deb-heat/heat/engine/constraints.py
Crag Wolfe f3103c6b21 Add OS::Heat::Value resource
The OS::Heat::Value resource provides a "value" attribute that may be
calculated through references to other parameters or resources. Other
resources may then reference this value.

The primary motivation is to avoid having to create nested templates
that only exist to execute these kinds of data transformations.

This resource's "type" property is optional but recommended.

Change-Id: I72fdab8261336644a9545668fac6f4e49a785e36
2016-08-20 12:33:21 -04:00

651 lines
22 KiB
Python

#
# Licensed under the Apache License, Version 2.0 (the "License"); you may
# not use this file except in compliance with the License. You may obtain
# a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
# WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. See the
# License for the specific language governing permissions and limitations
# under the License.
import collections
import numbers
import re
from oslo_cache import core
from oslo_config import cfg
from oslo_log import log
from oslo_utils import reflection
from oslo_utils import strutils
import six
from heat.common import cache
from heat.common import exception
from heat.common.i18n import _
from heat.common.i18n import _LW
from heat.engine import resources
# decorator that allows to cache the value
# of the function based on input arguments
MEMOIZE = core.get_memoization_decorator(conf=cfg.CONF,
region=cache.get_cache_region(),
group="constraint_validation_cache")
LOG = log.getLogger(__name__)
class Schema(collections.Mapping):
"""Schema base class for validating properties or parameters.
Schema objects are serializable to dictionaries following a superset of
the HOT input Parameter schema using dict().
Serialises to JSON in the form::
{
'type': 'list',
'required': False
'constraints': [
{
'length': {'min': 1},
'description': 'List must not be empty'
}
],
'schema': {
'*': {
'type': 'string'
}
},
'description': 'An example list property.'
}
"""
KEYS = (
TYPE, DESCRIPTION, DEFAULT, SCHEMA, REQUIRED, CONSTRAINTS,
IMMUTABLE,
) = (
'type', 'description', 'default', 'schema', 'required', 'constraints',
'immutable',
)
# Keywords for data types; each Schema subclass can define its respective
# type name used in templates
TYPE_KEYS = (
INTEGER_TYPE, STRING_TYPE, NUMBER_TYPE, BOOLEAN_TYPE, MAP_TYPE,
LIST_TYPE,
) = (
'INTEGER', 'STRING', 'NUMBER', 'BOOLEAN', 'MAP',
'LIST',
)
# Default type names for data types used in templates; can be overridden by
# subclasses
TYPES = (
INTEGER, STRING, NUMBER, BOOLEAN, MAP, LIST, ANY,
) = (
'Integer', 'String', 'Number', 'Boolean', 'Map', 'List', 'Any',
)
def __init__(self, data_type, description=None,
default=None, schema=None,
required=False, constraints=None, label=None,
immutable=False):
self._len = None
self.label = label
self.type = data_type
if self.type not in self.TYPES:
raise exception.InvalidSchemaError(
message=_('Invalid type (%s)') % self.type)
if required and default is not None:
LOG.warning(_LW("Option 'required=True' should not be used with "
"any 'default' value (%s)") % default)
self.description = description
self.required = required
self.immutable = immutable
if isinstance(schema, type(self)):
if self.type != self.LIST:
msg = _('Single schema valid only for '
'%(ltype)s, not %(utype)s') % dict(ltype=self.LIST,
utype=self.type)
raise exception.InvalidSchemaError(message=msg)
self.schema = AnyIndexDict(schema)
else:
self.schema = schema
if self.schema is not None and self.type not in (self.LIST,
self.MAP):
msg = _('Schema valid only for %(ltype)s or '
'%(mtype)s, not %(utype)s') % dict(ltype=self.LIST,
mtype=self.MAP,
utype=self.type)
raise exception.InvalidSchemaError(message=msg)
self.constraints = constraints or []
self.default = default
def validate(self, context=None):
"""Validates the schema.
