from copy import copy
from django.conf import settings
from django.db.models.expressions import Func, Value
from django.db.models.fields import (
DateField, DateTimeField, Field, IntegerField, TimeField,
)
from django.db.models.query_utils import RegisterLookupMixin
from django.utils import timezone
from django.utils.functional import cached_property
from django.utils.six.moves import range
class Lookup(object):
lookup_name = None
def __init__(self, lhs, rhs):
self.lhs, self.rhs = lhs, rhs
self.rhs = self.get_prep_lookup()
if hasattr(self.lhs, 'get_bilateral_transforms'):
bilateral_transforms = self.lhs.get_bilateral_transforms()
else:
bilateral_transforms = []
if bilateral_transforms:
# We should warn the user as soon as possible if he is trying to apply
# a bilateral transformation on a nested QuerySet: that won't work.
# We need to import QuerySet here so as to avoid circular
from django.db.models.query import QuerySet
if isinstance(rhs, QuerySet):
raise NotImplementedError("Bilateral transformations on nested querysets are not supported.")
self.bilateral_transforms = bilateral_transforms
def apply_bilateral_transforms(self, value):
for transform in self.bilateral_transforms:
value = transform(value)
return value
def batch_process_rhs(self, compiler, connection, rhs=None):
if rhs is None:
rhs = self.rhs
if self.bilateral_transforms:
sqls, sqls_params = [], []
for p in rhs:
value = Value(p, output_field=self.lhs.output_field)
value = self.apply_bilateral_transforms(value)
value = value.resolve_expression(compiler.query)
sql, sql_params = compiler.compile(value)
sqls.append(sql)
sqls_params.extend(sql_params)
else:
params = self.lhs.output_field.get_db_prep_lookup(
self.lookup_name, rhs, connection, prepared=True)
sqls, sqls_params = ['%s'] * len(params), params
return sqls, sqls_params
def get_prep_lookup(self):
return self.lhs.output_field.get_prep_lookup(self.lookup_name, self.rhs)
def get_db_prep_lookup(self, value, connection):
return (
'%s', self.lhs.output_field.get_db_prep_lookup(
self.lookup_name, value, connection, prepared=True))
def process_lhs(self, compiler, connection, lhs=None):
lhs = lhs or self.lhs
return compiler.compile(lhs)
def process_rhs(self, compiler, connection):
value = self.rhs
if self.bilateral_transforms:
if self.rhs_is_direct_value():
# Do not call get_db_prep_lookup here as the value will be
# transformed before being used for lookup
value = Value(value, output_field=self.lhs.output_field)
value = self.apply_bilateral_transforms(value)
value = value.resolve_expression(compiler.query)
# Due to historical reasons there are a couple of different
# ways to produce sql here. get_compiler is likely a Query
# instance, _as_sql QuerySet and as_sql just something with
# as_sql. Finally the value can of course be just plain
# Python value.
if hasattr(value, 'get_compiler'):
value = value.get_compiler(connection=connection)
if hasattr(value, 'as_sql'):
sql, params = compiler.compile(value)
return '(' + sql + ')', params
if hasattr(value, '_as_sql'):
sql, params = value._as_sql(connection=connection)
return '(' + sql + ')', params
else:
return self.get_db_prep_lookup(value, connection)
def rhs_is_direct_value(self):
return not(
hasattr(self.rhs, 'as_sql') or
hasattr(self.rhs, '_as_sql') or
hasattr(self.rhs, 'get_compiler'))
def relabeled_clone(self, relabels):
new = copy(self)
new.lhs = new.lhs.relabeled_clone(relabels)
if hasattr(new.rhs, 'relabeled_clone'):
new.rhs = new.rhs.relabeled_clone(relabels)
return new
def get_group_by_cols(self):
cols = self.lhs.get_group_by_cols()
if hasattr(self.rhs, 'get_group_by_cols'):
cols.extend(self.rhs.get_group_by_cols())
return cols
def as_sql(self, compiler, connection):
raise NotImplementedError
@cached_property
def contains_aggregate(self):
return self.lhs.contains_aggregate or getattr(self.rhs, 'contains_aggregate', False)
class Transform(RegisterLookupMixin, Func):
"""
RegisterLookupMixin() is first so that get_lookup() and get_transform()
first examine self and then check output_field.
