Updated beets lib ( + mutagen ) and added unidecode

This commit is contained in:
rembo10
2014-03-31 16:21:50 -07:00
parent f0b4f2f3c8
commit 2ba476c6ef
256 changed files with 63907 additions and 8702 deletions
+20
View File
@@ -0,0 +1,20 @@
# 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
+727
View File
@@ -0,0 +1,727 @@
# 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 beets
from beets.util.functemplate import Template
from .query import MatchQuery
# 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.
"""
_bytes_keys = ()
"""Keys whose values should be stored as raw bytes blobs rather than
strings.
"""
_search_fields = ()
"""The fields that should be queried by default by unqualified query
terms.
"""
@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()
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.
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.
"""
source = self._values_fixed if key in self._fields \
else self._values_flex
old_value = source.get(key)
source[key] = value
if 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 += key + '=?,'
value = self[key]
# Wrap path strings in buffers so they get stored
# "in the raw".
if key in self._bytes_keys and isinstance(value, str):
value = buffer(value)
subvars.append(value)
assignments = assignments[:-1] # Knock off last ,
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.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.
@classmethod
def _format(cls, key, value, for_path=False):
"""Format a value as the given field for this model.
"""
# Format the value as a string according to its type, if any.
if key in cls._fields:
value = cls._fields[key].format(value)
# Formatting must result in a string. To deal with
# Python2isms, implicitly convert ASCII strings.
assert isinstance(value, basestring), \
u'field formatter must produce strings'
if isinstance(value, bytes):
value = value.decode('utf8', 'ignore')
elif not isinstance(value, unicode):
# Fallback formatter. Convert to unicode at all cost.
if value is None:
value = u''
elif isinstance(value, basestring):
if isinstance(value, bytes):
value = value.decode('utf8', 'ignore')
else:
value = unicode(value)
if 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
def _get_formatted(self, key, for_path=False):
"""Get a field value formatted as a string (`unicode` object)
for display to the user. If `for_path` is true, then the value
will be sanitized for inclusion in a pathname (i.e., path
separators will be removed from the value).
"""
return self._format(key, self.get(key), for_path)
def _formatted_mapping(self, for_path=False):
"""Get a mapping containing all values on this object formatted
as human-readable strings.
"""
# In the future, this could be made "lazy" to avoid computing
# fields unnecessarily.
out = {}
for key in self.keys(True):
out[key] = self._get_formatted(key, for_path)
return out
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.
"""
# Build value mapping.
mapping = self._formatted_mapping(for_path)
# Get template functions.
funcs = self._template_funcs()
# Perform substitution.
if isinstance(template, basestring):
template = Template(template)
return template.substitute(mapping, 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")
typ = cls._fields.get(key)
if typ:
return typ.parse(string)
else:
# Fall back to unparsed string.
return 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):
"""Create a result set that will construct objects of type
`model_class`, which should be a subclass of `LibModel`, out of
the query result mapping in `rows`. The new objects are
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.
"""
self.model_class = model_class
self.rows = rows
self.db = db
self.query = query
def __iter__(self):
"""Construct Python objects for all rows that pass the query
predicate.
"""
for row in self.rows:
# 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'],)
)
values = dict(row)
values.update(
dict((row['key'], row['value']) for row in flex_rows)
)
# Construct the Python object and yield it if it passes the
# predicate.
obj = self.model_class(self.db, **values)
if not self.query or self.query.match(obj):
yield obj
def __len__(self):
"""Get the number of matching objects.
"""
if 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 len(self.rows)
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.
"""
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, order_by=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). If provided,
`order_by` is a SQLite ORDER BY clause for sorting.
"""
where, subvals = query.clause()
sql = "SELECT * FROM {0} WHERE {1}".format(
model_cls._table,
where or '1',
)
if order_by:
sql += " ORDER BY {0}".format(order_by)
with self.transaction() as tx:
rows = tx.query(sql, subvals)
return Results(model_cls, rows, self, None if where else query)
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()
+494
View File
@@ -0,0 +1,494 @@
# 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 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 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):
search = '%' + (self.pattern.replace('\\','\\\\').replace('%','\\%')
.replace('_','\\_')) + '%'
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):
value = getattr(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):
raise ValueError('date is not in one of the formats '
+ ', '.join(cls.date_formats))
date_format = cls.date_formats[ordinal]
date = datetime.strptime(string, date_format)
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
+140
View File
@@ -0,0 +1,140 @@
# 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 the value in the database, query,
format, and parse a given field.
"""
sql = None
"""The SQLite column type for the value.
"""
query = None
"""The `Query` subclass to be used when querying the field.
"""
def format(self, value):
"""Given a value of this type, produce a Unicode string
representing the value. This is used in template evaluation.
"""
raise NotImplementedError()
def parse(self, string):
"""Parse a (possibly human-written) string and return the
indicated value of this type.
"""
raise NotImplementedError()
# Reusable types.
class Integer(Type):
"""A basic integer type.
"""
sql = u'INTEGER'
query = query.NumericQuery
def format(self, value):
return unicode(value or 0)
def parse(self, string):
try:
return int(string)
except ValueError:
return 0
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 key for a SQLite table.
"""
sql = u'INTEGER PRIMARY KEY'
class Float(Type):
"""A basic floating-point type.
"""
sql = u'REAL'
query = query.NumericQuery
def format(self, value):
return u'{0:.1f}'.format(value or 0.0)
def parse(self, string):
try:
return float(string)
except ValueError:
return 0.0
class String(Type):
"""A Unicode string type.
"""
sql = u'TEXT'
query = query.SubstringQuery
def format(self, value):
return unicode(value) if value else u''
def parse(self, string):
return string
class Boolean(Type):
"""A boolean type.
"""
sql = u'INTEGER'
query = query.BooleanQuery
def format(self, value):
return unicode(bool(value))
def parse(self, string):
return str2bool(string)