lock ensures only one SQLite transaction
# is active at a time.
self._db_lock = threading.Lock()
# Set up database schema.
for model_cls in self._models:
self._make_table(model_cls._table, model_cls._fields)
self._make_attribute_table(model_cls._flex_table)
# Primitive access control: connections and transactions.
def _connection(self):
"""Get a SQLite connection object to the underlying database.
One connection object is created per thread.
"""
thread_id = threading.current_thread().ident
with self._shared_map_lock:
if thread_id in self._connections:
return self._connections[thread_id]
else:
# Make a new connection.
conn = sqlite3.connect(
self.path,
timeout=beets.config['timeout'].as_number(),
)
# Access SELECT results like dictionaries.
conn.row_factory = sqlite3.Row
self._connections[thread_id] = conn
return conn
@contextlib.contextmanager
def _tx_stack(self):
"""A context manager providing access to the current thread's
transaction stack. The context manager synchronizes access to
the stack map. Transactions should never migrate across threads.
"""
thread_id = threading.current_thread().ident
with self._shared_map_lock:
yield self._tx_stacks[thread_id]
def transaction(self):
"""Get a :class:`Transaction` object for interacting directly
with the underlying SQLite database.
"""
return Transaction(self)
# Schema setup and migration.
def _make_table(self, table, fields):
"""Set up the schema of the database. `fields` is a mapping
from field names to `Type`s. Columns are added if necessary.
"""
# Get current schema.
with self.transaction() as tx:
rows = tx.query('PRAGMA table_info(%s)' % table)
current_fields = set([row[1] for row in rows])
field_names = set(fields.keys())
if current_fields.issuperset(field_names):
# Table exists and has all the required columns.
return
if not current_fields:
# No table exists.
columns = []
for name, typ in fields.items():
columns.append('{0} {1}'.format(name, typ.sql))
setup_sql = 'CREATE TABLE {0} ({1});\n'.format(table,
', '.join(columns))
else:
# Table exists does not match the field set.
setup_sql = ''
for name, typ in fields.items():
if name in current_fields:
continue
setup_sql += 'ALTER TABLE {0} ADD COLUMN {1} {2};\n'.format(
table, name, typ.sql
)
with self.transaction() as tx:
tx.script(setup_sql)
def _make_attribute_table(self, flex_table):
"""Create a table and associated index for flexible attributes
for the given entity (if they don't exist).
"""
with self.transaction() as tx:
tx.script("""
CREATE TABLE IF NOT EXISTS {0} (
id INTEGER PRIMARY KEY,
entity_id INTEGER,
key TEXT,
value TEXT,
UNIQUE(entity_id, key) ON CONFLICT REPLACE);
CREATE INDEX IF NOT EXISTS {0}_by_entity
ON {0} (entity_id);
""".format(flex_table))
# Querying.
def _fetch(self, model_cls, query=None, sort=None):
"""Fetch the objects of type `model_cls` matching the given
query. The query may be given as a string, string sequence, a
Query object, or None (to fetch everything). `sort` is an
`Sort` object.
"""
query = query or TrueQuery() # A null query.
sort = sort or NullSort() # Unsorted.
where, subvals = query.clause()
order_by = sort.order_clause()
sql = ("SELECT * FROM {0} WHERE {1} {2}").format(
model_cls._table,
where or '1',
"ORDER BY {0}".format(order_by) if order_by else '',
)
with self.transaction() as tx:
rows = tx.query(sql, subvals)
return Results(
model_cls, rows, self,
None if where else query, # Slow query component.
sort if sort.is_slow() else None, # Slow sort component.
)
def _get(self, model_cls, id):
"""Get a Model object by its id or None if the id does not
exist.
"""
return self._fetch(model_cls, MatchQuery('id', id)).get()
lib/beets/dbcore/query.py
-654
View File
@@ -1,654 +0,0 @@
# This file is part of beets.
# Copyright 2014, Adrian Sampson.
#
# Permission is hereby granted, free of charge, to any person obtaining
# a copy of this software and associated documentation files (the
# "Software"), to deal in the Software without restriction, including
# without limitation the rights to use, copy, modify, merge, publish,
# distribute, sublicense, and/or sell copies of the Software, and to
# permit persons to whom the Software is furnished to do so, subject to
# the following conditions:
#
# The above copyright notice and this permission notice shall be
# included in all copies or substantial portions of the Software.
"""The Query type hierarchy for DBCore.
"""
import re
from operator import attrgetter
from beets import util
from datetime import datetime, timedelta
class Query(object):
"""An abstract class representing a query into the item database.
"""
def clause(self):
"""Generate an SQLite expression implementing the query.
Return a clause string, a sequence of substitution values for
the clause, and a Query object representing the "remainder"
Returns (clause, subvals) where clause is a valid sqlite
WHERE clause implementing the query and subvals is a list of
items to be substituted for ?s in the clause.
"""
return None, ()
def match(self, item):
"""Check whether this query matches a given Item. Can be used to
perform queries on arbitrary sets of Items.
"""
raise NotImplementedError
class FieldQuery(Query):
"""An abstract query that searches in a specific field for a
pattern. Subclasses must provide a `value_match` class method, which
determines whether a certain pattern string matches a certain value
string. Subclasses may also provide `col_clause` to implement the
same matching functionality in SQLite.
"""
def __init__(self, field, pattern, fast=True):
self.field = field
self.pattern = pattern
self.fast = fast
def col_clause(self):
return None, ()
def clause(self):
if self.fast:
return self.col_clause()
else:
# Matching a flexattr. This is a slow query.
return None, ()
@classmethod
def value_match(cls, pattern, value):
"""Determine whether the value matches the pattern. Both
arguments are strings.
"""
raise NotImplementedError()
def match(self, item):
return self.value_match(self.pattern, item.get(self.field))
class MatchQuery(FieldQuery):
"""A query that looks for exact matches in an item field."""
def col_clause(self):
return self.field + " = ?", [self.pattern]
@classmethod
def value_match(cls, pattern, value):
return pattern == value
class NoneQuery(FieldQuery):
def __init__(self, field, fast=True):
self.field = field
self.fast = fast
def col_clause(self):
return self.field + " IS NULL", ()
@classmethod
def match(self, item):
try:
return item[self.field] is None
except KeyError:
return True
class StringFieldQuery(FieldQuery):
"""A FieldQuery that converts values to strings before matching
them.
"""
@classmethod
def value_match(cls, pattern, value):
"""Determine whether the value matches the pattern. The value
may have any type.
"""
return cls.string_match(pattern, util.as_string(value))
@classmethod
def string_match(cls, pattern, value):
"""Determine whether the value matches the pattern. Both
arguments are strings. Subclasses implement this method.
"""
raise NotImplementedError()
class SubstringQuery(StringFieldQuery):
"""A query that matches a substring in a specific item field."""
def col_clause(self):
pattern = (self.pattern
.replace('\\', '\\\\')
.replace('%', '\\%')
.replace('_', '\\_'))
search = '%' + pattern + '%'
clause = self.field + " like ? escape '\\'"
subvals = [search]
return clause, subvals
@classmethod
def string_match(cls, pattern, value):
return pattern.lower() in value.lower()
class RegexpQuery(StringFieldQuery):
"""A query that matches a regular expression in a specific item
field.
"""
@classmethod
def string_match(cls, pattern, value):
try:
res = re.search(pattern, value)
except re.error:
# Invalid regular expression.
return False
return res is not None
class BooleanQuery(MatchQuery):
"""Matches a boolean field. Pattern should either be a boolean or a
string reflecting a boolean.
"""
def __init__(self, field, pattern, fast=True):
super(BooleanQuery, self).__init__(field, pattern, fast)
if isinstance(pattern, basestring):
self.pattern = util.str2bool(pattern)
self.pattern = int(self.pattern)
class BytesQuery(MatchQuery):
"""Match a raw bytes field (i.e., a path). This is a necessary hack
to work around the `sqlite3` module's desire to treat `str` and
`unicode` equivalently in Python 2. Always use this query instead of
`MatchQuery` when matching on BLOB values.
"""
def __init__(self, field, pattern):
super(BytesQuery, self).__init__(field, pattern)
# Use a buffer representation of the pattern for SQLite
# matching. This instructs SQLite to treat the blob as binary
# rather than encoded Unicode.
if isinstance(self.pattern, basestring):
# Implicitly coerce Unicode strings to their bytes
# equivalents.
if isinstance(self.pattern, unicode):
self.pattern = self.pattern.encode('utf8')
self.buf_pattern = buffer(self.pattern)
elif isinstance(self.pattern, buffer):
self.buf_pattern = self.pattern
self.pattern = bytes(self.pattern)
def col_clause(self):
return self.field + " = ?", [self.buf_pattern]
class NumericQuery(FieldQuery):
"""Matches numeric fields. A syntax using Ruby-style range ellipses
(``..``) lets users specify one- or two-sided ranges. For example,
``year:2001..`` finds music released since the turn of the century.
"""
def _convert(self, s):
"""Convert a string to a numeric type (float or int). If the
string cannot be converted, return None.
"""
# This is really just a bit of fun premature optimization.
try:
return int(s)
except ValueError:
try:
return float(s)
except ValueError:
return None
def __init__(self, field, pattern, fast=True):
super(NumericQuery, self).__init__(field, pattern, fast)
parts = pattern.split('..', 1)
if len(parts) == 1:
# No range.
self.point = self._convert(parts[0])
self.rangemin = None
self.rangemax = None
else:
# One- or two-sided range.
self.point = None
self.rangemin = self._convert(parts[0])
self.rangemax = self._convert(parts[1])
def match(self, item):
if self.field not in item:
return False
value = item[self.field]
if isinstance(value, basestring):
value = self._convert(value)
if self.point is not None:
return value == self.point
else:
if self.rangemin is not None and value < self.rangemin:
return False
if self.rangemax is not None and value > self.rangemax:
return False
return True
def col_clause(self):
if self.point is not None:
return self.field + '=?', (self.point,)
else:
if self.rangemin is not None and self.rangemax is not None:
return (u'{0} >= ? AND {0} <= ?'.format(self.field),
(self.rangemin, self.rangemax))
elif self.rangemin is not None:
return u'{0} >= ?'.format(self.field), (self.rangemin,)
elif self.rangemax is not None:
return u'{0} <= ?'.format(self.field), (self.rangemax,)
else:
return '1', ()
class CollectionQuery(Query):
"""An abstract query class that aggregates other queries. Can be
indexed like a list to access the sub-queries.
"""
def __init__(self, subqueries=()):
self.subqueries = subqueries
# Act like a sequence.
def __len__(self):
return len(self.subqueries)
def __getitem__(self, key):
return self.subqueries[key]
def __iter__(self):
return iter(self.subqueries)
def __contains__(self, item):
return item in self.subqueries
def clause_with_joiner(self, joiner):
"""Returns a clause created by joining together the clauses of
all subqueries with the string joiner (padded by spaces).
"""
clause_parts = []
subvals = []
for subq in self.subqueries:
subq_clause, subq_subvals = subq.clause()
if not subq_clause:
# Fall back to slow query.
return None, ()
clause_parts.append('(' + subq_clause + ')')
subvals += subq_subvals
clause = (' ' + joiner + ' ').join(clause_parts)
return clause, subvals
class AnyFieldQuery(CollectionQuery):
"""A query that matches if a given FieldQuery subclass matches in
any field. The individual field query class is provided to the
constructor.
"""
def __init__(self, pattern, fields, cls):
self.pattern = pattern
self.fields = fields
self.query_class = cls
subqueries = []
for field in self.fields:
subqueries.append(cls(field, pattern, True))
super(AnyFieldQuery, self).__init__(subqueries)
def clause(self):
return self.clause_with_joiner('or')
def match(self, item):
for subq in self.subqueries:
if subq.match(item):
return True
return False
class MutableCollectionQuery(CollectionQuery):
"""A collection query whose subqueries may be modified after the
query is initialized.
"""
def __setitem__(self, key, value):
self.subqueries[key] = value
def __delitem__(self, key):
del self.subqueries[key]
class AndQuery(MutableCollectionQuery):
"""A conjunction of a list of other queries."""
def clause(self):
return self.clause_with_joiner('and')
def match(self, item):
return all([q.match(item) for q in self.subqueries])
class OrQuery(MutableCollectionQuery):
"""A conjunction of a list of other queries."""
def clause(self):
return self.clause_with_joiner('or')
def match(self, item):
return any([q.match(item) for q in self.subqueries])
class TrueQuery(Query):
"""A query that always matches."""
def clause(self):
return '1', ()
def match(self, item):
return True
class FalseQuery(Query):
"""A query that never matches."""
def clause(self):
return '0', ()
def match(self, item):
return False
# Time/date queries.
def _to_epoch_time(date):
"""Convert a `datetime` object to an integer number of seconds since
the (local) Unix epoch.
"""
epoch = datetime.fromtimestamp(0)
delta = date - epoch
try:
return int(delta.total_seconds())
except AttributeError:
# datetime.timedelta.total_seconds() is not available on Python 2.6
return delta.seconds + delta.days * 24 * 3600
def _parse_periods(pattern):
"""Parse a string containing two dates separated by two dots (..).
Return a pair of `Period` objects.
"""
parts = pattern.split('..', 1)
if len(parts) == 1:
instant = Period.parse(parts[0])
return (instant, instant)
else:
start = Period.parse(parts[0])
end = Period.parse(parts[1])
return (start, end)
class Period(object):
"""A period of time given by a date, time and precision.
Example: 2014-01-01 10:50:30 with precision 'month' represents all
instants of time during January 2014.
"""
precisions = ('year', 'month', 'day')
date_formats = ('%Y', '%Y-%m', '%Y-%m-%d')
def __init__(self, date, precision):
"""Create a period with the given date (a `datetime` object) and
precision (a string, one of "year", "month", or "day").
"""
if precision not in Period.precisions:
raise ValueError('Invalid precision ' + str(precision))
self.date = date
self.precision = precision
@classmethod
def parse(cls, string):
"""Parse a date and return a `Period` object or `None` if the
string is empty.
"""
if not string:
return None
ordinal = string.count('-')
if ordinal >= len(cls.date_formats):
# Too many components.
return None
date_format = cls.date_formats[ordinal]
try:
date = datetime.strptime(string, date_format)
except ValueError:
# Parsing failed.
return None
precision = cls.precisions[ordinal]
return cls(date, precision)
def open_right_endpoint(self):
"""Based on the precision, convert the period to a precise
`datetime` for use as a right endpoint in a right-open interval.
"""
precision = self.precision
date = self.date
if 'year' == self.precision:
return date.replace(year=date.year + 1, month=1)
elif 'month' == precision:
if (date.month < 12):
return date.replace(month=date.month + 1)
else:
return date.replace(year=date.year + 1, month=1)
elif 'day' == precision:
return date + timedelta(days=1)
else:
raise ValueError('unhandled precision ' + str(precision))
class DateInterval(object):
"""A closed-open interval of dates.
A left endpoint of None means since the beginning of time.
A right endpoint of None means towards infinity.
"""
def __init__(self, start, end):
if start is not None and end is not None and not start < end:
raise ValueError("start date {0} is not before end date {1}"
.format(start, end))
self.start = start
self.end = end
@classmethod
def from_periods(cls, start, end):
"""Create an interval with two Periods as the endpoints.
"""
end_date = end.open_right_endpoint() if end is not None else None
start_date = start.date if start is not None else None
return cls(start_date, end_date)
def contains(self, date):
if self.start is not None and date < self.start:
return False
if self.end is not None and date >= self.end:
return False
return True
def __str__(self):
return'[{0}, {1})'.format(self.start, self.end)
class DateQuery(FieldQuery):
"""Matches date fields stored as seconds since Unix epoch time.
Dates can be specified as ``year-month-day`` strings where only year
is mandatory.
The value of a date field can be matched against a date interval by
using an ellipsis interval syntax similar to that of NumericQuery.
"""
def __init__(self, field, pattern, fast=True):
super(DateQuery, self).__init__(field, pattern, fast)
start, end = _parse_periods(pattern)
self.interval = DateInterval.from_periods(start, end)
def match(self, item):
timestamp = float(item[self.field])
date = datetime.utcfromtimestamp(timestamp)
return self.interval.contains(date)
_clause_tmpl = "{0} {1} ?"
def col_clause(self):
clause_parts = []
subvals = []
if self.interval.start:
clause_parts.append(self._clause_tmpl.format(self.field, ">="))
subvals.append(_to_epoch_time(self.interval.start))
if self.interval.end:
clause_parts.append(self._clause_tmpl.format(self.field, "<"))
subvals.append(_to_epoch_time(self.interval.end))
if clause_parts:
# One- or two-sided interval.
clause = ' AND '.join(clause_parts)
else:
# Match any date.
clause = '1'
return clause, subvals
# Sorting.
class Sort(object):
"""An abstract class representing a sort operation for a query into
the item database.
"""
def order_clause(self):
"""Generates a SQL fragment to be used in a ORDER BY clause, or
None if no fragment is used (i.e., this is a slow sort).
"""
return None
def sort(self, items):
"""Sort the list of objects and return a list.
"""
return sorted(items)
def is_slow(self):
"""Indicate whether this query is *slow*, meaning that it cannot
be executed in SQL and must be executed in Python.
"""
return False
class MultipleSort(Sort):
"""Sort that encapsulates multiple sub-sorts.
"""
def __init__(self, sorts=None):
self.sorts = sorts or []
def add_sort(self, sort):
self.sorts.append(sort)
def _sql_sorts(self):
"""Return the list of sub-sorts for which we can be (at least
partially) fast.