"""
bilateral = False
def __init__(self, expression, **extra):
# Restrict Transform to allow only a single expression.
super(Transform, self).__init__(expression, **extra)
@property
def lhs(self):
return self.get_source_expressions()[0]
def get_bilateral_transforms(self):
if hasattr(self.lhs, 'get_bilateral_transforms'):
bilateral_transforms = self.lhs.get_bilateral_transforms()
else:
bilateral_transforms = []
if self.bilateral:
bilateral_transforms.append(self.__class__)
return bilateral_transforms
class BuiltinLookup(Lookup):
def process_lhs(self, compiler, connection, lhs=None):
lhs_sql, params = super(BuiltinLookup, self).process_lhs(
compiler, connection, lhs)
field_internal_type = self.lhs.output_field.get_internal_type()
db_type = self.lhs.output_field.db_type(connection=connection)
lhs_sql = connection.ops.field_cast_sql(
db_type, field_internal_type) % lhs_sql
lhs_sql = connection.ops.lookup_cast(self.lookup_name, field_internal_type) % lhs_sql
return lhs_sql, list(params)
def as_sql(self, compiler, connection):
lhs_sql, params = self.process_lhs(compiler, connection)
rhs_sql, rhs_params = self.process_rhs(compiler, connection)
params.extend(rhs_params)
rhs_sql = self.get_rhs_op(connection, rhs_sql)
return '%s %s' % (lhs_sql, rhs_sql), params
def get_rhs_op(self, connection, rhs):
return connection.operators[self.lookup_name] % rhs
class Exact(BuiltinLookup):
lookup_name = 'exact'
Field.register_lookup(Exact)
class IExact(BuiltinLookup):
lookup_name = 'iexact'
def process_rhs(self, qn, connection):
rhs, params = super(IExact, self).process_rhs(qn, connection)
if params:
params[0] = connection.ops.prep_for_iexact_query(params[0])
return rhs, params
Field.register_lookup(IExact)
class GreaterThan(BuiltinLookup):
lookup_name = 'gt'
Field.register_lookup(GreaterThan)
class GreaterThanOrEqual(BuiltinLookup):
lookup_name = 'gte'
Field.register_lookup(GreaterThanOrEqual)
class LessThan(BuiltinLookup):
lookup_name = 'lt'
Field.register_lookup(LessThan)
class LessThanOrEqual(BuiltinLookup):
lookup_name = 'lte'
Field.register_lookup(LessThanOrEqual)
class In(BuiltinLookup):
lookup_name = 'in'
def process_rhs(self, compiler, connection):
if self.rhs_is_direct_value():
# rhs should be an iterable, we use batch_process_rhs
# to prepare/transform those values
rhs = list(self.rhs)
if not rhs:
from django.db.models.sql.datastructures import EmptyResultSet
raise EmptyResultSet
sqls, sqls_params = self.batch_process_rhs(compiler, connection, rhs)
placeholder = '(' + ', '.join(sqls) + ')'
return (placeholder, sqls_params)
else:
return super(In, self).process_rhs(compiler, connection)
def get_rhs_op(self, connection, rhs):
return 'IN %s' % rhs
def as_sql(self, compiler, connection):
max_in_list_size = connection.ops.max_in_list_size()
if self.rhs_is_direct_value() and max_in_list_size and len(self.rhs) > max_in_list_size:
return self.split_parameter_list_as_sql(compiler, connection)
return super(In, self).as_sql(compiler, connection)
def split_parameter_list_as_sql(self, compiler, connection):