This method checks if the schema itself is valid, and if the
default value - if present - complies to the schema's constraints.
"""
for c in self.constraints:
if not self._is_valid_constraint(c):
err_msg = _('%(name)s constraint '
'invalid for %(utype)s') % dict(
name=type(c).__name__,
utype=self.type)
raise exception.InvalidSchemaError(message=err_msg)
self._validate_default(context)
# validated nested schema(ta)
if self.schema:
if isinstance(self.schema, AnyIndexDict):
self.schema.value.validate(context)
else:
for nested_schema in six.itervalues(self.schema):
nested_schema.validate(context)
def _validate_default(self, context):
if self.default is not None:
try:
self.validate_constraints(self.default, context,
[CustomConstraint])
except (ValueError, TypeError) as exc:
raise exception.InvalidSchemaError(
message=_('Invalid default %(default)s (%(exc)s)') %
dict(default=self.default, exc=exc))
def set_default(self, default=None):
"""Set the default value for this Schema object."""
self.default = default
def _is_valid_constraint(self, constraint):
valid_types = getattr(constraint, 'valid_types', [])
return any(self.type == getattr(self, t, None) for t in valid_types)
@staticmethod
def str_to_num(value):
"""Convert a string representation of a number into a numeric type."""
if isinstance(value, numbers.Number):
return value
try:
return int(value)
except ValueError:
return float(value)
def to_schema_type(self, value):
"""Returns the value in the schema's data type."""
try:
# We have to be backwards-compatible for Integer and Number
# Schema types and try to convert string representations of
# number into "real" number types, therefore calling
# str_to_num below.
if self.type == self.INTEGER:
num = Schema.str_to_num(value)
if isinstance(num, float):
raise ValueError(_('%s is not an integer.') % num)
return num
elif self.type == self.NUMBER:
return Schema.str_to_num(value)
elif self.type == self.STRING:
return six.text_type(value)
elif self.type == self.BOOLEAN:
return strutils.bool_from_string(str(value), strict=True)
except ValueError:
raise ValueError(_('Value "%(val)s" is invalid for data type '
'"%(type)s".')
% {'val': value, 'type': self.type})
return value
def validate_constraints(self, value, context=None, skipped=None,
template=None):
if not skipped:
skipped = []
try:
for constraint in self.constraints:
if type(constraint) not in skipped:
constraint.validate(value, self, context, template)
except ValueError as ex:
raise exception.StackValidationFailed(message=six.text_type(ex))
def __getitem__(self, key):
if key == self.TYPE:
return self.type.lower()
elif key == self.DESCRIPTION:
if self.description is not None:
return self.description
elif key == self.DEFAULT:
if self.default is not None:
return self.default
elif key == self.SCHEMA:
if self.schema is not None:
return dict((n, dict(s)) for n, s in self.schema.items())
elif key == self.REQUIRED:
return self.required
elif key == self.CONSTRAINTS:
if self.constraints:
return [dict(c) for c in self.constraints]
raise KeyError(key)
def __iter__(self):
for k in self.KEYS:
try:
self[k]
except KeyError:
pass
else:
yield k
def __len__(self):
if self._len is None:
self._len = len(list(iter(self)))
return self._len
class AnyIndexDict(collections.Mapping):
"""A Mapping that returns the same value for any integer index.
Used for storing the schema for a list. When converted to a dictionary,
it contains a single item with the key '*'.
"""
ANYTHING = '*'
def __init__(self, value):
self.value = value
def __getitem__(self, key):
if key != self.ANYTHING and not isinstance(key, six.integer_types):
raise KeyError(_('Invalid key %s') % str(key))
return self.value
def __iter__(self):
yield self.ANYTHING
def __len__(self):
return 1
class Constraint(collections.Mapping):
"""Parent class for constraints on allowable values for a Property.
Constraints are serializable to dictionaries following the HOT input
Parameter constraints schema using dict().