A contiguous suffix of fast (SQL-capable) sub-sorts are
executable in SQL. The remaining, even if they are fast
independently, must be executed slowly.
"""
sql_sorts = []
for sort in reversed(self.sorts):
if not sort.order_clause() is None:
sql_sorts.append(sort)
else:
break
sql_sorts.reverse()
return sql_sorts
def order_clause(self):
order_strings = []
for sort in self._sql_sorts():
order = sort.order_clause()
order_strings.append(order)
return ", ".join(order_strings)
def is_slow(self):
for sort in self.sorts:
if sort.is_slow():
return True
return False
def sort(self, items):
slow_sorts = []
switch_slow = False
for sort in reversed(self.sorts):
if switch_slow:
slow_sorts.append(sort)
elif sort.order_clause() is None:
switch_slow = True
slow_sorts.append(sort)
else:
pass
for sort in slow_sorts:
items = sort.sort(items)
return items
def __repr__(self):
return u'MultipleSort({0})'.format(repr(self.sorts))
class FieldSort(Sort):
"""An abstract sort criterion that orders by a specific field (of
any kind).
"""
def __init__(self, field, ascending=True):
self.field = field
self.ascending = ascending
def sort(self, objs):
# TODO: Conversion and null-detection here. In Python 3,
# comparisons with None fail. We should also support flexible
# attributes with different types without falling over.
return sorted(objs, key=attrgetter(self.field),
reverse=not self.ascending)
def __repr__(self):
return u'<{0}: {1}{2}>'.format(
type(self).__name__,
self.field,
'+' if self.ascending else '-',
)
class FixedFieldSort(FieldSort):
"""Sort object to sort on a fixed field.
"""
def order_clause(self):
order = "ASC" if self.ascending else "DESC"
return "{0} {1}".format(self.field, order)
class SlowFieldSort(FieldSort):
"""A sort criterion by some model field other than a fixed field:
i.e., a computed or flexible field.
"""
def is_slow(self):
return True
class NullSort(Sort):
"""No sorting. Leave results unsorted."""
def sort(items):
return items
lib/beets/dbcore/queryparse.py
-180
View File
@@ -1,180 +0,0 @@
# This file is part of beets.
# Copyright 2014, Adrian Sampson.
#
# Permission is hereby granted, free of charge, to any person obtaining
# a copy of this software and associated documentation files (the
# "Software"), to deal in the Software without restriction, including
# without limitation the rights to use, copy, modify, merge, publish,
# distribute, sublicense, and/or sell copies of the Software, and to
# permit persons to whom the Software is furnished to do so, subject to
# the following conditions:
#
# The above copyright notice and this permission notice shall be
# included in all copies or substantial portions of the Software.
"""Parsing of strings into DBCore queries.
"""
import re
import itertools
from . import query
PARSE_QUERY_PART_REGEX = re.compile(
# Non-capturing optional segment for the keyword.
r'(?:'
r'(\S+?)' # The field key.
r'(?<!\\):' # Unescaped :
r')?'
r'(.*)', # The term itself.
re.I # Case-insensitive.
)
def parse_query_part(part, query_classes={}, prefixes={},
default_class=query.SubstringQuery):
"""Take a query in the form of a key/value pair separated by a
colon and return a tuple of `(key, value, cls)`. `key` may be None,
indicating that any field may be matched. `cls` is a subclass of
`FieldQuery`.
The optional `query_classes` parameter maps field names to default
query types; `default_class` is the fallback. `prefixes` is a map
from query prefix markers and query types. Prefix-indicated queries
take precedence over type-based queries.
To determine the query class, two factors are used: prefixes and
field types. For example, the colon prefix denotes a regular
expression query and a type map might provide a special kind of
query for numeric values. If neither a prefix nor a specific query
class is available, `default_class` is used.
For instance,
'stapler' -> (None, 'stapler', SubstringQuery)
'color:red' -> ('color', 'red', SubstringQuery)
':^Quiet' -> (None, '^Quiet', RegexpQuery)
'color::b..e' -> ('color', 'b..e', RegexpQuery)
Prefixes may be "escaped" with a backslash to disable the keying
behavior.
"""
part = part.strip()
match = PARSE_QUERY_PART_REGEX.match(part)
assert match # Regex should always match.
key = match.group(1)
term = match.group(2).replace('\:', ':')
# Match the search term against the list of prefixes.
for pre, query_class in prefixes.items():
if term.startswith(pre):
return key, term[len(pre):], query_class
# No matching prefix: use type-based or fallback/default query.
query_class = query_classes.get(key, default_class)
return key, term, query_class
def construct_query_part(model_cls, prefixes, query_part):
"""Create a query from a single query component, `query_part`, for
querying instances of `model_cls`. Return a `Query` instance.
"""
# Shortcut for empty query parts.
if not query_part:
return query.TrueQuery()
# Get the query classes for each possible field.
query_classes = {}
for k, t in itertools.chain(model_cls._fields.items(),
model_cls._types.items()):
query_classes[k] = t.query
# Parse the string.
key, pattern, query_class = \
parse_query_part(query_part, query_classes, prefixes)
# No key specified.
if key is None:
if issubclass(query_class, query.FieldQuery):
# The query type matches a specific field, but none was
# specified. So we use a version of the query that matches
# any field.
return query.AnyFieldQuery(pattern, model_cls._search_fields,
query_class)
else:
# Other query type.
return query_class(pattern)
key = key.lower()
return query_class(key.lower(), pattern, key in model_cls._fields)
def query_from_strings(query_cls, model_cls, prefixes, query_parts):
"""Creates a collection query of type `query_cls` from a list of
strings in the format used by parse_query_part. `model_cls`
determines how queries are constructed from strings.
"""
subqueries = []
for part in query_parts:
subqueries.append(construct_query_part(model_cls, prefixes, part))
if not subqueries: # No terms in query.
subqueries = [query.TrueQuery()]
return query_cls(subqueries)
def construct_sort_part(model_cls, part):
"""Create a `Sort` from a single string criterion.
`model_cls` is the `Model` being queried. `part` is a single string
ending in ``+`` or ``-`` indicating the sort.
"""
assert part, "part must be a field name and + or -"
field = part[:-1]
assert field, "field is missing"
direction = part[-1]
assert direction in ('+', '-'), "part must end with + or -"
is_ascending = direction == '+'
if field in model_cls._sorts:
sort = model_cls._sorts[field](model_cls, is_ascending)
elif field in model_cls._fields:
sort = query.FixedFieldSort(field, is_ascending)
else:
# Flexible or computed.
sort = query.SlowFieldSort(field, is_ascending)
return sort
def sort_from_strings(model_cls, sort_parts):
"""Create a `Sort` from a list of sort criteria (strings).
"""
if not sort_parts:
return query.NullSort()
else:
sort = query.MultipleSort()
for part in sort_parts:
sort.add_sort(construct_sort_part(model_cls, part))
return sort
def parse_sorted_query(model_cls, parts, prefixes={},
query_cls=query.AndQuery):
"""Given a list of strings, create the `Query` and `Sort` that they
represent.
"""
# Separate query token and sort token.
query_parts = []
sort_parts = []
for part in parts:
if part.endswith((u'+', u'-')) and u':' not in part:
sort_parts.append(part)
else:
query_parts.append(part)
# Parse each.
q = query_from_strings(
query_cls, model_cls, prefixes, query_parts
)
s = sort_from_strings(model_cls, sort_parts)
return q, s
lib/beets/dbcore/types.py
-208
View File
@@ -1,208 +0,0 @@
# This file is part of beets.
# Copyright 2014, Adrian Sampson.
#
# Permission is hereby granted, free of charge, to any person obtaining
# a copy of this software and associated documentation files (the
# "Software"), to deal in the Software without restriction, including
# without limitation the rights to use, copy, modify, merge, publish,
# distribute, sublicense, and/or sell copies of the Software, and to
# permit persons to whom the Software is furnished to do so, subject to
# the following conditions:
#
# The above copyright notice and this permission notice shall be
# included in all copies or substantial portions of the Software.
"""Representation of type information for DBCore model fields.
"""
from . import query
from beets.util import str2bool
# Abstract base.
class Type(object):
"""An object encapsulating the type of a model field. Includes
information about how to store, query, format, and parse a given
field.
"""
sql = u'TEXT'
"""The SQLite column type for the value.
"""
query = query.SubstringQuery
"""The `Query` subclass to be used when querying the field.
"""
model_type = unicode
"""The Python type that is used to represent the value in the model.
The model is guaranteed to return a value of this type if the field
is accessed. To this end, the constructor is used by the `normalize`
and `from_sql` methods and the `default` property.
"""
@property
def null(self):
"""The value to be exposed when the underlying value is None.
"""
return self.model_type()
def format(self, value):
"""Given a value of this type, produce a Unicode string
representing the value. This is used in template evaluation.
"""
if value is None:
value = self.null
# `self.null` might be `None`
if value is None:
value = u''
if isinstance(value, bytes):
value = value.decode('utf8', 'ignore')
return unicode(value)
def parse(self, string):
"""Parse a (possibly human-written) string and return the
indicated value of this type.
"""
try:
return self.model_type(string)
except ValueError:
return self.null
def normalize(self, value):
"""Given a value that will be assigned into a field of this
type, normalize the value to have the appropriate type. This
base implementation only reinterprets `None`.
"""
if value is None:
return self.null
else:
# TODO This should eventually be replaced by
# `self.model_type(value)`
return value
def from_sql(self, sql_value):
"""Receives the value stored in the SQL backend and return the
value to be stored in the model.
For fixed fields the type of `value` is determined by the column
type affinity given in the `sql` property and the SQL to Python
mapping of the database adapter. For more information see:
http://www.sqlite.org/datatype3.html
https://docs.python.org/2/library/sqlite3.html#sqlite-and-python-types
Flexible fields have the type afinity `TEXT`. This means the
`sql_value` is either a `buffer` or a `unicode` object` and the
method must handle these in addition.
"""
if isinstance(sql_value, buffer):
sql_value = bytes(sql_value).decode('utf8', 'ignore')
if isinstance(sql_value, unicode):
return self.parse(sql_value)
else:
return self.normalize(sql_value)
def to_sql(self, model_value):
"""Convert a value as stored in the model object to a value used
by the database adapter.
"""
return model_value
# Reusable types.
class Default(Type):
null = None
class Integer(Type):
"""A basic integer type.
"""
sql = u'INTEGER'
query = query.NumericQuery
model_type = int
class PaddedInt(Integer):
"""An integer field that is formatted with a given number of digits,
padded with zeroes.
"""
def __init__(self, digits):
self.digits = digits
def format(self, value):
return u'{0:0{1}d}'.format(value or 0, self.digits)
class ScaledInt(Integer):
"""An integer whose formatting operation scales the number by a
constant and adds a suffix. Good for units with large magnitudes.
"""
def __init__(self, unit, suffix=u''):
self.unit = unit
self.suffix = suffix
def format(self, value):
return u'{0}{1}'.format((value or 0) // self.unit, self.suffix)
class Id(Integer):
"""An integer used as the row id or a foreign key in a SQLite table.
This type is nullable: None values are not translated to zero.
"""
null = None
def __init__(self, primary=True):
if primary:
self.sql = u'INTEGER PRIMARY KEY'
class Float(Type):
"""A basic floating-point type.
"""
sql = u'REAL'
query = query.NumericQuery
model_type = float
def format(self, value):
return u'{0:.1f}'.format(value or 0.0)
class NullFloat(Float):
"""Same as `Float`, but does not normalize `None` to `0.0`.
"""
null = None
class String(Type):
"""A Unicode string type.
"""
sql = u'TEXT'
query = query.SubstringQuery
class Boolean(Type):
"""A boolean type.
"""
sql = u'INTEGER'
query = query.BooleanQuery
model_type = bool
def format(self, value):
return unicode(bool(value))
def parse(self, string):
return str2bool(string)
# Shared instances of common types.
DEFAULT = Default()
INTEGER = Integer()
PRIMARY_ID = Id(True)
FOREIGN_ID = Id(False)
FLOAT = Float()
NULL_FLOAT = NullFloat()
STRING = String()
BOOLEAN = Boolean()
lib/beets/importer.py
-1428
View File
File diff suppressed because it is too large Load Diff
lib/beets/library.py
-1299
View File
File diff suppressed because it is too large Load Diff
lib/beets/mediafile.py
-1929
View File
File diff suppressed because it is too large Load Diff
lib/beets/plugins.py
-435
View File
@@ -1,435 +0,0 @@
# This file is part of beets.
# Copyright 2013, Adrian Sampson.
#
# Permission is hereby granted, free of charge, to any person obtaining
# a copy of this software and associated documentation files (the
# "Software"), to deal in the Software without restriction, including
# without limitation the rights to use, copy, modify, merge, publish,
# distribute, sublicense, and/or sell copies of the Software, and to
# permit persons to whom the Software is furnished to do so, subject to
# the following conditions:
#
# The above copyright notice and this permission notice shall be
# included in all copies or substantial portions of the Software.
"""Support for beets plugins."""
import logging
import traceback
import inspect
import re
from collections import defaultdict
import beets
from beets import mediafile
PLUGIN_NAMESPACE = 'beetsplug'
# Plugins using the Last.fm API can share the same API key.
LASTFM_KEY = '2dc3914abf35f0d9c92d97d8f8e42b43'
# Global logger.
log = logging.getLogger('beets')
class PluginConflictException(Exception):
"""Indicates that the services provided by one plugin conflict with
those of another.
For example two plugins may define different types for flexible fields.
"""
# Managing the plugins themselves.
class BeetsPlugin(object):
"""The base class for all beets plugins. Plugins provide
functionality by defining a subclass of BeetsPlugin and overriding
the abstract methods defined here.
"""
def __init__(self, name=None):
"""Perform one-time plugin setup.
"""
self.import_stages = []
self.name = name or self.__module__.split('.')[-1]
self.config = beets.config[self.name]
if not self.template_funcs:
self.template_funcs = {}
if not self.template_fields:
self.template_fields = {}
if not self.album_template_fields:
self.album_template_fields = {}
def commands(self):
"""Should return a list of beets.ui.Subcommand objects for
commands that should be added to beets' CLI.
"""
return ()
def queries(self):
"""Should return a dict mapping prefixes to Query subclasses.
"""
return {}
def track_distance(self, item, info):
"""Should return a Distance object to be added to the
distance for every track comparison.
"""
return beets.autotag.hooks.Distance()
def album_distance(self, items, album_info, mapping):
"""Should return a Distance object to be added to the
distance for every album-level comparison.
"""
return beets.autotag.hooks.Distance()
def candidates(self, items, artist, album, va_likely):
"""Should return a sequence of AlbumInfo objects that match the
album whose items are provided.
"""
return ()
def item_candidates(self, item, artist, title):
"""Should return a sequence of TrackInfo objects that match the
item provided.
"""
return ()
def album_for_id(self, album_id):
"""Return an AlbumInfo object or None if no matching release was
found.
"""
return None
def track_for_id(self, track_id):
"""Return a TrackInfo object or None if no matching release was
found.
"""
return None
def add_media_field(self, name, descriptor):
"""Add a field that is synchronized between media files and items.
When a media field is added ``item.write()`` will set the name
property of the item's MediaFile to ``item[name]`` and save the
changes. Similarly ``item.read()`` will set ``item[name]`` to
the value of the name property of the media file.
``descriptor`` must be an instance of ``mediafile.MediaField``.
"""
# Defer impor to prevent circular dependency
from beets import library
mediafile.MediaFile.add_field(name, descriptor)
library.Item._media_fields.add(name)
listeners = None
@classmethod
def register_listener(cls, event, func):
"""Add a function as a listener for the specified event. (An
imperative alternative to the @listen decorator.)
"""
if cls.listeners is None:
cls.listeners = defaultdict(list)
cls.listeners[event].append(func)
@classmethod
def listen(cls, event):
"""Decorator that adds a function as an event handler for the
specified event (as a string). The parameters passed to function
will vary depending on what event occurred.
The function should respond to named parameters.
function(**kwargs) will trap all arguments in a dictionary.
Example:
>>> @MyPlugin.listen("imported")
>>> def importListener(**kwargs):
... pass
"""
def helper(func):
if cls.listeners is None:
cls.listeners = defaultdict(list)
cls.listeners[event].append(func)
return func
return helper
template_funcs = None
template_fields = None
album_template_fields = None
@classmethod
def template_func(cls, name):
"""Decorator that registers a path template function. The
function will be invoked as ``%name{}`` from path format
strings.
"""
def helper(func):
if cls.template_funcs is None:
cls.template_funcs = {}
cls.template_funcs[name] = func
return func
return helper
@classmethod
def template_field(cls, name):
"""Decorator that registers a path template field computation.
The value will be referenced as ``$name`` from path format
strings. The function must accept a single parameter, the Item
being formatted.
"""
def helper(func):
if cls.template_fields is None:
cls.template_fields = {}
cls.template_fields[name] = func
return func
return helper
_classes = set()
def load_plugins(names=()):
"""Imports the modules for a sequence of plugin names. Each name
must be the name of a Python module under the "beetsplug" namespace
package in sys.path; the module indicated should contain the
BeetsPlugin subclasses desired.
"""
for name in names:
modname = '%s.%s' % (PLUGIN_NAMESPACE, name)
try:
try:
namespace = __import__(modname, None, None)
except ImportError as exc:
# Again, this is hacky:
if exc.args[0].endswith(' ' + name):
log.warn(u'** plugin {0} not found'.format(name))
else:
raise
else:
for obj in getattr(namespace, name).__dict__.values():
if isinstance(obj, type) and issubclass(obj, BeetsPlugin) \
and obj != BeetsPlugin and obj not in _classes:
_classes.add(obj)
except:
log.warn(u'** error loading plugin {0}'.format(name))
log.warn(traceback.format_exc())
_instances = {}
def find_plugins():
"""Returns a list of BeetsPlugin subclass instances from all
currently loaded beets plugins. Loads the default plugin set
first.