# This is a special case for databases which limit the number of
# elements which can appear in an 'IN' clause.
max_in_list_size = connection.ops.max_in_list_size()
lhs, lhs_params = self.process_lhs(compiler, connection)
rhs, rhs_params = self.batch_process_rhs(compiler, connection)
in_clause_elements = ['(']
params = []
for offset in range(0, len(rhs_params), max_in_list_size):
if offset > 0:
in_clause_elements.append(' OR ')
in_clause_elements.append('%s IN (' % lhs)
params.extend(lhs_params)
sqls = rhs[offset: offset + max_in_list_size]
sqls_params = rhs_params[offset: offset + max_in_list_size]
param_group = ', '.join(sqls)
in_clause_elements.append(param_group)
in_clause_elements.append(')')
params.extend(sqls_params)
in_clause_elements.append(')')
return ''.join(in_clause_elements), params
Field.register_lookup(In)
class PatternLookup(BuiltinLookup):
def get_rhs_op(self, connection, rhs):
# Assume we are in startswith. We need to produce SQL like:
# col LIKE %s, ['thevalue%']
# For python values we can (and should) do that directly in Python,
# but if the value is for example reference to other column, then
# we need to add the % pattern match to the lookup by something like
# col LIKE othercol || '%%'
# So, for Python values we don't need any special pattern, but for
# SQL reference values or SQL transformations we need the correct
# pattern added.
if (hasattr(self.rhs, 'get_compiler') or hasattr(self.rhs, 'as_sql')
or hasattr(self.rhs, '_as_sql') or self.bilateral_transforms):
pattern = connection.pattern_ops[self.lookup_name].format(connection.pattern_esc)
return pattern.format(rhs)
else:
return super(PatternLookup, self).get_rhs_op(connection, rhs)
class Contains(PatternLookup):
lookup_name = 'contains'
def process_rhs(self, qn, connection):
rhs, params = super(Contains, self).process_rhs(qn, connection)
if params and not self.bilateral_transforms:
params[0] = "%%%s%%" % connection.ops.prep_for_like_query(params[0])
return rhs, params
Field.register_lookup(Contains)
class IContains(Contains):
lookup_name = 'icontains'
Field.register_lookup(IContains)
class StartsWith(PatternLookup):
lookup_name = 'startswith'
def process_rhs(self, qn, connection):
rhs, params = super(StartsWith, self).process_rhs(qn, connection)
if params and not self.bilateral_transforms:
params[0] = "%s%%" % connection.ops.prep_for_like_query(params[0])
return rhs, params
Field.register_lookup(StartsWith)
class IStartsWith(PatternLookup):
lookup_name = 'istartswith'
def process_rhs(self, qn, connection):
rhs, params = super(IStartsWith, self).process_rhs(qn, connection)
if params and not self.bilateral_transforms:
params[0] = "%s%%" % connection.ops.prep_for_like_query(params[0])
return rhs, params
Field.register_lookup(IStartsWith)
class EndsWith(PatternLookup):
lookup_name = 'endswith'
def process_rhs(self, qn, connection):
rhs, params = super(EndsWith, self).process_rhs(qn, connection)
if params and not self.bilateral_transforms:
params[0] = "%%%s" % connection.ops.prep_for_like_query(params[0])
return rhs, params
Field.register_lookup(EndsWith)
class IEndsWith(PatternLookup):
lookup_name = 'iendswith'
def process_rhs(self, qn, connection):
rhs, params = super(IEndsWith, self).process_rhs(qn, connection)
if params and not self.bilateral_transforms:
params[0] = "%%%s" % connection.ops.prep_for_like_query(params[0])
return rhs, params
Field.register_lookup(IEndsWith)
class Between(BuiltinLookup):
def get_rhs_op(self, connection, rhs):
return "BETWEEN %s AND %s" % (rhs, rhs)
class Range(BuiltinLookup):
lookup_name = 'range'
def get_rhs_op(self, connection, rhs):
return "BETWEEN %s AND %s" % (rhs[0], rhs[1])
Loading ...