"""
(DESCRIPTION,) = ('description',)
def __init__(self, description=None):
self.description = description
def __str__(self):
def desc():
if self.description:
yield self.description
yield self._str()
return '\n'.join(desc())
def validate(self, value, schema=None, context=None, template=None):
if not self._is_valid(value, schema, context, template):
if self.description:
err_msg = self.description
else:
err_msg = self._err_msg(value)
raise ValueError(err_msg)
@classmethod
def _name(cls):
return '_'.join(w.lower() for w in re.findall('[A-Z]?[a-z]+',
cls.__name__))
def __getitem__(self, key):
if key == self.DESCRIPTION:
if self.description is None:
raise KeyError(key)
return self.description
if key == self._name():
return self._constraint()
raise KeyError(key)
def __iter__(self):
if self.description is not None:
yield self.DESCRIPTION
yield self._name()
def __len__(self):
return 2 if self.description is not None else 1
class Range(Constraint):
"""Constrain values within a range.
Serializes to JSON as::
{
'range': {'min': <min>, 'max': <max>},
'description': <description>
}
"""
(MIN, MAX) = ('min', 'max')
valid_types = (Schema.INTEGER_TYPE, Schema.NUMBER_TYPE,)
def __init__(self, min=None, max=None, description=None):
super(Range, self).__init__(description)
self.min = min
self.max = max
for param in (min, max):
if not isinstance(param, (float, six.integer_types, type(None))):
raise exception.InvalidSchemaError(
message=_('min/max must be numeric'))
if min is max is None:
raise exception.InvalidSchemaError(
message=_('A range constraint must have a min value and/or '
'a max value specified.'))
def _str(self):
if self.max is None:
fmt = _('The value must be at least %(min)s.')
elif self.min is None:
fmt = _('The value must be no greater than %(max)s.')
else:
fmt = _('The value must be in the range %(min)s to %(max)s.')
return fmt % self._constraint()
def _err_msg(self, value):
return '%s is out of range (min: %s, max: %s)' % (value,
self.min,
self.max)
def _is_valid(self, value, schema, context, template):
value = Schema.str_to_num(value)
if self.min is not None:
if value < self.min:
return False
if self.max is not None:
if value > self.max:
return False
return True
def _constraint(self):
def constraints():
if self.min is not None:
yield self.MIN, self.min
if self.max is not None:
yield self.MAX, self.max
return dict(constraints())
class Length(Range):
"""Constrain the length of values within a range.
Serializes to JSON as::
{
'length': {'min': <min>, 'max': <max>},
'description': <description>
}
"""
valid_types = (Schema.STRING_TYPE, Schema.LIST_TYPE, Schema.MAP_TYPE,)
def __init__(self, min=None, max=None, description=None):
if min is max is None:
raise exception.InvalidSchemaError(
message=_('A length constraint must have a min value and/or '
'a max value specified.'))
super(Length, self).__init__(min, max, description)
for param in (min, max):
if not isinstance(param, (six.integer_types, type(None))):
msg = _('min/max length must be integral')
raise exception.InvalidSchemaError(message=msg)
def _str(self):
if self.max is None:
fmt = _('The length must be at least %(min)s.')
elif self.min is None:
fmt = _('The length must be no greater than %(max)s.')
else:
fmt = _('The length must be in the range %(min)s to %(max)s.')
return fmt % self._constraint()
def _err_msg(self, value):
return 'length (%d) is out of range (min: %s, max: %s)' % (len(value),
self.min,
self.max)
def _is_valid(self, value, schema, context, template):
return super(Length, self)._is_valid(len(value), schema, context,
template)
class AllowedValues(Constraint):
"""Constrain values to a predefined set.