"""
load_plugins()
plugins = []
for cls in _classes:
# Only instantiate each plugin class once.
if cls not in _instances:
_instances[cls] = cls()
plugins.append(_instances[cls])
return plugins
# Communication with plugins.
def commands():
"""Returns a list of Subcommand objects from all loaded plugins.
"""
out = []
for plugin in find_plugins():
out += plugin.commands()
return out
def queries():
"""Returns a dict mapping prefix strings to Query subclasses all loaded
plugins.
"""
out = {}
for plugin in find_plugins():
out.update(plugin.queries())
return out
def types(model_cls):
# Gives us `item_types` and `album_types`
attr_name = '{0}_types'.format(model_cls.__name__.lower())
types = {}
for plugin in find_plugins():
plugin_types = getattr(plugin, attr_name, {})
for field in plugin_types:
if field in types and plugin_types[field] != types[field]:
raise PluginConflictException(
u'Plugin {0} defines flexible field {1} '
'which has already been defined with '
'another type.'.format(plugin.name, field)
)
types.update(plugin_types)
return types
def track_distance(item, info):
"""Gets the track distance calculated by all loaded plugins.
Returns a Distance object.
"""
from beets.autotag.hooks import Distance
dist = Distance()
for plugin in find_plugins():
dist.update(plugin.track_distance(item, info))
return dist
def album_distance(items, album_info, mapping):
"""Returns the album distance calculated by plugins."""
from beets.autotag.hooks import Distance
dist = Distance()
for plugin in find_plugins():
dist.update(plugin.album_distance(items, album_info, mapping))
return dist
def candidates(items, artist, album, va_likely):
"""Gets MusicBrainz candidates for an album from each plugin.
"""
out = []
for plugin in find_plugins():
out.extend(plugin.candidates(items, artist, album, va_likely))
return out
def item_candidates(item, artist, title):
"""Gets MusicBrainz candidates for an item from the plugins.
"""
out = []
for plugin in find_plugins():
out.extend(plugin.item_candidates(item, artist, title))
return out
def album_for_id(album_id):
"""Get AlbumInfo objects for a given ID string.
"""
out = []
for plugin in find_plugins():
res = plugin.album_for_id(album_id)
if res:
out.append(res)
return out
def track_for_id(track_id):
"""Get TrackInfo objects for a given ID string.
"""
out = []
for plugin in find_plugins():
res = plugin.track_for_id(track_id)
if res:
out.append(res)
return out
def template_funcs():
"""Get all the template functions declared by plugins as a
dictionary.
"""
funcs = {}
for plugin in find_plugins():
if plugin.template_funcs:
funcs.update(plugin.template_funcs)
return funcs
def import_stages():
"""Get a list of import stage functions defined by plugins."""
stages = []
for plugin in find_plugins():
if hasattr(plugin, 'import_stages'):
stages += plugin.import_stages
return stages
# New-style (lazy) plugin-provided fields.
def item_field_getters():
"""Get a dictionary mapping field names to unary functions that
compute the field's value.
"""
funcs = {}
for plugin in find_plugins():
if plugin.template_fields:
funcs.update(plugin.template_fields)
return funcs
def album_field_getters():
"""As above, for album fields.
"""
funcs = {}
for plugin in find_plugins():
if plugin.album_template_fields:
funcs.update(plugin.album_template_fields)
return funcs
# Event dispatch.
def event_handlers():
"""Find all event handlers from plugins as a dictionary mapping
event names to sequences of callables.
"""
all_handlers = defaultdict(list)
for plugin in find_plugins():
if plugin.listeners:
for event, handlers in plugin.listeners.items():
all_handlers[event] += handlers
return all_handlers
def send(event, **arguments):
"""Sends an event to all assigned event listeners. Event is the
name of the event to send, all other named arguments go to the
event handler(s).
Returns a list of return values from the handlers.
"""
log.debug(u'Sending event: {0}'.format(event))
for handler in event_handlers()[event]:
# Don't break legacy plugins if we want to pass more arguments
argspec = inspect.getargspec(handler).args
args = dict((k, v) for k, v in arguments.items() if k in argspec)
handler(**args)
def feat_tokens(for_artist=True):
"""Return a regular expression that matches phrases like "featuring"
that separate a main artist or a song title from secondary artists.
The `for_artist` option determines whether the regex should be
suitable for matching artist fields (the default) or title fields.
"""
feat_words = ['ft', 'featuring', 'feat', 'feat.', 'ft.']
if for_artist:
feat_words += ['with', 'vs', 'and', 'con', '&']
return '(?<=\s)(?:{0})(?=\s)'.format(
'|'.join(re.escape(x) for x in feat_words)
)
def sanitize_choices(choices, choices_all):
"""Clean up a stringlist configuration attribute: keep only choices
elements present in choices_all, remove duplicate elements, expand '*'
wildcard while keeping original stringlist order.
"""
seen = set()
others = [x for x in choices_all if x not in choices]
res = []
for s in choices:
if s in list(choices_all) + ['*']:
if not (s in seen or seen.add(s)):
res.extend(list(others) if s == '*' else [s])
return res
lib/beets/ui/__init__.py
-970
View File
@@ -1,970 +0,0 @@
# This file is part of beets.
# Copyright 2014, Adrian Sampson.
#
# Permission is hereby granted, free of charge, to any person obtaining
# a copy of this software and associated documentation files (the
# "Software"), to deal in the Software without restriction, including
# without limitation the rights to use, copy, modify, merge, publish,
# distribute, sublicense, and/or sell copies of the Software, and to
# permit persons to whom the Software is furnished to do so, subject to
# the following conditions:
#
# The above copyright notice and this permission notice shall be
# included in all copies or substantial portions of the Software.
"""This module contains all of the core logic for beets' command-line
interface. To invoke the CLI, just call beets.ui.main(). The actual
CLI commands are implemented in the ui.commands module.
"""
from __future__ import print_function
import locale
import optparse
import textwrap
import sys
from difflib import SequenceMatcher
import logging
import sqlite3
import errno
import re
import struct
import traceback
import os.path
from beets import library
from beets import plugins
from beets import util
from beets.util.functemplate import Template
from beets import config
from beets.util import confit
from beets.autotag import mb
# On Windows platforms, use colorama to support "ANSI" terminal colors.
if sys.platform == 'win32':
try:
import colorama
except ImportError:
pass
else:
colorama.init()
log = logging.getLogger('beets')
if not log.handlers:
log.addHandler(logging.StreamHandler())
log.propagate = False # Don't propagate to root handler.
PF_KEY_QUERIES = {
'comp': 'comp:true',
'singleton': 'singleton:true',
}
class UserError(Exception):
"""UI exception. Commands should throw this in order to display
nonrecoverable errors to the user.
"""
# Utilities.
def _encoding():
"""Tries to guess the encoding used by the terminal."""
# Configured override?
encoding = config['terminal_encoding'].get()
if encoding:
return encoding
# Determine from locale settings.
try:
return locale.getdefaultlocale()[1] or 'utf8'
except ValueError:
# Invalid locale environment variable setting. To avoid
# failing entirely for no good reason, assume UTF-8.
return 'utf8'
def decargs(arglist):
"""Given a list of command-line argument bytestrings, attempts to
decode them to Unicode strings.
"""
return [s.decode(_encoding()) for s in arglist]
def print_(*strings):
"""Like print, but rather than raising an error when a character
is not in the terminal's encoding's character set, just silently
replaces it.
"""
if strings:
if isinstance(strings[0], unicode):
txt = u' '.join(strings)
else:
txt = ' '.join(strings)
else:
txt = u''
if isinstance(txt, unicode):
txt = txt.encode(_encoding(), 'replace')
print(txt)
def input_(prompt=None):
"""Like `raw_input`, but decodes the result to a Unicode string.
Raises a UserError if stdin is not available. The prompt is sent to
stdout rather than stderr. A printed between the prompt and the
input cursor.
"""
# raw_input incorrectly sends prompts to stderr, not stdout, so we
# use print() explicitly to display prompts.
# http://bugs.python.org/issue1927
if prompt:
if isinstance(prompt, unicode):
prompt = prompt.encode(_encoding(), 'replace')
print(prompt, end=' ')
try:
resp = raw_input()
except EOFError:
raise UserError('stdin stream ended while input required')
return resp.decode(sys.stdin.encoding or 'utf8', 'ignore')
def input_options(options, require=False, prompt=None, fallback_prompt=None,
numrange=None, default=None, max_width=72):
"""Prompts a user for input. The sequence of `options` defines the
choices the user has. A single-letter shortcut is inferred for each
option; the user's choice is returned as that single, lower-case
letter. The options should be provided as lower-case strings unless
a particular shortcut is desired; in that case, only that letter
should be capitalized.
By default, the first option is the default. `default` can be provided to
override this. If `require` is provided, then there is no default. The
prompt and fallback prompt are also inferred but can be overridden.
If numrange is provided, it is a pair of `(high, low)` (both ints)
indicating that, in addition to `options`, the user may enter an
integer in that inclusive range.
`max_width` specifies the maximum number of columns in the
automatically generated prompt string.
"""
# Assign single letters to each option. Also capitalize the options
# to indicate the letter.
letters = {}
display_letters = []
capitalized = []
first = True
for option in options:
# Is a letter already capitalized?
for letter in option:
if letter.isalpha() and letter.upper() == letter:
found_letter = letter
break
else:
# Infer a letter.
for letter in option:
if not letter.isalpha():
continue # Don't use punctuation.
if letter not in letters:
found_letter = letter
break
else:
raise ValueError('no unambiguous lettering found')
letters[found_letter.lower()] = option
index = option.index(found_letter)
# Mark the option's shortcut letter for display.
if not require and (
(default is None and not numrange and first) or
(isinstance(default, basestring) and
found_letter.lower() == default.lower())):
# The first option is the default; mark it.
show_letter = '[%s]' % found_letter.upper()
is_default = True
else:
show_letter = found_letter.upper()
is_default = False
# Colorize the letter shortcut.
show_letter = colorize('turquoise' if is_default else 'blue',
show_letter)
# Insert the highlighted letter back into the word.
capitalized.append(
option[:index] + show_letter + option[index + 1:]
)
display_letters.append(found_letter.upper())
first = False
# The default is just the first option if unspecified.
if require:
default = None
elif default is None:
if numrange:
default = numrange[0]
else:
default = display_letters[0].lower()
# Make a prompt if one is not provided.
if not prompt:
prompt_parts = []
prompt_part_lengths = []
if numrange:
if isinstance(default, int):
default_name = str(default)
default_name = colorize('turquoise', default_name)
tmpl = '# selection (default %s)'
prompt_parts.append(tmpl % default_name)
prompt_part_lengths.append(len(tmpl % str(default)))
else:
prompt_parts.append('# selection')
prompt_part_lengths.append(len(prompt_parts[-1]))
prompt_parts += capitalized
prompt_part_lengths += [len(s) for s in options]
# Wrap the query text.
prompt = ''
line_length = 0
for i, (part, length) in enumerate(zip(prompt_parts,
prompt_part_lengths)):
# Add punctuation.
if i == len(prompt_parts) - 1:
part += '?'
else:
part += ','
length += 1
# Choose either the current line or the beginning of the next.
if line_length + length + 1 > max_width:
prompt += '\n'
line_length = 0
if line_length != 0:
# Not the beginning of the line; need a space.
part = ' ' + part
length += 1
prompt += part
line_length += length
# Make a fallback prompt too. This is displayed if the user enters
# something that is not recognized.
if not fallback_prompt:
fallback_prompt = 'Enter one of '
if numrange:
fallback_prompt += '%i-%i, ' % numrange
fallback_prompt += ', '.join(display_letters) + ':'
resp = input_(prompt)
while True:
resp = resp.strip().lower()
# Try default option.
if default is not None and not resp:
resp = default
# Try an integer input if available.
if numrange:
try:
resp = int(resp)
except ValueError:
pass
else:
low, high = numrange
if low <= resp <= high:
return resp
else:
resp = None
# Try a normal letter input.
if resp:
resp = resp[0]
if resp in letters:
return resp
# Prompt for new input.
resp = input_(fallback_prompt)
def input_yn(prompt, require=False):
"""Prompts the user for a "yes" or "no" response. The default is
"yes" unless `require` is `True`, in which case there is no default.
"""
sel = input_options(
('y', 'n'), require, prompt, 'Enter Y or N:'
)
return sel == 'y'
def human_bytes(size):
"""Formats size, a number of bytes, in a human-readable way."""
suffices = ['B', 'KB', 'MB', 'GB', 'TB', 'PB', 'EB', 'ZB', 'YB', 'HB']
for suffix in suffices:
if size < 1024:
return "%3.1f %s" % (size, suffix)
size /= 1024.0
return "big"
def human_seconds(interval):
"""Formats interval, a number of seconds, as a human-readable time
interval using English words.
"""
units = [
(1, 'second'),
(60, 'minute'),
(60, 'hour'),
(24, 'day'),
(7, 'week'),
(52, 'year'),
(10, 'decade'),
]
for i in range(len(units) - 1):
increment, suffix = units[i]
next_increment, _ = units[i + 1]
interval /= float(increment)
if interval < next_increment:
break
else:
# Last unit.
increment, suffix = units[-1]
interval /= float(increment)
return "%3.1f %ss" % (interval, suffix)
def human_seconds_short(interval):
"""Formats a number of seconds as a short human-readable M:SS
string.
"""
interval = int(interval)
return u'%i:%02i' % (interval // 60, interval % 60)
# ANSI terminal colorization code heavily inspired by pygments:
# http://dev.pocoo.org/hg/pygments-main/file/b2deea5b5030/pygments/console.py
# (pygments is by Tim Hatch, Armin Ronacher, et al.)
COLOR_ESCAPE = "\x1b["
DARK_COLORS = ["black", "darkred", "darkgreen", "brown", "darkblue",
"purple", "teal", "lightgray"]
LIGHT_COLORS = ["darkgray", "red", "green", "yellow", "blue",
"fuchsia", "turquoise", "white"]
RESET_COLOR = COLOR_ESCAPE + "39;49;00m"
def _colorize(color, text):
"""Returns a string that prints the given text in the given color
in a terminal that is ANSI color-aware. The color must be something
in DARK_COLORS or LIGHT_COLORS.
"""
if color in DARK_COLORS:
escape = COLOR_ESCAPE + "%im" % (DARK_COLORS.index(color) + 30)
elif color in LIGHT_COLORS:
escape = COLOR_ESCAPE + "%i;01m" % (LIGHT_COLORS.index(color) + 30)
else:
raise ValueError('no such color %s', color)
return escape + text + RESET_COLOR
def colorize(color, text):
"""Colorize text if colored output is enabled. (Like _colorize but
conditional.)
"""
if config['color']:
return _colorize(color, text)
else:
return text
def _colordiff(a, b, highlight='red', minor_highlight='lightgray'):
"""Given two values, return the same pair of strings except with
their differences highlighted in the specified color. Strings are
highlighted intelligently to show differences; other values are
stringified and highlighted in their entirety.
"""
if not isinstance(a, basestring) or not isinstance(b, basestring):
# Non-strings: use ordinary equality.
a = unicode(a)
b = unicode(b)
if a == b:
return a, b
else:
return colorize(highlight, a), colorize(highlight, b)
if isinstance(a, bytes) or isinstance(b, bytes):
# A path field.
a = util.displayable_path(a)
b = util.displayable_path(b)
a_out = []
b_out = []
matcher = SequenceMatcher(lambda x: False, a, b)
for op, a_start, a_end, b_start, b_end in matcher.get_opcodes():
if op == 'equal':
# In both strings.
a_out.append(a[a_start:a_end])
b_out.append(b[b_start:b_end])
elif op == 'insert':
# Right only.
b_out.append(colorize(highlight, b[b_start:b_end]))
elif op == 'delete':
# Left only.
a_out.append(colorize(highlight, a[a_start:a_end]))
elif op == 'replace':
# Right and left differ. Colorise with second highlight if
# it's just a case change.
if a[a_start:a_end].lower() != b[b_start:b_end].lower():
color = highlight
else:
color = minor_highlight
a_out.append(colorize(color, a[a_start:a_end]))
b_out.append(colorize(color, b[b_start:b_end]))
else:
assert(False)
return u''.join(a_out), u''.join(b_out)
def colordiff(a, b, highlight='red'):
"""Colorize differences between two values if color is enabled.
(Like _colordiff but conditional.)
"""
if config['color']:
return _colordiff(a, b, highlight)
else:
return unicode(a), unicode(b)
def get_path_formats(subview=None):
"""Get the configuration's path formats as a list of query/template
pairs.
"""
path_formats = []
subview = subview or config['paths']
for query, view in subview.items():
query = PF_KEY_QUERIES.get(query, query) # Expand common queries.
path_formats.append((query, Template(view.get(unicode))))
return path_formats
def get_replacements():
"""Confit validation function that reads regex/string pairs.
"""
replacements = []
for pattern, repl in config['replace'].get(dict).items():
repl = repl or ''
try:
replacements.append((re.compile(pattern), repl))
except re.error:
raise UserError(
u'malformed regular expression in replace: {0}'.format(
pattern
)
)
return replacements
def _pick_format(album, fmt=None):
"""Pick a format string for printing Album or Item objects,
falling back to config options and defaults.
"""
if fmt:
return fmt
if album:
return config['list_format_album'].get(unicode)
else:
return config['list_format_item'].get(unicode)
def print_obj(obj, lib, fmt=None):
"""Print an Album or Item object. If `fmt` is specified, use that
format string. Otherwise, use the configured template.
"""
album = isinstance(obj, library.Album)
fmt = _pick_format(album, fmt)
if isinstance(fmt, Template):
template = fmt
else:
template = Template(fmt)
print_(obj.evaluate_template(template))
def term_width():
"""Get the width (columns) of the terminal."""
fallback = config['ui']['terminal_width'].get(int)
# The fcntl and termios modules are not available on non-Unix
# platforms, so we fall back to a constant.
try:
import fcntl
import termios
except ImportError:
return fallback
try:
buf = fcntl.ioctl(0, termios.TIOCGWINSZ, ' ' * 4)
except IOError:
return fallback
try:
height, width = struct.unpack('hh', buf)
except struct.error:
return fallback
return width
FLOAT_EPSILON = 0.01
def _field_diff(field, old, new):
"""Given two Model objects, format their values for `field` and
highlight changes among them. Return a human-readable string. If the
value has not changed, return None instead.
"""
oldval = old.get(field)
newval = new.get(field)
# If no change, abort.
if isinstance(oldval, float) and isinstance(newval, float) and \
abs(oldval - newval) < FLOAT_EPSILON:
return None
elif oldval == newval:
return None
# Get formatted values for output.
oldstr = old.formatted().get(field, u'')
newstr = new.formatted().get(field, u'')
# For strings, highlight changes. For others, colorize the whole
# thing.
if isinstance(oldval, basestring):
oldstr, newstr = colordiff(oldval, newstr)
else:
oldstr, newstr = colorize('red', oldstr), colorize('red', newstr)
return u'{0} -> {1}'.format(oldstr, newstr)
def show_model_changes(new, old=None, fields=None, always=False):
"""Given a Model object, print a list of changes from its pristine
version stored in the database. Return a boolean indicating whether
any changes were found.
`old` may be the "original" object to avoid using the pristine
version from the database. `fields` may be a list of fields to
restrict the detection to. `always` indicates whether the object is
always identified, regardless of whether any changes are present.
"""
old = old or new._db._get(type(new), new.id)
# Build up lines showing changed fields.
changes = []
for field in old:
# Subset of the fields. Never show mtime.
if field == 'mtime' or (fields and field not in fields):
continue
# Detect and show difference for this field.
line = _field_diff(field, old, new)
if line:
changes.append(u' {0}: {1}'.format(field, line))
# New fields.
for field in set(new) - set(old):
if fields and field not in fields:
continue
changes.append(u' {0}: {1}'.format(
field,
colorize('red', new.formatted()[field])
))
# Print changes.
if changes or always:
print_obj(old, old._db)
if changes:
print_(u'\n'.join(changes))
return bool(changes)
# Subcommand parsing infrastructure.
#
# This is a fairly generic subcommand parser for optparse. It is
# maintained externally here:
# http://gist.github.com/462717
# There you will also find a better description of the code and a more
# succinct example program.
class Subcommand(object):
"""A subcommand of a root command-line application that may be
invoked by a SubcommandOptionParser.
"""
def __init__(self, name, parser=None, help='', aliases=(), hide=False):
"""Creates a new subcommand. name is the primary way to invoke
the subcommand; aliases are alternate names. parser is an
OptionParser responsible for parsing the subcommand's options.
help is a short description of the command. If no parser is
given, it defaults to a new, empty OptionParser.
"""
self.name = name
self.parser = parser or optparse.OptionParser()
self.aliases = aliases
self.help = help
self.hide = hide
self._root_parser = None
def print_help(self):
self.parser.print_help()
def parse_args(self, args):
return self.parser.parse_args(args)
@property
def root_parser(self):
return self._root_parser
@root_parser.setter
def root_parser(self, root_parser):
self._root_parser = root_parser
self.parser.prog = '{0} {1}'.format(root_parser.get_prog_name(),
self.name)
class SubcommandsOptionParser(optparse.OptionParser):
"""A variant of OptionParser that parses subcommands and their
arguments.
"""
def __init__(self, *args, **kwargs):
"""Create a new subcommand-aware option parser. All of the
options to OptionParser.__init__ are supported in addition
to subcommands, a sequence of Subcommand objects.
"""
# A more helpful default usage.
if 'usage' not in kwargs:
kwargs['usage'] = """
%prog COMMAND [ARGS...]
%prog help COMMAND"""
kwargs['add_help_option'] = False
# Super constructor.
optparse.OptionParser.__init__(self, *args, **kwargs)
# Our root parser needs to stop on the first unrecognized argument.
self.disable_interspersed_args()
self.subcommands = []
def add_subcommand(self, *cmds):
"""Adds a Subcommand object to the parser's list of commands.
"""
for cmd in cmds:
cmd.root_parser = self
self.subcommands.append(cmd)
# Add the list of subcommands to the help message.
def format_help(self, formatter=None):
# Get the original help message, to which we will append.
out = optparse.OptionParser.format_help(self, formatter)
if formatter is None:
formatter = self.formatter
# Subcommands header.
result = ["\n"]
result.append(formatter.format_heading('Commands'))
formatter.indent()
# Generate the display names (including aliases).
# Also determine the help position.
disp_names = []
help_position = 0
subcommands = [c for c in self.subcommands if not c.hide]
subcommands.sort(key=lambda c: c.name)
for subcommand in subcommands:
name = subcommand.name
if subcommand.aliases:
name += ' (%s)' % ', '.join(subcommand.aliases)
disp_names.append(name)
# Set the help position based on the max width.
proposed_help_position = len(name) + formatter.current_indent + 2
if proposed_help_position <= formatter.max_help_position:
help_position = max(help_position, proposed_help_position)
# Add each subcommand to the output.
for subcommand, name in zip(subcommands, disp_names):
# Lifted directly from optparse.py.
name_width = help_position - formatter.current_indent - 2
if len(name) > name_width:
name = "%*s%s\n" % (formatter.current_indent, "", name)
indent_first = help_position
else:
name = "%*s%-*s " % (formatter.current_indent, "",
name_width, name)
indent_first = 0
result.append(name)
help_width = formatter.width - help_position
help_lines = textwrap.wrap(subcommand.help, help_width)
result.append("%*s%s\n" % (indent_first, "", help_lines[0]))
result.extend(["%*s%s\n" % (help_position, "", line)
for line in help_lines[1:]])
formatter.dedent()
# Concatenate the original help message with the subcommand
# list.
return out + "".join(result)
def _subcommand_for_name(self, name):
"""Return the subcommand in self.subcommands matching the
given name. The name may either be the name of a subcommand or
an alias. If no subcommand matches, returns None.
"""
for subcommand in self.subcommands:
if name == subcommand.name or \
name in subcommand.aliases:
return subcommand
return None
def parse_global_options(self, args):
"""Parse options up to the subcommand argument. Returns a tuple
of the options object and the remaining arguments.
"""
options, subargs = self.parse_args(args)
# Force the help command
if options.help:
subargs = ['help']
elif options.version:
subargs = ['version']
return options, subargs
def parse_subcommand(self, args):
"""Given the `args` left unused by a `parse_global_options`,
return the invoked subcommand, the subcommand options, and the
subcommand arguments.
"""
# Help is default command
if not args:
args = ['help']
cmdname = args.pop(0)
subcommand = self._subcommand_for_name(cmdname)
if not subcommand:
raise UserError("unknown command '{0}'".format(cmdname))
suboptions, subargs = subcommand.parse_args(args)
return subcommand, suboptions, subargs
optparse.Option.ALWAYS_TYPED_ACTIONS += ('callback',)
def vararg_callback(option, opt_str, value, parser):
"""Callback for an option with variable arguments.
Manually collect arguments right of a callback-action
option (ie. with action="callback"), and add the resulting
list to the destination var.
Usage:
parser.add_option("-c", "--callback", dest="vararg_attr",
action="callback", callback=vararg_callback)
Details:
http://docs.python.org/2/library/optparse.html#callback-example-6-variable
-arguments
"""
value = [value]
def floatable(str):
try:
float(str)
return True
except ValueError:
return False
for arg in parser.rargs:
# stop on --foo like options
if arg[:2] == "--" and len(arg) > 2:
break
# stop on -a, but not on -3 or -3.0
if arg[:1] == "-" and len(arg) > 1 and not floatable(arg):
break
value.append(arg)
del parser.rargs[:len(value) - 1]
setattr(parser.values, option.dest, value)
# The main entry point and bootstrapping.
def _load_plugins(config):
"""Load the plugins specified in the configuration.
"""
paths = config['pluginpath'].get(confit.StrSeq(split=False))
paths = map(util.normpath, paths)
import beetsplug
beetsplug.__path__ = paths + beetsplug.__path__
# For backwards compatibility.
sys.path += paths
plugins.load_plugins(config['plugins'].as_str_seq())
plugins.send("pluginload")
return plugins
def _setup(options, lib=None):
"""Prepare and global state and updates it with command line options.
Returns a list of subcommands, a list of plugins, and a library instance.
"""
# Configure the MusicBrainz API.
mb.configure()
config = _configure(options)
plugins = _load_plugins(config)
# Get the default subcommands.
from beets.ui.commands import default_commands
subcommands = list(default_commands)
subcommands.extend(plugins.commands())
if lib is None:
lib = _open_library(config)
plugins.send("library_opened", lib=lib)
library.Item._types = plugins.types(library.Item)
library.Album._types = plugins.types(library.Album)
return subcommands, plugins, lib
def _configure(options):
"""Amend the global configuration object with command line options.
"""
# Add any additional config files specified with --config. This
# special handling lets specified plugins get loaded before we
# finish parsing the command line.
if getattr(options, 'config', None) is not None:
config_path = options.config
del options.config
config.set_file(config_path)
config.set_args(options)
# Configure the logger.
if config['verbose'].get(bool):
log.setLevel(logging.DEBUG)
else:
log.setLevel(logging.INFO)
config_path = config.user_config_path()
if os.path.isfile(config_path):
log.debug(u'user configuration: {0}'.format(
util.displayable_path(config_path)))
else:
log.debug(u'no user configuration found at {0}'.format(
util.displayable_path(config_path)))
log.debug(u'data directory: {0}'
.format(util.displayable_path(config.config_dir())))
return config
def _open_library(config):
"""Create a new library instance from the configuration.
"""
dbpath = config['library'].as_filename()
try:
lib = library.Library(
dbpath,
config['directory'].as_filename(),
get_path_formats(),
get_replacements(),
)
lib.get_item(0) # Test database connection.
except (sqlite3.OperationalError, sqlite3.DatabaseError):
log.debug(traceback.format_exc())
raise UserError(u"database file {0} could not be opened".format(
util.displayable_path(dbpath)
))
log.debug(u'library database: {0}\n'
u'library directory: {1}'
.format(util.displayable_path(lib.path),
util.displayable_path(lib.directory)))
return lib
def _raw_main(args, lib=None):
"""A helper function for `main` without top-level exception
handling.
"""
parser = SubcommandsOptionParser()
parser.add_option('-l', '--library', dest='library',
help='library database file to use')
parser.add_option('-d', '--directory', dest='directory',
help="destination music directory")
parser.add_option('-v', '--verbose', dest='verbose', action='store_true',
help='print debugging information')
parser.add_option('-c', '--config', dest='config',
help='path to configuration file')
parser.add_option('-h', '--help', dest='help', action='store_true',
help='how this help message and exit')
parser.add_option('--version', dest='version', action='store_true',
help=optparse.SUPPRESS_HELP)
options, subargs = parser.parse_global_options(args)
# Special case for the `config --edit` command: bypass _setup so
# that an invalid configuration does not prevent the editor from
# starting.
if subargs[0] == 'config' and ('-e' in subargs or '--edit' in subargs):
from beets.ui.commands import config_edit
return config_edit()
subcommands, plugins, lib = _setup(options, lib)
parser.add_subcommand(*subcommands)
subcommand, suboptions, subargs = parser.parse_subcommand(subargs)
subcommand.func(lib, suboptions, subargs)
plugins.send('cli_exit', lib=lib)
def main(args=None):
"""Run the main command-line interface for beets. Includes top-level
exception handlers that print friendly error messages.
"""
try:
_raw_main(args)
except UserError as exc:
message = exc.args[0] if exc.args else None
log.error(u'error: {0}'.format(message))
sys.exit(1)
except util.HumanReadableException as exc:
exc.log(log)
sys.exit(1)
except library.FileOperationError as exc:
# These errors have reasonable human-readable descriptions, but
# we still want to log their tracebacks for debugging.
log.debug(traceback.format_exc())
log.error(exc)
sys.exit(1)
except confit.ConfigError as exc:
log.error(u'configuration error: {0}'.format(exc))
sys.exit(1)
except IOError as exc:
if exc.errno == errno.EPIPE:
# "Broken pipe". End silently.
pass
else:
raise
except KeyboardInterrupt:
# Silently ignore ^C except in verbose mode.
log.debug(traceback.format_exc())
lib/beets/ui/commands.py
-1646
View File
File diff suppressed because it is too large Load Diff
lib/beets/ui/completion_base.sh
-162
View File
@@ -1,162 +0,0 @@
# This file is part of beets.
# Copyright (c) 2014, Thomas Scholtes.
#
# Permission is hereby granted, free of charge, to any person obtaining
# a copy of this software and associated documentation files (the
# "Software"), to deal in the Software without restriction, including
# without limitation the rights to use, copy, modify, merge, publish,
# distribute, sublicense, and/or sell copies of the Software, and to
# permit persons to whom the Software is furnished to do so, subject to
# the following conditions:
#
# The above copyright notice and this permission notice shall be
# included in all copies or substantial portions of the Software.
# Completion for the `beet` command
# =================================
#
# Load this script to complete beets subcommands, options, and
# queries.
#
# If a beets command is found on the command line it completes filenames and
# the subcommand's options. Otherwise it will complete global options and
# subcommands. If the previous option on the command line expects an argument,
# it also completes filenames or directories. Options are only
# completed if '-' has already been typed on the command line.
#
# Note that completion of plugin commands only works for those plugins
# that were enabled when running `beet completion`. It does not check
# plugins dynamically
#
# Currently, only Bash 3.2 and newer is supported and the
# `bash-completion` package is requied.
#
# TODO
# ----
#
# * There are some issues with arguments that are quoted on the command line.
#
# * Complete arguments for the `--format` option by expanding field variables.
#
# beet ls -f "$tit[TAB]
# beet ls -f "$title
#
# * Support long options with `=`, e.g. `--config=file`. Debian's bash
# completion package can handle this.
#
# Determines the beets subcommand and dispatches the completion
# accordingly.
_beet_dispatch() {
local cur prev cmd=
COMPREPLY=()
_get_comp_words_by_ref -n : cur prev
# Look for the beets subcommand
local arg
for (( i=1; i < COMP_CWORD; i++ )); do
arg="${COMP_WORDS[i]}"
if _list_include_item "${opts___global}" $arg; then
((i++))
elif [[ "$arg" != -* ]]; then
cmd="$arg"
break
fi
done
# Replace command shortcuts
if [[ -n $cmd ]] && _list_include_item "$aliases" "$cmd"; then
eval "cmd=\$alias__$cmd"
fi
case $cmd in
help)
COMPREPLY+=( $(compgen -W "$commands" -- $cur) )
;;
list|remove|move|update|write|stats)
_beet_complete_query
;;
"")
_beet_complete_global
;;
*)
_beet_complete
;;
esac
}
# Adds option and file completion to COMPREPLY for the subcommand $cmd
_beet_complete() {
if [[ $cur == -* ]]; then
local opts flags completions
eval "opts=\$opts__$cmd"
eval "flags=\$flags__$cmd"
completions="${flags___common} ${opts} ${flags}"
COMPREPLY+=( $(compgen -W "$completions" -- $cur) )
else
_filedir
fi
}
# Add global options and subcommands to the completion
_beet_complete_global() {
case $prev in
-h|--help)
# Complete commands
COMPREPLY+=( $(compgen -W "$commands" -- $cur) )
return
;;
-l|--library|-c|--config)
# Filename completion
_filedir
return
;;
-d|--directory)
# Directory completion
_filedir -d
return
;;
esac
if [[ $cur == -* ]]; then
local completions="$opts___global $flags___global"
COMPREPLY+=( $(compgen -W "$completions" -- $cur) )
elif [[ -n $cur ]] && _list_include_item "$aliases" "$cur"; then
local cmd
eval "cmd=\$alias__$cur"
COMPREPLY+=( "$cmd" )
else
COMPREPLY+=( $(compgen -W "$commands" -- $cur) )
fi
}
_beet_complete_query() {
local opts
eval "opts=\$opts__$cmd"
if [[ $cur == -* ]] || _list_include_item "$opts" "$prev"; then
_beet_complete
elif [[ $cur != \'* && $cur != \"* &&
$cur != *:* ]]; then
# Do not complete quoted queries or those who already have a field
# set.
compopt -o nospace
COMPREPLY+=( $(compgen -S : -W "$fields" -- $cur) )
return 0
fi
}
# Returns true if the space separated list $1 includes $2
_list_include_item() {
[[ " $1 " == *[[:space:]]$2[[:space:]]* ]]
}
# This is where beets dynamically adds the _beet function. This
# function sets the variables $flags, $opts, $commands, and $aliases.
complete -o filenames -F _beet beet
lib/beets/util/__init__.py
-686
View File
@@ -1,686 +0,0 @@
# This file is part of beets.
# Copyright 2013, Adrian Sampson.
#
# Permission is hereby granted, free of charge, to any person obtaining
# a copy of this software and associated documentation files (the
# "Software"), to deal in the Software without restriction, including
# without limitation the rights to use, copy, modify, merge, publish,
# distribute, sublicense, and/or sell copies of the Software, and to
# permit persons to whom the Software is furnished to do so, subject to
# the following conditions:
#
# The above copyright notice and this permission notice shall be
# included in all copies or substantial portions of the Software.
"""Miscellaneous utility functions."""
from __future__ import division
import os
import sys
import re
import shutil
import fnmatch
from collections import defaultdict
import traceback
import subprocess
import platform
MAX_FILENAME_LENGTH = 200
WINDOWS_MAGIC_PREFIX = u'\\\\?\\'
class HumanReadableException(Exception):
"""An Exception that can include a human-readable error message to
be logged without a traceback. Can preserve a traceback for
debugging purposes as well.
Has at least two fields: `reason`, the underlying exception or a
string describing the problem; and `verb`, the action being
performed during the error.
If `tb` is provided, it is a string containing a traceback for the
associated exception. (Note that this is not necessary in Python 3.x
and should be removed when we make the transition.)
"""
error_kind = 'Error' # Human-readable description of error type.
def __init__(self, reason, verb, tb=None):
self.reason = reason
self.verb = verb
self.tb = tb
super(HumanReadableException, self).__init__(self.get_message())
def _gerund(self):
"""Generate a (likely) gerund form of the English verb.
"""
if ' ' in self.verb:
return self.verb
gerund = self.verb[:-1] if self.verb.endswith('e') else self.verb
gerund += 'ing'
return gerund
def _reasonstr(self):
"""Get the reason as a string."""
if isinstance(self.reason, unicode):
return self.reason
elif isinstance(self.reason, basestring): # Byte string.
return self.reason.decode('utf8', 'ignore')
elif hasattr(self.reason, 'strerror'): # i.e., EnvironmentError
return self.reason.strerror
else:
return u'"{0}"'.format(unicode(self.reason))
def get_message(self):
"""Create the human-readable description of the error, sans
introduction.
"""
raise NotImplementedError
def log(self, logger):
"""Log to the provided `logger` a human-readable message as an
error and a verbose traceback as a debug message.
"""
if self.tb:
logger.debug(self.tb)
logger.error(u'{0}: {1}'.format(self.error_kind, self.args[0]))
class FilesystemError(HumanReadableException):
"""An error that occurred while performing a filesystem manipulation
via a function in this module. The `paths` field is a sequence of
pathnames involved in the operation.
"""
def __init__(self, reason, verb, paths, tb=None):
self.paths = paths
super(FilesystemError, self).__init__(reason, verb, tb)
def get_message(self):
# Use a nicer English phrasing for some specific verbs.
if self.verb in ('move', 'copy', 'rename'):
clause = u'while {0} {1} to {2}'.format(
self._gerund(),
displayable_path(self.paths[0]),
displayable_path(self.paths[1])
)
elif self.verb in ('delete', 'write', 'create', 'read'):
clause = u'while {0} {1}'.format(
self._gerund(),
displayable_path(self.paths[0])
)
else:
clause = u'during {0} of paths {1}'.format(
self.verb, u', '.join(displayable_path(p) for p in self.paths)
)
return u'{0} {1}'.format(self._reasonstr(), clause)
def normpath(path):
"""Provide the canonical form of the path suitable for storing in
the database.
"""
path = syspath(path, prefix=False)
path = os.path.normpath(os.path.abspath(os.path.expanduser(path)))
return bytestring_path(path)
def ancestry(path):
"""Return a list consisting of path's parent directory, its
grandparent, and so on. For instance:
>>> ancestry('/a/b/c')
['/', '/a', '/a/b']
The argument should *not* be the result of a call to `syspath`.
"""
out = []
last_path = None
while path:
path = os.path.dirname(path)
if path == last_path:
break
last_path = path
if path:
# don't yield ''
out.insert(0, path)
return out
def sorted_walk(path, ignore=(), logger=None):
"""Like `os.walk`, but yields things in case-insensitive sorted,
breadth-first order. Directory and file names matching any glob
pattern in `ignore` are skipped. If `logger` is provided, then
warning messages are logged there when a directory cannot be listed.
"""
# Make sure the path isn't a Unicode string.
path = bytestring_path(path)
# Get all the directories and files at this level.
try:
contents = os.listdir(syspath(path))
except OSError as exc:
if logger:
logger.warn(u'could not list directory {0}: {1}'.format(
displayable_path(path), exc.strerror
))
return
dirs = []
files = []
for base in contents:
base = bytestring_path(base)
# Skip ignored filenames.
skip = False
for pat in ignore:
if fnmatch.fnmatch(base, pat):
skip = True
break
if skip:
continue
# Add to output as either a file or a directory.
cur = os.path.join(path, base)
if os.path.isdir(syspath(cur)):
dirs.append(base)
else:
files.append(base)
# Sort lists (case-insensitive) and yield the current level.
dirs.sort(key=bytes.lower)
files.sort(key=bytes.lower)
yield (path, dirs, files)
# Recurse into directories.
for base in dirs:
cur = os.path.join(path, base)
# yield from sorted_walk(...)
for res in sorted_walk(cur, ignore, logger):
yield res
def mkdirall(path):
"""Make all the enclosing directories of path (like mkdir -p on the
parent).
"""
for ancestor in ancestry(path):
if not os.path.isdir(syspath(ancestor)):
try:
os.mkdir(syspath(ancestor))
except (OSError, IOError) as exc:
raise FilesystemError(exc, 'create', (ancestor,),
traceback.format_exc())
def fnmatch_all(names, patterns):
"""Determine whether all strings in `names` match at least one of
the `patterns`, which should be shell glob expressions.
"""
for name in names:
matches = False
for pattern in patterns:
matches = fnmatch.fnmatch(name, pattern)
if matches:
break
if not matches:
return False
return True
def prune_dirs(path, root=None, clutter=('.DS_Store', 'Thumbs.db')):
"""If path is an empty directory, then remove it. Recursively remove
path's ancestry up to root (which is never removed) where there are
empty directories. If path is not contained in root, then nothing is
removed. Glob patterns in clutter are ignored when determining
emptiness. If root is not provided, then only path may be removed
(i.e., no recursive removal).
"""
path = normpath(path)
if root is not None:
root = normpath(root)
ancestors = ancestry(path)
if root is None:
# Only remove the top directory.
ancestors = []
elif root in ancestors:
# Only remove directories below the root.
ancestors = ancestors[ancestors.index(root) + 1:]
else:
# Remove nothing.
return
# Traverse upward from path.
ancestors.append(path)
ancestors.reverse()
for directory in ancestors:
directory = syspath(directory)
if not os.path.exists(directory):
# Directory gone already.
continue
if fnmatch_all(os.listdir(directory), clutter):
# Directory contains only clutter (or nothing).
try:
shutil.rmtree(directory)
except OSError:
break
else:
break
def components(path):
"""Return a list of the path components in path. For instance:
>>> components('/a/b/c')
['a', 'b', 'c']
The argument should *not* be the result of a call to `syspath`.
"""
comps = []
ances = ancestry(path)
for anc in ances:
comp = os.path.basename(anc)
if comp:
comps.append(comp)
else: # root
comps.append(anc)
last = os.path.basename(path)
if last:
comps.append(last)
return comps
def _fsencoding():
"""Get the system's filesystem encoding. On Windows, this is always
UTF-8 (not MBCS).
"""
encoding = sys.getfilesystemencoding() or sys.getdefaultencoding()
if encoding == 'mbcs':
# On Windows, a broken encoding known to Python as "MBCS" is
# used for the filesystem. However, we only use the Unicode API
# for Windows paths, so the encoding is actually immaterial so
# we can avoid dealing with this nastiness. We arbitrarily
# choose UTF-8.
encoding = 'utf8'
return encoding
def bytestring_path(path):
"""Given a path, which is either a str or a unicode, returns a str
path (ensuring that we never deal with Unicode pathnames).
"""
# Pass through bytestrings.
if isinstance(path, str):
return path
# On Windows, remove the magic prefix added by `syspath`. This makes
# ``bytestring_path(syspath(X)) == X``, i.e., we can safely
# round-trip through `syspath`.
if os.path.__name__ == 'ntpath' and path.startswith(WINDOWS_MAGIC_PREFIX):
path = path[len(WINDOWS_MAGIC_PREFIX):]
# Try to encode with default encodings, but fall back to UTF8.
try:
return path.encode(_fsencoding())
except (UnicodeError, LookupError):
return path.encode('utf8')
def displayable_path(path, separator=u'; '):
"""Attempts to decode a bytestring path to a unicode object for the
purpose of displaying it to the user. If the `path` argument is a
list or a tuple, the elements are joined with `separator`.
"""
if isinstance(path, (list, tuple)):
return separator.join(displayable_path(p) for p in path)
elif isinstance(path, unicode):
return path
elif not isinstance(path, str):
# A non-string object: just get its unicode representation.
return unicode(path)
try:
return path.decode(_fsencoding(), 'ignore')
except (UnicodeError, LookupError):
return path.decode('utf8', 'ignore')
def syspath(path, prefix=True):
"""Convert a path for use by the operating system. In particular,
paths on Windows must receive a magic prefix and must be converted
to Unicode before they are sent to the OS. To disable the magic
prefix on Windows, set `prefix` to False---but only do this if you
*really* know what you're doing.
"""
# Don't do anything if we're not on windows
if os.path.__name__ != 'ntpath':
return path
if not isinstance(path, unicode):
# Beets currently represents Windows paths internally with UTF-8
# arbitrarily. But earlier versions used MBCS because it is
# reported as the FS encoding by Windows. Try both.
try:
path = path.decode('utf8')
except UnicodeError:
# The encoding should always be MBCS, Windows' broken
# Unicode representation.
encoding = sys.getfilesystemencoding() or sys.getdefaultencoding()
path = path.decode(encoding, 'replace')
# Add the magic prefix if it isn't already there.
# http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247.aspx
if prefix and not path.startswith(WINDOWS_MAGIC_PREFIX):
if path.startswith(u'\\\\'):
# UNC path. Final path should look like \\?\UNC\...
path = u'UNC' + path[1:]
path = WINDOWS_MAGIC_PREFIX + path
return path
def samefile(p1, p2):
"""Safer equality for paths."""
return shutil._samefile(syspath(p1), syspath(p2))
def remove(path, soft=True):
"""Remove the file. If `soft`, then no error will be raised if the
file does not exist.
"""
path = syspath(path)
if soft and not os.path.exists(path):
return
try:
os.remove(path)
except (OSError, IOError) as exc:
raise FilesystemError(exc, 'delete', (path,), traceback.format_exc())
def copy(path, dest, replace=False):
"""Copy a plain file. Permissions are not copied. If `dest` already
exists, raises a FilesystemError unless `replace` is True. Has no
effect if `path` is the same as `dest`. Paths are translated to
system paths before the syscall.
"""
if samefile(path, dest):
return
path = syspath(path)
dest = syspath(dest)
if not replace and os.path.exists(dest):
raise FilesystemError('file exists', 'copy', (path, dest))
try:
shutil.copyfile(path, dest)
except (OSError, IOError) as exc:
raise FilesystemError(exc, 'copy', (path, dest),
traceback.format_exc())
def move(path, dest, replace=False):
"""Rename a file. `dest` may not be a directory. If `dest` already
exists, raises an OSError unless `replace` is True. Has no effect if
`path` is the same as `dest`. If the paths are on different
filesystems (or the rename otherwise fails), a copy is attempted
instead, in which case metadata will *not* be preserved. Paths are
translated to system paths.
"""
if samefile(path, dest):
return
path = syspath(path)
dest = syspath(dest)
if os.path.exists(dest) and not replace:
raise FilesystemError('file exists', 'rename', (path, dest),
traceback.format_exc())
# First, try renaming the file.
try:
os.rename(path, dest)
except OSError:
# Otherwise, copy and delete the original.
try:
shutil.copyfile(path, dest)
os.remove(path)
except (OSError, IOError) as exc:
raise FilesystemError(exc, 'move', (path, dest),
traceback.format_exc())
def link(path, dest, replace=False):
"""Create a symbolic link from path to `dest`. Raises an OSError if
`dest` already exists, unless `replace` is True. Does nothing if
`path` == `dest`."""
if (samefile(path, dest)):
return
path = syspath(path)
dest = syspath(dest)
if os.path.exists(dest) and not replace:
raise FilesystemError('file exists', 'rename', (path, dest),
traceback.format_exc())
try:
os.symlink(path, dest)
except OSError:
raise FilesystemError('Operating system does not support symbolic '
'links.', 'link', (path, dest),
traceback.format_exc())
def unique_path(path):
"""Returns a version of ``path`` that does not exist on the
filesystem. Specifically, if ``path` itself already exists, then
something unique is appended to the path.
"""
if not os.path.exists(syspath(path)):
return path
base, ext = os.path.splitext(path)
match = re.search(r'\.(\d)+$', base)
if match:
num = int(match.group(1))
base = base[:match.start()]
else:
num = 0
while True:
num += 1
new_path = '%s.%i%s' % (base, num, ext)
if not os.path.exists(new_path):
return new_path
# Note: The Windows "reserved characters" are, of course, allowed on
# Unix. They are forbidden here because they cause problems on Samba
# shares, which are sufficiently common as to cause frequent problems.
# http://msdn.microsoft.com/en-us/library/windows/desktop/aa365247.aspx
CHAR_REPLACE = [
(re.compile(ur'[\\/]'), u'_'), # / and \ -- forbidden everywhere.
(re.compile(ur'^\.'), u'_'), # Leading dot (hidden files on Unix).
(re.compile(ur'[\x00-\x1f]'), u''), # Control characters.
(re.compile(ur'[<>:"\?\*\|]'), u'_'), # Windows "reserved characters".
(re.compile(ur'\.$'), u'_'), # Trailing dots.
(re.compile(ur'\s+$'), u''), # Trailing whitespace.
]
def sanitize_path(path, replacements=None):
"""Takes a path (as a Unicode string) and makes sure that it is
legal. Returns a new path. Only works with fragments; won't work
reliably on Windows when a path begins with a drive letter. Path
separators (including altsep!) should already be cleaned from the
path components. If replacements is specified, it is used *instead*
of the default set of replacements; it must be a list of (compiled
regex, replacement string) pairs.
"""
replacements = replacements or CHAR_REPLACE
comps = components(path)
if not comps:
return ''
for i, comp in enumerate(comps):
for regex, repl in replacements:
comp = regex.sub(repl, comp)
comps[i] = comp
return os.path.join(*comps)
def truncate_path(path, length=MAX_FILENAME_LENGTH):
"""Given a bytestring path or a Unicode path fragment, truncate the
components to a legal length. In the last component, the extension
is preserved.
"""
comps = components(path)
out = [c[:length] for c in comps]
base, ext = os.path.splitext(comps[-1])
if ext:
# Last component has an extension.
base = base[:length - len(ext)]
out[-1] = base + ext
return os.path.join(*out)
def str2bool(value):
"""Returns a boolean reflecting a human-entered string."""
if value.lower() in ('yes', '1', 'true', 't', 'y'):
return True
else:
return False
def as_string(value):
"""Convert a value to a Unicode object for matching with a query.
None becomes the empty string. Bytestrings are silently decoded.
"""
if value is None:
return u''
elif isinstance(value, buffer):
return str(value).decode('utf8', 'ignore')
elif isinstance(value, str):
return value.decode('utf8', 'ignore')
else:
return unicode(value)
def levenshtein(s1, s2):
"""A nice DP edit distance implementation from Wikibooks:
http://en.wikibooks.org/wiki/Algorithm_implementation/Strings/
Levenshtein_distance#Python
"""
if len(s1) < len(s2):
return levenshtein(s2, s1)
if not s1:
return len(s2)
previous_row = xrange(len(s2) + 1)
for i, c1 in enumerate(s1):
current_row = [i + 1]
for j, c2 in enumerate(s2):
insertions = previous_row[j + 1] + 1
deletions = current_row[j] + 1
substitutions = previous_row[j] + (c1 != c2)
current_row.append(min(insertions, deletions, substitutions))
previous_row = current_row
return previous_row[-1]
def plurality(objs):
"""Given a sequence of comparable objects, returns the object that
is most common in the set and the frequency of that object. The
sequence must contain at least one object.
"""
# Calculate frequencies.
freqs = defaultdict(int)
for obj in objs:
freqs[obj] += 1
if not freqs:
raise ValueError('sequence must be non-empty')
# Find object with maximum frequency.
max_freq = 0
res = None
for obj, freq in freqs.items():
if freq > max_freq:
max_freq = freq
res = obj
return res, max_freq
def cpu_count():
"""Return the number of hardware thread contexts (cores or SMT
threads) in the system.
"""
# Adapted from the soundconverter project:
# https://github.com/kassoulet/soundconverter
if sys.platform == 'win32':
try:
num = int(os.environ['NUMBER_OF_PROCESSORS'])
except (ValueError, KeyError):
num = 0
elif sys.platform == 'darwin':
try:
num = int(command_output(['sysctl', '-n', 'hw.ncpu']))
except ValueError:
num = 0
else:
try:
num = os.sysconf('SC_NPROCESSORS_ONLN')
except (ValueError, OSError, AttributeError):
num = 0
if num >= 1:
return num
else:
return 1
def command_output(cmd, shell=False):
"""Runs the command and returns its output after it has exited.
``cmd`` is a list of arguments starting with the command names. If
``shell`` is true, ``cmd`` is assumed to be a string and passed to a
shell to execute.
If the process exits with a non-zero return code
``subprocess.CalledProcessError`` is raised. May also raise
``OSError``.
This replaces `subprocess.check_output`, which isn't available in
Python 2.6 and which can have problems if lots of output is sent to
stderr.
"""
proc = subprocess.Popen(
cmd,
stdout=subprocess.PIPE,
stderr=subprocess.PIPE,
close_fds=platform.system() != 'Windows',
shell=shell
)
stdout, stderr = proc.communicate()
if proc.returncode:
raise subprocess.CalledProcessError(
returncode=proc.returncode,
cmd=' '.join(cmd),
)
return stdout
def max_filename_length(path, limit=MAX_FILENAME_LENGTH):
"""Attempt to determine the maximum filename length for the
filesystem containing `path`. If the value is greater than `limit`,
then `limit` is used instead (to prevent errors when a filesystem
misreports its capacity). If it cannot be determined (e.g., on
Windows), return `limit`.
"""
if hasattr(os, 'statvfs'):
try:
res = os.statvfs(path)
except OSError:
return limit
return min(res[9], limit)
else:
return limit
lib/beets/util/artresizer.py
-211
View File
@@ -1,211 +0,0 @@
# This file is part of beets.
# Copyright 2014, Fabrice Laporte
#
# Permission is hereby granted, free of charge, to any person obtaining
# a copy of this software and associated documentation files (the
# "Software"), to deal in the Software without restriction, including
# without limitation the rights to use, copy, modify, merge, publish,
# distribute, sublicense, and/or sell copies of the Software, and to
# permit persons to whom the Software is furnished to do so, subject to
# the following conditions:
#
# The above copyright notice and this permission notice shall be
# included in all copies or substantial portions of the Software.
"""Abstraction layer to resize images using PIL, ImageMagick, or a
public resizing proxy if neither is available.
"""
import urllib
import subprocess
import os
import re
from tempfile import NamedTemporaryFile
import logging
from beets import util
# Resizing methods
PIL = 1
IMAGEMAGICK = 2
WEBPROXY = 3
PROXY_URL = 'http://images.weserv.nl/'
log = logging.getLogger('beets')
def resize_url(url, maxwidth):
"""Return a proxied image URL that resizes the original image to
maxwidth (preserving aspect ratio).
"""
return '{0}?{1}'.format(PROXY_URL, urllib.urlencode({
'url': url.replace('http://', ''),
'w': str(maxwidth),
}))
def temp_file_for(path):
"""Return an unused filename with the same extension as the
specified path.
"""
ext = os.path.splitext(path)[1]
with NamedTemporaryFile(suffix=ext, delete=False) as f:
return f.name
def pil_resize(maxwidth, path_in, path_out=None):
"""Resize using Python Imaging Library (PIL). Return the output path
of resized image.
"""
path_out = path_out or temp_file_for(path_in)
from PIL import Image
log.debug(u'artresizer: PIL resizing {0} to {1}'.format(
util.displayable_path(path_in), util.displayable_path(path_out)
))
try:
im = Image.open(util.syspath(path_in))
size = maxwidth, maxwidth
im.thumbnail(size, Image.ANTIALIAS)
im.save(path_out)
return path_out
except IOError:
log.error(u"PIL cannot create thumbnail for '{0}'".format(
util.displayable_path(path_in)
))
return path_in
def im_resize(maxwidth, path_in, path_out=None):
"""Resize using ImageMagick's ``convert`` tool.
Return the output path of resized image.
"""
path_out = path_out or temp_file_for(path_in)
log.debug(u'artresizer: ImageMagick resizing {0} to {1}'.format(
util.displayable_path(path_in), util.displayable_path(path_out)
))
# "-resize widthxheight>" shrinks images with dimension(s) larger
# than the corresponding width and/or height dimension(s). The >
# "only shrink" flag is prefixed by ^ escape char for Windows
# compatibility.
try:
util.command_output([
'convert', util.syspath(path_in),
'-resize', '{0}x^>'.format(maxwidth), path_out
])
except subprocess.CalledProcessError:
log.warn(u'artresizer: IM convert failed for {0}'.format(
util.displayable_path(path_in)
))
return path_in
return path_out
BACKEND_FUNCS = {
PIL: pil_resize,
IMAGEMAGICK: im_resize,
}
class Shareable(type):
"""A pseudo-singleton metaclass that allows both shared and
non-shared instances. The ``MyClass.shared`` property holds a
lazily-created shared instance of ``MyClass`` while calling
``MyClass()`` to construct a new object works as usual.
"""
def __init__(cls, name, bases, dict):
super(Shareable, cls).__init__(name, bases, dict)
cls._instance = None
@property
def shared(cls):
if cls._instance is None:
cls._instance = cls()
return cls._instance
class ArtResizer(object):
"""A singleton class that performs image resizes.
"""
__metaclass__ = Shareable
def __init__(self, method=None):
"""Create a resizer object for the given method or, if none is
specified, with an inferred method.
"""
self.method = self._check_method(method)
log.debug(u"artresizer: method is {0}".format(self.method))
self.can_compare = self._can_compare()
def resize(self, maxwidth, path_in, path_out=None):
"""Manipulate an image file according to the method, returning a
new path. For PIL or IMAGEMAGIC methods, resizes the image to a
temporary file. For WEBPROXY, returns `path_in` unmodified.
"""
if self.local:
func = BACKEND_FUNCS[self.method[0]]
return func(maxwidth, path_in, path_out)
else:
return path_in
def proxy_url(self, maxwidth, url):
"""Modifies an image URL according the method, returning a new
URL. For WEBPROXY, a URL on the proxy server is returned.
Otherwise, the URL is returned unmodified.
"""
if self.local:
return url
else:
return resize_url(url, maxwidth)
@property
def local(self):
"""A boolean indicating whether the resizing method is performed
locally (i.e., PIL or ImageMagick).
"""
return self.method[0] in BACKEND_FUNCS
def _can_compare(self):
"""A boolean indicating whether image comparison is available"""
return self.method[0] == IMAGEMAGICK and self.method[1] > (6, 8, 7)
@staticmethod
def _check_method(method=None):
"""A tuple indicating whether current method is available and its
version. If no method is given, it returns a supported one.
"""
# Guess available method
if not method:
for m in [IMAGEMAGICK, PIL]:
_, version = ArtResizer._check_method(m)
if version:
return (m, version)
return (WEBPROXY, (0))
if method == IMAGEMAGICK:
# Try invoking ImageMagick's "convert".
try:
out = util.command_output(['identify', '--version'])
if 'imagemagick' in out.lower():
pattern = r".+ (\d+)\.(\d+)\.(\d+).*"
match = re.search(pattern, out)
if match:
return (IMAGEMAGICK,
(int(match.group(1)),
int(match.group(2)),
int(match.group(3))))
return (IMAGEMAGICK, (0))
except (subprocess.CalledProcessError, OSError):
return (IMAGEMAGICK, None)
if method == PIL:
# Try importing PIL.
try:
__import__('PIL', fromlist=['Image'])
return (PIL, (0))
except ImportError:
return (PIL, None)
lib/beets/util/bluelet.py
-647
View File
@@ -1,647 +0,0 @@
"""Extremely simple pure-Python implementation of coroutine-style
asynchronous socket I/O. Inspired by, but inferior to, Eventlet.
Bluelet can also be thought of as a less-terrible replacement for
asyncore.
Bluelet: easy concurrency without all the messy parallelism.
"""
import socket
import select
import sys
import types
import errno
import traceback
import time
import collections
# A little bit of "six" (Python 2/3 compatibility): cope with PEP 3109 syntax
# changes.
PY3 = sys.version_info[0] == 3
if PY3:
def _reraise(typ, exc, tb):
raise exc.with_traceback(tb)
else:
exec("""
def _reraise(typ, exc, tb):
raise typ, exc, tb
""")
# Basic events used for thread scheduling.
class Event(object):
"""Just a base class identifying Bluelet events. An event is an
object yielded from a Bluelet thread coroutine to suspend operation
and communicate with the scheduler.
"""
pass
class WaitableEvent(Event):
"""A waitable event is one encapsulating an action that can be
waited for using a select() call. That is, it's an event with an
associated file descriptor.
"""
def waitables(self):
"""Return "waitable" objects to pass to select(). Should return
three iterables for input readiness, output readiness, and
exceptional conditions (i.e., the three lists passed to
select()).
"""
return (), (), ()
def fire(self):
"""Called when an associated file descriptor becomes ready
(i.e., is returned from a select() call).
"""
pass
class ValueEvent(Event):
"""An event that does nothing but return a fixed value."""
def __init__(self, value):
self.value = value
class ExceptionEvent(Event):
"""Raise an exception at the yield point. Used internally."""
def __init__(self, exc_info):
self.exc_info = exc_info
class SpawnEvent(Event):
"""Add a new coroutine thread to the scheduler."""
def __init__(self, coro):
self.spawned = coro
class JoinEvent(Event):
"""Suspend the thread until the specified child thread has
completed.
"""
def __init__(self, child):
self.child = child
class KillEvent(Event):
"""Unschedule a child thread."""
def __init__(self, child):
self.child = child
class DelegationEvent(Event):
"""Suspend execution of the current thread, start a new thread and,
once the child thread finished, return control to the parent
thread.
"""
def __init__(self, coro):
self.spawned = coro
class ReturnEvent(Event):
"""Return a value the current thread's delegator at the point of
delegation. Ends the current (delegate) thread.
"""
def __init__(self, value):
self.value = value
class SleepEvent(WaitableEvent):
"""Suspend the thread for a given duration.
"""
def __init__(self, duration):
self.wakeup_time = time.time() + duration
def time_left(self):
return max(self.wakeup_time - time.time(), 0.0)
class ReadEvent(WaitableEvent):
"""Reads from a file-like object."""
def __init__(self, fd, bufsize):
self.fd = fd
self.bufsize = bufsize
def waitables(self):
return (self.fd,), (), ()
def fire(self):
return self.fd.read(self.bufsize)
class WriteEvent(WaitableEvent):
"""Writes to a file-like object."""
def __init__(self, fd, data):
self.fd = fd
self.data = data
def waitable(self):
return (), (self.fd,), ()
def fire(self):
self.fd.write(self.data)
# Core logic for executing and scheduling threads.
def _event_select(events):
"""Perform a select() over all the Events provided, returning the
ones ready to be fired. Only WaitableEvents (including SleepEvents)
matter here; all other events are ignored (and thus postponed).
"""
# Gather waitables and wakeup times.
waitable_to_event = {}
rlist, wlist, xlist = [], [], []
earliest_wakeup = None
for event in events:
if isinstance(event, SleepEvent):
if not earliest_wakeup:
earliest_wakeup = event.wakeup_time
else:
earliest_wakeup = min(earliest_wakeup, event.wakeup_time)
elif isinstance(event, WaitableEvent):
r, w, x = event.waitables()
rlist += r
wlist += w
xlist += x
for waitable in r:
waitable_to_event[('r', waitable)] = event
for waitable in w:
waitable_to_event[('w', waitable)] = event
for waitable in x:
waitable_to_event[('x', waitable)] = event
# If we have a any sleeping threads, determine how long to sleep.
if earliest_wakeup:
timeout = max(earliest_wakeup - time.time(), 0.0)
else:
timeout = None
# Perform select() if we have any waitables.
if rlist or wlist or xlist:
rready, wready, xready = select.select(rlist, wlist, xlist, timeout)
else:
rready, wready, xready = (), (), ()
if timeout:
time.sleep(timeout)
# Gather ready events corresponding to the ready waitables.
ready_events = set()
for ready in rready:
ready_events.add(waitable_to_event[('r', ready)])
for ready in wready:
ready_events.add(waitable_to_event[('w', ready)])
for ready in xready:
ready_events.add(waitable_to_event[('x', ready)])
# Gather any finished sleeps.
for event in events:
if isinstance(event, SleepEvent) and event.time_left() == 0.0:
ready_events.add(event)
return ready_events
class ThreadException(Exception):
def __init__(self, coro, exc_info):
self.coro = coro
self.exc_info = exc_info
def reraise(self):
_reraise(self.exc_info[0], self.exc_info[1], self.exc_info[2])
SUSPENDED = Event() # Special sentinel placeholder for suspended threads.
class Delegated(Event):
"""Placeholder indicating that a thread has delegated execution to a
different thread.
"""
def __init__(self, child):
self.child = child
def run(root_coro):
"""Schedules a coroutine, running it to completion. This
encapsulates the Bluelet scheduler, which the root coroutine can
add to by spawning new coroutines.
"""
# The "threads" dictionary keeps track of all the currently-
# executing and suspended coroutines. It maps coroutines to their
# currently "blocking" event. The event value may be SUSPENDED if
# the coroutine is waiting on some other condition: namely, a
# delegated coroutine or a joined coroutine. In this case, the
# coroutine should *also* appear as a value in one of the below
# dictionaries `delegators` or `joiners`.
threads = {root_coro: ValueEvent(None)}
# Maps child coroutines to delegating parents.
delegators = {}
# Maps child coroutines to joining (exit-waiting) parents.
joiners = collections.defaultdict(list)
def complete_thread(coro, return_value):
"""Remove a coroutine from the scheduling pool, awaking
delegators and joiners as necessary and returning the specified
value to any delegating parent.
"""
del threads[coro]
# Resume delegator.
if coro in delegators:
threads[delegators[coro]] = ValueEvent(return_value)
del delegators[coro]
# Resume joiners.
if coro in joiners:
for parent in joiners[coro]:
threads[parent] = ValueEvent(None)
del joiners[coro]
def advance_thread(coro, value, is_exc=False):
"""After an event is fired, run a given coroutine associated with
it in the threads dict until it yields again. If the coroutine
exits, then the thread is removed from the pool. If the coroutine
raises an exception, it is reraised in a ThreadException. If
is_exc is True, then the value must be an exc_info tuple and the
exception is thrown into the coroutine.
"""
try:
if is_exc:
next_event = coro.throw(*value)
else:
next_event = coro.send(value)
except StopIteration:
# Thread is done.
complete_thread(coro, None)
except:
# Thread raised some other exception.
del threads[coro]
raise ThreadException(coro, sys.exc_info())
else:
if isinstance(next_event, types.GeneratorType):
# Automatically invoke sub-coroutines. (Shorthand for
# explicit bluelet.call().)
next_event = DelegationEvent(next_event)
threads[coro] = next_event
def kill_thread(coro):
"""Unschedule this thread and its (recursive) delegates.
"""
# Collect all coroutines in the delegation stack.
coros = [coro]
while isinstance(threads[coro], Delegated):
coro = threads[coro].child
coros.append(coro)
# Complete each coroutine from the top to the bottom of the
# stack.
for coro in reversed(coros):
complete_thread(coro, None)
# Continue advancing threads until root thread exits.
exit_te = None
while threads:
try:
# Look for events that can be run immediately. Continue
# running immediate events until nothing is ready.
while True:
have_ready = False
for coro, event in list(threads.items()):
if isinstance(event, SpawnEvent):
threads[event.spawned] = ValueEvent(None) # Spawn.
advance_thread(coro, None)
have_ready = True
elif isinstance(event, ValueEvent):
advance_thread(coro, event.value)
have_ready = True
elif isinstance(event, ExceptionEvent):
advance_thread(coro, event.exc_info, True)
have_ready = True
elif isinstance(event, DelegationEvent):
threads[coro] = Delegated(event.spawned) # Suspend.
threads[event.spawned] = ValueEvent(None) # Spawn.
delegators[event.spawned] = coro
have_ready = True
elif isinstance(event, ReturnEvent):
# Thread is done.
complete_thread(coro, event.value)
have_ready = True
elif isinstance(event, JoinEvent):
threads[coro] = SUSPENDED # Suspend.
joiners[event.child].append(coro)
have_ready = True
elif isinstance(event, KillEvent):
threads[coro] = ValueEvent(None)
kill_thread(event.child)
have_ready = True
# Only start the select when nothing else is ready.
if not have_ready:
break
# Wait and fire.
event2coro = dict((v, k) for k, v in threads.items())
for event in _event_select(threads.values()):
# Run the IO operation, but catch socket errors.
try:
value = event.fire()
except socket.error as exc:
if isinstance(exc.args, tuple) and \
exc.args[0] == errno.EPIPE:
# Broken pipe. Remote host disconnected.
pass
else:
traceback.print_exc()
# Abort the coroutine.
threads[event2coro[event]] = ReturnEvent(None)
else:
advance_thread(event2coro[event], value)
except ThreadException as te:
# Exception raised from inside a thread.
event = ExceptionEvent(te.exc_info)
if te.coro in delegators:
# The thread is a delegate. Raise exception in its
# delegator.
threads[delegators[te.coro]] = event
del delegators[te.coro]
else:
# The thread is root-level. Raise in client code.
exit_te = te
break
except:
# For instance, KeyboardInterrupt during select(). Raise
# into root thread and terminate others.
threads = {root_coro: ExceptionEvent(sys.exc_info())}
# If any threads still remain, kill them.
for coro in threads:
coro.close()
# If we're exiting with an exception, raise it in the client.
if exit_te:
exit_te.reraise()
# Sockets and their associated events.
class SocketClosedError(Exception):
pass
class Listener(object):
"""A socket wrapper object for listening sockets.
"""
def __init__(self, host, port):
"""Create a listening socket on the given hostname and port.
"""
self._closed = False
self.host = host
self.port = port
self.sock = socket.socket(socket.AF_INET, socket.SOCK_STREAM)
self.sock.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1)
self.sock.bind((host, port))
self.sock.listen(5)
def accept(self):
"""An event that waits for a connection on the listening socket.
When a connection is made, the event returns a Connection
object.
"""
if self._closed:
raise SocketClosedError()
return AcceptEvent(self)
def close(self):
"""Immediately close the listening socket. (Not an event.)
"""
self._closed = True
self.sock.close()
class Connection(object):
"""A socket wrapper object for connected sockets.
"""
def __init__(self, sock, addr):
self.sock = sock
self.addr = addr
self._buf = b''
self._closed = False
def close(self):
"""Close the connection."""
self._closed = True
self.sock.close()
def recv(self, size):
"""Read at most size bytes of data from the socket."""
if self._closed:
raise SocketClosedError()
if self._buf:
# We already have data read previously.
out = self._buf[:size]
self._buf = self._buf[size:]
return ValueEvent(out)
else:
return ReceiveEvent(self, size)
def send(self, data):
"""Sends data on the socket, returning the number of bytes
successfully sent.
"""
if self._closed:
raise SocketClosedError()
return SendEvent(self, data)
def sendall(self, data):
"""Send all of data on the socket."""
if self._closed:
raise SocketClosedError()
return SendEvent(self, data, True)
def readline(self, terminator=b"\n", bufsize=1024):
"""Reads a line (delimited by terminator) from the socket."""
if self._closed:
raise SocketClosedError()
while True:
if terminator in self._buf:
line, self._buf = self._buf.split(terminator, 1)
line += terminator
yield ReturnEvent(line)
break
data = yield ReceiveEvent(self, bufsize)
if data:
self._buf += data
else:
line = self._buf
self._buf = b''
yield ReturnEvent(line)
break
class AcceptEvent(WaitableEvent):
"""An event for Listener objects (listening sockets) that suspends
execution until the socket gets a connection.
"""
def __init__(self, listener):
self.listener = listener
def waitables(self):
return (self.listener.sock,), (), ()
def fire(self):
sock, addr = self.listener.sock.accept()
return Connection(sock, addr)
class ReceiveEvent(WaitableEvent):
"""An event for Connection objects (connected sockets) for
asynchronously reading data.
"""
def __init__(self, conn, bufsize):
self.conn = conn
self.bufsize = bufsize
def waitables(self):
return (self.conn.sock,), (), ()
def fire(self):
return self.conn.sock.recv(self.bufsize)
class SendEvent(WaitableEvent):
"""An event for Connection objects (connected sockets) for
asynchronously writing data.
"""
def __init__(self, conn, data, sendall=False):
self.conn = conn
self.data = data
self.sendall = sendall
def waitables(self):
return (), (self.conn.sock,), ()
def fire(self):
if self.sendall:
return self.conn.sock.sendall(self.data)
else:
return self.conn.sock.send(self.data)
# Public interface for threads; each returns an event object that
# can immediately be "yield"ed.
def null():
"""Event: yield to the scheduler without doing anything special.
"""
return ValueEvent(None)
def spawn(coro):
"""Event: add another coroutine to the scheduler. Both the parent
and child coroutines run concurrently.
"""
if not isinstance(coro, types.GeneratorType):
raise ValueError('%s is not a coroutine' % str(coro))
return SpawnEvent(coro)
def call(coro):
"""Event: delegate to another coroutine. The current coroutine
is resumed once the sub-coroutine finishes. If the sub-coroutine
returns a value using end(), then this event returns that value.
"""
if not isinstance(coro, types.GeneratorType):
raise ValueError('%s is not a coroutine' % str(coro))
return DelegationEvent(coro)
def end(value=None):
"""Event: ends the coroutine and returns a value to its
delegator.
"""
return ReturnEvent(value)
def read(fd, bufsize=None):
"""Event: read from a file descriptor asynchronously."""
if bufsize is None:
# Read all.
def reader():
buf = []
while True:
data = yield read(fd, 1024)
if not data:
break
buf.append(data)
yield ReturnEvent(''.join(buf))
return DelegationEvent(reader())
else:
return ReadEvent(fd, bufsize)
def write(fd, data):
"""Event: write to a file descriptor asynchronously."""
return WriteEvent(fd, data)
def connect(host, port):
"""Event: connect to a network address and return a Connection
object for communicating on the socket.
"""
addr = (host, port)
sock = socket.create_connection(addr)
return ValueEvent(Connection(sock, addr))
def sleep(duration):
"""Event: suspend the thread for ``duration`` seconds.
"""
return SleepEvent(duration)
def join(coro):
"""Suspend the thread until another, previously `spawn`ed thread
completes.
"""
return JoinEvent(coro)
def kill(coro):
"""Halt the execution of a different `spawn`ed thread.
"""
return KillEvent(coro)
# Convenience function for running socket servers.
def server(host, port, func):
"""A coroutine that runs a network server. Host and port specify the
listening address. func should be a coroutine that takes a single
parameter, a Connection object. The coroutine is invoked for every
incoming connection on the listening socket.
"""
def handler(conn):
try:
yield func(conn)
finally:
conn.close()
listener = Listener(host, port)
try:
while True:
conn = yield listener.accept()
yield spawn(handler(conn))
except KeyboardInterrupt:
pass
finally:
listener.close()
lib/beets/util/confit.py
-1281
View File
File diff suppressed because it is too large Load Diff
lib/beets/util/enumeration.py
-40
View File
@@ -1,40 +0,0 @@
# This file is part of beets.
# Copyright 2013, Adrian Sampson.
#
# Permission is hereby granted, free of charge, to any person obtaining
# a copy of this software and associated documentation files (the
# "Software"), to deal in the Software without restriction, including
# without limitation the rights to use, copy, modify, merge, publish,
# distribute, sublicense, and/or sell copies of the Software, and to
# permit persons to whom the Software is furnished to do so, subject to
# the following conditions:
#
# The above copyright notice and this permission notice shall be
# included in all copies or substantial portions of the Software.
from enum import Enum
class OrderedEnum(Enum):
"""
An Enum subclass that allows comparison of members.
"""
def __ge__(self, other):
if self.__class__ is other.__class__:
return self.value >= other.value
return NotImplemented
def __gt__(self, other):
if self.__class__ is other.__class__:
return self.value > other.value
return NotImplemented
def __le__(self, other):
if self.__class__ is other.__class__:
return self.value <= other.value
return NotImplemented
def __lt__(self, other):
if self.__class__ is other.__class__:
return self.value < other.value
return NotImplemented
lib/beets/util/functemplate.py
-571
View File
@@ -1,571 +0,0 @@
# This file is part of beets.
# Copyright 2013, Adrian Sampson.
#
# Permission is hereby granted, free of charge, to any person obtaining
# a copy of this software and associated documentation files (the
# "Software"), to deal in the Software without restriction, including
# without limitation the rights to use, copy, modify, merge, publish,
# distribute, sublicense, and/or sell copies of the Software, and to
# permit persons to whom the Software is furnished to do so, subject to
# the following conditions:
#
# The above copyright notice and this permission notice shall be
# included in all copies or substantial portions of the Software.
"""This module implements a string formatter based on the standard PEP
292 string.Template class extended with function calls. Variables, as
with string.Template, are indicated with $ and functions are delimited
with %.
This module assumes that everything is Unicode: the template and the
substitution values. Bytestrings are not supported. Also, the templates
always behave like the ``safe_substitute`` method in the standard
library: unknown symbols are left intact.
This is sort of like a tiny, horrible degeneration of a real templating
engine like Jinja2 or Mustache.
"""
from __future__ import print_function
import re
import ast
import dis
import types
SYMBOL_DELIM = u'$'
FUNC_DELIM = u'%'
GROUP_OPEN = u'{'
GROUP_CLOSE = u'}'
ARG_SEP = u','
ESCAPE_CHAR = u'$'
VARIABLE_PREFIX = '__var_'
FUNCTION_PREFIX = '__func_'
class Environment(object):
"""Contains the values and functions to be substituted into a
template.
"""
def __init__(self, values, functions):
self.values = values
self.functions = functions
# Code generation helpers.
def ex_lvalue(name):
"""A variable load expression."""
return ast.Name(name, ast.Store())
def ex_rvalue(name):
"""A variable store expression."""
return ast.Name(name, ast.Load())
def ex_literal(val):
"""An int, float, long, bool, string, or None literal with the given
value.
"""
if val is None:
return ast.Name('None', ast.Load())
elif isinstance(val, (int, float, long)):
return ast.Num(val)
elif isinstance(val, bool):
return ast.Name(str(val), ast.Load())
elif isinstance(val, basestring):
return ast.Str(val)
raise TypeError('no literal for {0}'.format(type(val)))
def ex_varassign(name, expr):
"""Assign an expression into a single variable. The expression may
either be an `ast.expr` object or a value to be used as a literal.
"""
if not isinstance(expr, ast.expr):
expr = ex_literal(expr)
return ast.Assign([ex_lvalue(name)], expr)
def ex_call(func, args):
"""A function-call expression with only positional parameters. The
function may be an expression or the name of a function. Each
argument may be an expression or a value to be used as a literal.
"""
if isinstance(func, basestring):
func = ex_rvalue(func)
args = list(args)
for i in range(len(args)):
if not isinstance(args[i], ast.expr):
args[i] = ex_literal(args[i])
return ast.Call(func, args, [], None, None)
def compile_func(arg_names, statements, name='_the_func', debug=False):
"""Compile a list of statements as the body of a function and return
the resulting Python function. If `debug`, then print out the
bytecode of the compiled function.
"""
func_def = ast.FunctionDef(
name,
ast.arguments(
[ast.Name(n, ast.Param()) for n in arg_names],
None, None,
[ex_literal(None) for _ in arg_names],
),
statements,
[],
)
mod = ast.Module([func_def])
ast.fix_missing_locations(mod)
prog = compile(mod, '<generated>', 'exec')
# Debug: show bytecode.
if debug:
dis.dis(prog)
for const in prog.co_consts:
if isinstance(const, types.CodeType):
dis.dis(const)
the_locals = {}
exec prog in {}, the_locals
return the_locals[name]
# AST nodes for the template language.
class Symbol(object):
"""A variable-substitution symbol in a template."""
def __init__(self, ident, original):
self.ident = ident
self.original = original
def __repr__(self):
return u'Symbol(%s)' % repr(self.ident)
def evaluate(self, env):
"""Evaluate the symbol in the environment, returning a Unicode
string.
"""
if self.ident in env.values:
# Substitute for a value.
return env.values[self.ident]
else:
# Keep original text.
return self.original
def translate(self):
"""Compile the variable lookup."""
expr = ex_rvalue(VARIABLE_PREFIX + self.ident.encode('utf8'))
return [expr], set([self.ident.encode('utf8')]), set()
class Call(object):
"""A function call in a template."""
def __init__(self, ident, args, original):
self.ident = ident
self.args = args
self.original = original
def __repr__(self):
return u'Call(%s, %s, %s)' % (repr(self.ident), repr(self.args),
repr(self.original))
def evaluate(self, env):
"""Evaluate the function call in the environment, returning a
Unicode string.
"""
if self.ident in env.functions:
arg_vals = [expr.evaluate(env) for expr in self.args]
try:
out = env.functions[self.ident](*arg_vals)
except Exception as exc:
# Function raised exception! Maybe inlining the name of
# the exception will help debug.
return u'<%s>' % unicode(exc)
return unicode(out)
else:
return self.original
def translate(self):
"""Compile the function call."""
varnames = set()
funcnames = set([self.ident.encode('utf8')])
arg_exprs = []
for arg in self.args:
subexprs, subvars, subfuncs = arg.translate()
varnames.update(subvars)
funcnames.update(subfuncs)
# Create a subexpression that joins the result components of
# the arguments.
arg_exprs.append(ex_call(
ast.Attribute(ex_literal(u''), 'join', ast.Load()),
[ex_call(
'map',
[
ex_rvalue('unicode'),
ast.List(subexprs, ast.Load()),
]
)],
))
subexpr_call = ex_call(
FUNCTION_PREFIX + self.ident.encode('utf8'),
arg_exprs
)
return [subexpr_call], varnames, funcnames
class Expression(object):
"""Top-level template construct: contains a list of text blobs,
Symbols, and Calls.
"""
def __init__(self, parts):
self.parts = parts
def __repr__(self):
return u'Expression(%s)' % (repr(self.parts))
def evaluate(self, env):
"""Evaluate the entire expression in the environment, returning
a Unicode string.
"""
out = []
for part in self.parts:
if isinstance(part, basestring):
out.append(part)
else:
out.append(part.evaluate(env))
return u''.join(map(unicode, out))
def translate(self):
"""Compile the expression to a list of Python AST expressions, a
set of variable names used, and a set of function names.
"""
expressions = []
varnames = set()
funcnames = set()
for part in self.parts:
if isinstance(part, basestring):
expressions.append(ex_literal(part))
else:
e, v, f = part.translate()
expressions.extend(e)
varnames.update(v)
funcnames.update(f)
return expressions, varnames, funcnames
# Parser.
class ParseError(Exception):
pass
class Parser(object):
"""Parses a template expression string. Instantiate the class with
the template source and call ``parse_expression``. The ``pos`` field
will indicate the character after the expression finished and
``parts`` will contain a list of Unicode strings, Symbols, and Calls
reflecting the concatenated portions of the expression.
This is a terrible, ad-hoc parser implementation based on a
left-to-right scan with no lexing step to speak of; it's probably
both inefficient and incorrect. Maybe this should eventually be
replaced with a real, accepted parsing technique (PEG, parser
generator, etc.).
"""
def __init__(self, string):
self.string = string
self.pos = 0
self.parts = []
# Common parsing resources.
special_chars = (SYMBOL_DELIM, FUNC_DELIM, GROUP_OPEN, GROUP_CLOSE,
ARG_SEP, ESCAPE_CHAR)
special_char_re = re.compile(ur'[%s]|$' %
u''.join(re.escape(c) for c in special_chars))
def parse_expression(self):
"""Parse a template expression starting at ``pos``. Resulting
components (Unicode strings, Symbols, and Calls) are added to
the ``parts`` field, a list. The ``pos`` field is updated to be
the next character after the expression.
"""
text_parts = []
while self.pos < len(self.string):
char = self.string[self.pos]
if char not in self.special_chars:
# A non-special character. Skip to the next special
# character, treating the interstice as literal text.
next_pos = (
self.special_char_re.search(self.string[self.pos:]).start()
+ self.pos
)
text_parts.append(self.string[self.pos:next_pos])
self.pos = next_pos
continue
if self.pos == len(self.string) - 1:
# The last character can never begin a structure, so we
# just interpret it as a literal character (unless it
# terminates the expression, as with , and }).
if char not in (GROUP_CLOSE, ARG_SEP):
text_parts.append(char)
self.pos += 1
break
next_char = self.string[self.pos + 1]
if char == ESCAPE_CHAR and next_char in \
(SYMBOL_DELIM, FUNC_DELIM, GROUP_CLOSE, ARG_SEP):
# An escaped special character ($$, $}, etc.). Note that
# ${ is not an escape sequence: this is ambiguous with
# the start of a symbol and it's not necessary (just
# using { suffices in all cases).
text_parts.append(next_char)
self.pos += 2 # Skip the next character.
continue
# Shift all characters collected so far into a single string.
if text_parts:
self.parts.append(u''.join(text_parts))
text_parts = []
if char == SYMBOL_DELIM:
# Parse a symbol.
self.parse_symbol()
elif char == FUNC_DELIM:
# Parse a function call.
self.parse_call()
elif char in (GROUP_CLOSE, ARG_SEP):
# Template terminated.
break
elif char == GROUP_OPEN:
# Start of a group has no meaning hear; just pass
# through the character.
text_parts.append(char)
self.pos += 1
else:
assert False
# If any parsed characters remain, shift them into a string.
if text_parts:
self.parts.append(u''.join(text_parts))
def parse_symbol(self):
"""Parse a variable reference (like ``$foo`` or ``${foo}``)
starting at ``pos``. Possibly appends a Symbol object (or,
failing that, text) to the ``parts`` field and updates ``pos``.
The character at ``pos`` must, as a precondition, be ``$``.
"""
assert self.pos < len(self.string)
assert self.string[self.pos] == SYMBOL_DELIM
if self.pos == len(self.string) - 1:
# Last character.
self.parts.append(SYMBOL_DELIM)
self.pos += 1
return
next_char = self.string[self.pos + 1]
start_pos = self.pos
self.pos += 1
if next_char == GROUP_OPEN:
# A symbol like ${this}.
self.pos += 1 # Skip opening.
closer = self.string.find(GROUP_CLOSE, self.pos)
if closer == -1 or closer == self.pos:
# No closing brace found or identifier is empty.
self.parts.append(self.string[start_pos:self.pos])
else:
# Closer found.
ident = self.string[self.pos:closer]
self.pos = closer + 1
self.parts.append(Symbol(ident,
self.string[start_pos:self.pos]))
else:
# A bare-word symbol.
ident = self._parse_ident()
if ident:
# Found a real symbol.
self.parts.append(Symbol(ident,
self.string[start_pos:self.pos]))
else:
# A standalone $.
self.parts.append(SYMBOL_DELIM)
def parse_call(self):
"""Parse a function call (like ``%foo{bar,baz}``) starting at
``pos``. Possibly appends a Call object to ``parts`` and update
``pos``. The character at ``pos`` must be ``%``.
"""
assert self.pos < len(self.string)
assert self.string[self.pos] == FUNC_DELIM
start_pos = self.pos
self.pos += 1
ident = self._parse_ident()
if not ident:
# No function name.
self.parts.append(FUNC_DELIM)
return
if self.pos >= len(self.string):
# Identifier terminates string.
self.parts.append(self.string[start_pos:self.pos])
return
if self.string[self.pos] != GROUP_OPEN:
# Argument list not opened.
self.parts.append(self.string[start_pos:self.pos])
return
# Skip past opening brace and try to parse an argument list.
self.pos += 1
args = self.parse_argument_list()
if self.pos >= len(self.string) or \
self.string[self.pos] != GROUP_CLOSE:
# Arguments unclosed.
self.parts.append(self.string[start_pos:self.pos])
return
self.pos += 1 # Move past closing brace.
self.parts.append(Call(ident, args, self.string[start_pos:self.pos]))
def parse_argument_list(self):
"""Parse a list of arguments starting at ``pos``, returning a
list of Expression objects. Does not modify ``parts``. Should
leave ``pos`` pointing to a } character or the end of the
string.
"""
# Try to parse a subexpression in a subparser.
expressions = []
while self.pos < len(self.string):
subparser = Parser(self.string[self.pos:])
subparser.parse_expression()
# Extract and advance past the parsed expression.
expressions.append(Expression(subparser.parts))
self.pos += subparser.pos
if self.pos >= len(self.string) or \
self.string[self.pos] == GROUP_CLOSE:
# Argument list terminated by EOF or closing brace.
break
# Only other way to terminate an expression is with ,.
# Continue to the next argument.
assert self.string[self.pos] == ARG_SEP
self.pos += 1
return expressions
def _parse_ident(self):
"""Parse an identifier and return it (possibly an empty string).
Updates ``pos``.
"""
remainder = self.string[self.pos:]
ident = re.match(ur'\w*', remainder).group(0)
self.pos += len(ident)
return ident
def _parse(template):
"""Parse a top-level template string Expression. Any extraneous text
is considered literal text.
"""
parser = Parser(template)
parser.parse_expression()
parts = parser.parts
remainder = parser.string[parser.pos:]
if remainder:
parts.append(remainder)
return Expression(parts)
# External interface.
class Template(object):
"""A string template, including text, Symbols, and Calls.
"""
def __init__(self, template):
self.expr = _parse(template)
self.original = template
self.compiled = self.translate()
def __eq__(self, other):
return self.original == other.original
def interpret(self, values={}, functions={}):
"""Like `substitute`, but forces the interpreter (rather than
the compiled version) to be used. The interpreter includes
exception-handling code for missing variables and buggy template
functions but is much slower.
"""
return self.expr.evaluate(Environment(values, functions))
def substitute(self, values={}, functions={}):
"""Evaluate the template given the values and functions.
"""
try:
res = self.compiled(values, functions)
except: # Handle any exceptions thrown by compiled version.
res = self.interpret(values, functions)
return res
def translate(self):
"""Compile the template to a Python function."""
expressions, varnames, funcnames = self.expr.translate()
argnames = []
for varname in varnames:
argnames.append(VARIABLE_PREFIX.encode('utf8') + varname)
for funcname in funcnames:
argnames.append(FUNCTION_PREFIX.encode('utf8') + funcname)
func = compile_func(
argnames,
[ast.Return(ast.List(expressions, ast.Load()))],
)
def wrapper_func(values={}, functions={}):
args = {}
for varname in varnames:
args[VARIABLE_PREFIX + varname] = values[varname]
for funcname in funcnames:
args[FUNCTION_PREFIX + funcname] = functions[funcname]
parts = func(**args)
return u''.join(parts)
return wrapper_func
# Performance tests.
if __name__ == '__main__':
import timeit
_tmpl = Template(u'foo $bar %baz{foozle $bar barzle} $bar')
_vars = {'bar': 'qux'}
_funcs = {'baz': unicode.upper}
interp_time = timeit.timeit('_tmpl.interpret(_vars, _funcs)',
'from __main__ import _tmpl, _vars, _funcs',
number=10000)
print(interp_time)
comp_time = timeit.timeit('_tmpl.substitute(_vars, _funcs)',
'from __main__ import _tmpl, _vars, _funcs',
number=10000)
print(comp_time)
print('Speedup:', interp_time / comp_time)
lib/beets/util/pipeline.py
-517
View File
@@ -1,517 +0,0 @@
# This file is part of beets.
# Copyright 2013, Adrian Sampson.
#
# Permission is hereby granted, free of charge, to any person obtaining
# a copy of this software and associated documentation files (the
# "Software"), to deal in the Software without restriction, including
# without limitation the rights to use, copy, modify, merge, publish,
# distribute, sublicense, and/or sell copies of the Software, and to
# permit persons to whom the Software is furnished to do so, subject to
# the following conditions:
#
# The above copyright notice and this permission notice shall be
# included in all copies or substantial portions of the Software.
"""Simple but robust implementation of generator/coroutine-based
pipelines in Python. The pipelines may be run either sequentially
(single-threaded) or in parallel (one thread per pipeline stage).
This implementation supports pipeline bubbles (indications that the
processing for a certain item should abort). To use them, yield the
BUBBLE constant from any stage coroutine except the last.
In the parallel case, the implementation transparently handles thread
shutdown when the processing is complete and when a stage raises an
exception. KeyboardInterrupts (^C) are also handled.
When running a parallel pipeline, it is also possible to use
multiple coroutines for the same pipeline stage; this lets you speed
up a bottleneck stage by dividing its work among multiple threads.
To do so, pass an iterable of coroutines to the Pipeline constructor
in place of any single coroutine.
"""
from __future__ import print_function
import Queue
from threading import Thread, Lock
import sys
BUBBLE = '__PIPELINE_BUBBLE__'
POISON = '__PIPELINE_POISON__'
DEFAULT_QUEUE_SIZE = 16
def _invalidate_queue(q, val=None, sync=True):
"""Breaks a Queue such that it never blocks, always has size 1,
and has no maximum size. get()ing from the queue returns `val`,
which defaults to None. `sync` controls whether a lock is
required (because it's not reentrant!).
"""
def _qsize(len=len):
return 1
def _put(item):
pass
def _get():
return val
if sync:
q.mutex.acquire()
try:
q.maxsize = 0
q._qsize = _qsize
q._put = _put
q._get = _get
q.not_empty.notifyAll()
q.not_full.notifyAll()
finally:
if sync:
q.mutex.release()
class CountedQueue(Queue.Queue):
"""A queue that keeps track of the number of threads that are
still feeding into it. The queue is poisoned when all threads are
finished with the queue.
"""
def __init__(self, maxsize=0):
Queue.Queue.__init__(self, maxsize)
self.nthreads = 0
self.poisoned = False
def acquire(self):
"""Indicate that a thread will start putting into this queue.
Should not be called after the queue is already poisoned.
"""
with self.mutex:
assert not self.poisoned
assert self.nthreads >= 0
self.nthreads += 1
def release(self):
"""Indicate that a thread that was putting into this queue has
exited. If this is the last thread using the queue, the queue
is poisoned.
"""
with self.mutex:
self.nthreads -= 1
assert self.nthreads >= 0
if self.nthreads == 0:
# All threads are done adding to this queue. Poison it
# when it becomes empty.
self.poisoned = True
# Replacement _get invalidates when no items remain.
_old_get = self._get
def _get():
out = _old_get()
if not self.queue:
_invalidate_queue(self, POISON, False)
return out
if self.queue:
# Items remain.
self._get = _get
else:
# No items. Invalidate immediately.
_invalidate_queue(self, POISON, False)
class MultiMessage(object):
"""A message yielded by a pipeline stage encapsulating multiple
values to be sent to the next stage.
"""
def __init__(self, messages):
self.messages = messages
def multiple(messages):
"""Yield multiple([message, ..]) from a pipeline stage to send
multiple values to the next pipeline stage.
"""
return MultiMessage(messages)
def stage(func):
"""Decorate a function to become a simple stage.
>>> @stage
... def add(n, i):
... return i + n
>>> pipe = Pipeline([
... iter([1, 2, 3]),
... add(2),
... ])
>>> list(pipe.pull())
[3, 4, 5]
"""
def coro(*args):
task = None
while True:
task = yield task
task = func(*(args + (task,)))
return coro
def mutator_stage(func):
"""Decorate a function that manipulates items in a coroutine to
become a simple stage.
>>> @mutator_stage
... def setkey(key, item):
... item[key] = True
>>> pipe = Pipeline([
... iter([{'x': False}, {'a': False}]),
... setkey('x'),
... ])
>>> list(pipe.pull())
[{'x': True}, {'a': False, 'x': True}]
"""
def coro(*args):
task = None
while True:
task = yield task
func(*(args + (task,)))
return coro
def _allmsgs(obj):
"""Returns a list of all the messages encapsulated in obj. If obj
is a MultiMessage, returns its enclosed messages. If obj is BUBBLE,
returns an empty list. Otherwise, returns a list containing obj.
"""
if isinstance(obj, MultiMessage):
return obj.messages
elif obj == BUBBLE:
return []
else:
return [obj]
class PipelineThread(Thread):
"""Abstract base class for pipeline-stage threads."""
def __init__(self, all_threads):
super(PipelineThread, self).__init__()
self.abort_lock = Lock()
self.abort_flag = False
self.all_threads = all_threads
self.exc_info = None
def abort(self):
"""Shut down the thread at the next chance possible.
"""
with self.abort_lock:
self.abort_flag = True
# Ensure that we are not blocking on a queue read or write.
if hasattr(self, 'in_queue'):
_invalidate_queue(self.in_queue, POISON)
if hasattr(self, 'out_queue'):
_invalidate_queue(self.out_queue, POISON)
def abort_all(self, exc_info):
"""Abort all other threads in the system for an exception.
"""
self.exc_info = exc_info
for thread in self.all_threads:
thread.abort()
class FirstPipelineThread(PipelineThread):
"""The thread running the first stage in a parallel pipeline setup.
The coroutine should just be a generator.
"""
def __init__(self, coro, out_queue, all_threads):
super(FirstPipelineThread, self).__init__(all_threads)
self.coro = coro
self.out_queue = out_queue
self.out_queue.acquire()
self.abort_lock = Lock()
self.abort_flag = False
def run(self):
try:
while True:
with self.abort_lock:
if self.abort_flag:
return
# Get the value from the generator.
try:
msg = self.coro.next()
except StopIteration:
break
# Send messages to the next stage.
for msg in _allmsgs(msg):
with self.abort_lock:
if self.abort_flag:
return
self.out_queue.put(msg)
except:
self.abort_all(sys.exc_info())
return
# Generator finished; shut down the pipeline.
self.out_queue.release()
class MiddlePipelineThread(PipelineThread):
"""A thread running any stage in the pipeline except the first or
last.
"""
def __init__(self, coro, in_queue, out_queue, all_threads):
super(MiddlePipelineThread, self).__init__(all_threads)
self.coro = coro
self.in_queue = in_queue
self.out_queue = out_queue
self.out_queue.acquire()
def run(self):
try:
# Prime the coroutine.
self.coro.next()
while True:
with self.abort_lock:
if self.abort_flag:
return
# Get the message from the previous stage.
msg = self.in_queue.get()
if msg is POISON:
break
with self.abort_lock:
if self.abort_flag:
return
# Invoke the current stage.
out = self.coro.send(msg)
# Send messages to next stage.
for msg in _allmsgs(out):
with self.abort_lock:
if self.abort_flag:
return
self.out_queue.put(msg)
except:
self.abort_all(sys.exc_info())
return
# Pipeline is shutting down normally.
self.out_queue.release()
class LastPipelineThread(PipelineThread):
"""A thread running the last stage in a pipeline. The coroutine
should yield nothing.
"""
def __init__(self, coro, in_queue, all_threads):
super(LastPipelineThread, self).__init__(all_threads)
self.coro = coro
self.in_queue = in_queue
def run(self):
# Prime the coroutine.
self.coro.next()
try:
while True:
with self.abort_lock:
if self.abort_flag:
return
# Get the message from the previous stage.
msg = self.in_queue.get()
if msg is POISON:
break
with self.abort_lock:
if self.abort_flag:
return
# Send to consumer.
self.coro.send(msg)
except:
self.abort_all(sys.exc_info())
return
class Pipeline(object):
"""Represents a staged pattern of work. Each stage in the pipeline
is a coroutine that receives messages from the previous stage and
yields messages to be sent to the next stage.
"""
def __init__(self, stages):
"""Makes a new pipeline from a list of coroutines. There must
be at least two stages.
"""
if len(stages) < 2:
raise ValueError('pipeline must have at least two stages')
self.stages = []
for stage in stages:
if isinstance(stage, (list, tuple)):
self.stages.append(stage)
else:
# Default to one thread per stage.
self.stages.append((stage,))
def run_sequential(self):
"""Run the pipeline sequentially in the current thread. The
stages are run one after the other. Only the first coroutine
in each stage is used.
"""
list(self.pull())
def run_parallel(self, queue_size=DEFAULT_QUEUE_SIZE):
"""Run the pipeline in parallel using one thread per stage. The
messages between the stages are stored in queues of the given
size.
"""
queue_count = len(self.stages) - 1
queues = [CountedQueue(queue_size) for i in range(queue_count)]
threads = []
# Set up first stage.
for coro in self.stages[0]:
threads.append(FirstPipelineThread(coro, queues[0], threads))
# Middle stages.
for i in range(1, queue_count):
for coro in self.stages[i]:
threads.append(MiddlePipelineThread(
coro, queues[i - 1], queues[i], threads
))
# Last stage.
for coro in self.stages[-1]:
threads.append(
LastPipelineThread(coro, queues[-1], threads)
)
# Start threads.
for thread in threads:
thread.start()
# Wait for termination. The final thread lasts the longest.
try:
# Using a timeout allows us to receive KeyboardInterrupt
# exceptions during the join().
while threads[-1].isAlive():
threads[-1].join(1)
except:
# Stop all the threads immediately.
for thread in threads:
thread.abort()
raise
finally:
# Make completely sure that all the threads have finished
# before we return. They should already be either finished,
# in normal operation, or aborted, in case of an exception.
for thread in threads[:-1]:
thread.join()
for thread in threads:
exc_info = thread.exc_info
if exc_info:
# Make the exception appear as it was raised originally.
raise exc_info[0], exc_info[1], exc_info[2]
def pull(self):
"""Yield elements from the end of the pipeline. Runs the stages
sequentially until the last yields some messages. Each of the messages
is then yielded by ``pulled.next()``. If the pipeline has a consumer,
that is the last stage does not yield any messages, then pull will not
yield any messages. Only the first coroutine in each stage is used
"""
coros = [stage[0] for stage in self.stages]
# "Prime" the coroutines.
for coro in coros[1:]:
coro.next()
# Begin the pipeline.
for out in coros[0]:
msgs = _allmsgs(out)
for coro in coros[1:]:
next_msgs = []
for msg in msgs:
out = coro.send(msg)
next_msgs.extend(_allmsgs(out))
msgs = next_msgs
for msg in msgs:
yield msg
# Smoke test.
if __name__ == '__main__':
import time
# Test a normally-terminating pipeline both in sequence and
# in parallel.
def produce():
for i in range(5):
print('generating %i' % i)
time.sleep(1)
yield i
def work():
num = yield
while True:
print('processing %i' % num)
time.sleep(2)
num = yield num * 2
def consume():
while True:
num = yield
time.sleep(1)
print('received %i' % num)
ts_start = time.time()
Pipeline([produce(), work(), consume()]).run_sequential()
ts_seq = time.time()
Pipeline([produce(), work(), consume()]).run_parallel()
ts_par = time.time()
Pipeline([produce(), (work(), work()), consume()]).run_parallel()
ts_end = time.time()
print('Sequential time:', ts_seq - ts_start)
print('Parallel time:', ts_par - ts_seq)
print('Multiply-parallel time:', ts_end - ts_par)
print()
# Test a pipeline that raises an exception.
def exc_produce():
for i in range(10):
print('generating %i' % i)
time.sleep(1)
yield i
def exc_work():
num = yield
while True:
print('processing %i' % num)
time.sleep(3)
if num == 3:
raise Exception()
num = yield num * 2
def exc_consume():
while True:
num = yield
print('received %i' % num)
Pipeline([exc_produce(), exc_work(), exc_consume()]).run_parallel(1)
lib/beets/vfs.py
-50
View File
@@ -1,50 +0,0 @@
# This file is part of beets.
# Copyright 2013, Adrian Sampson.
#
# Permission is hereby granted, free of charge, to any person obtaining
# a copy of this software and associated documentation files (the
# "Software"), to deal in the Software without restriction, including
# without limitation the rights to use, copy, modify, merge, publish,
# distribute, sublicense, and/or sell copies of the Software, and to
# permit persons to whom the Software is furnished to do so, subject to
# the following conditions:
#
# The above copyright notice and this permission notice shall be
# included in all copies or substantial portions of the Software.
"""A simple utility for constructing filesystem-like trees from beets
libraries.
"""
from collections import namedtuple
from beets import util
Node = namedtuple('Node', ['files', 'dirs'])
def _insert(node, path, itemid):
"""Insert an item into a virtual filesystem node."""
if len(path) == 1:
# Last component. Insert file.
node.files[path[0]] = itemid
else:
# In a directory.
dirname = path[0]
rest = path[1:]
if dirname not in node.dirs:
node.dirs[dirname] = Node({}, {})
_insert(node.dirs[dirname], rest, itemid)
def libtree(lib):
"""Generates a filesystem-like directory tree for the files
contained in `lib`. Filesystem nodes are (files, dirs) named
tuples in which both components are dictionaries. The first
maps filenames to Item ids. The second maps directory names to
child node tuples.
"""
root = Node({}, {})
for item in lib.items():
dest = item.destination(fragment=True)
parts = util.components(dest)
_insert(root, parts, item.id)
return root