Serializes to JSON as::
{
'allowed_values': [<allowed1>, <allowed2>, ...],
'description': <description>
}
"""
valid_types = (Schema.STRING_TYPE, Schema.INTEGER_TYPE, Schema.NUMBER_TYPE,
Schema.BOOLEAN_TYPE, Schema.LIST_TYPE,)
def __init__(self, allowed, description=None):
super(AllowedValues, self).__init__(description)
if (not isinstance(allowed, collections.Sequence) or
isinstance(allowed, six.string_types)):
raise exception.InvalidSchemaError(
message=_('AllowedValues must be a list'))
self.allowed = tuple(allowed)
def _str(self):
allowed = ', '.join(str(a) for a in self.allowed)
return _('Allowed values: %s') % allowed
def _err_msg(self, value):
allowed = '[%s]' % ', '.join(str(a) for a in self.allowed)
return '"%s" is not an allowed value %s' % (value, allowed)
def _is_valid(self, value, schema, context, template):
# For list values, check if all elements of the list are contained
# in allowed list.
if isinstance(value, list):
return all(v in self.allowed for v in value)
if schema is not None:
_allowed = tuple(schema.to_schema_type(v) for v in self.allowed)
return schema.to_schema_type(value) in _allowed
return value in self.allowed
def _constraint(self):
return list(self.allowed)
class AllowedPattern(Constraint):
"""Constrain values to a predefined regular expression pattern.
Serializes to JSON as::
{
'allowed_pattern': <pattern>,
'description': <description>
}
"""
valid_types = (Schema.STRING_TYPE,)
def __init__(self, pattern, description=None):
super(AllowedPattern, self).__init__(description)
if not isinstance(pattern, six.string_types):
raise exception.InvalidSchemaError(
message=_('AllowedPattern must be a string'))
self.pattern = pattern
self.match = re.compile(pattern).match
def _str(self):
return _('Value must match pattern: %s') % self.pattern
def _err_msg(self, value):
return '"%s" does not match pattern "%s"' % (value, self.pattern)
def _is_valid(self, value, schema, context, template):
match = self.match(value)
return match is not None and match.end() == len(value)
def _constraint(self):
return self.pattern
class CustomConstraint(Constraint):
"""A constraint delegating validation to an external class."""
valid_types = (Schema.STRING_TYPE, Schema.INTEGER_TYPE, Schema.NUMBER_TYPE,
Schema.BOOLEAN_TYPE, Schema.LIST_TYPE)
def __init__(self, name, description=None, environment=None):
super(CustomConstraint, self).__init__(description)
self.name = name
self._environment = environment
self._custom_constraint = None
def _constraint(self):
return self.name
@property
def custom_constraint(self):
if self._custom_constraint is None:
if self._environment is None:
self._environment = resources.global_env()
constraint_class = self._environment.get_constraint(self.name)
if constraint_class:
self._custom_constraint = constraint_class()
return self._custom_constraint
def _str(self):
message = getattr(self.custom_constraint, "message", None)
if not message:
message = _('Value must be of type %s') % self.name
return message
def _err_msg(self, value):
constraint = self.custom_constraint
if constraint is None:
return _('"%(value)s" does not validate %(name)s '
'(constraint not found)') % {
"value": value, "name": self.name}
error = getattr(constraint, "error", None)
if error:
return error(value)
return _('"%(value)s" does not validate %(name)s') % {
"value": value, "name": self.name}
def _is_valid(self, value, schema, context, template):
constraint = self.custom_constraint
if not constraint:
return False
try:
result = constraint.validate(value, context,
template=template)
except TypeError:
# for backwards compatibility with older service constraints
result = constraint.validate(value, context)
return result
class BaseCustomConstraint(object):
"""A base class for validation using API clients.
It will provide a better error message, and reduce a bit of duplication.
Subclass must provide `expected_exceptions` and implement
`validate_with_client`.
"""
expected_exceptions = (exception.EntityNotFound,)
resource_client_name = None
resource_getter_name = None
_error_message = None
def error(self, value):
if self._error_message is None:
return _("Error validating value '%(value)s'") % {"value": value}
return _("Error validating value '%(value)s': %(message)s") % {
"value": value, "message": self._error_message}
def validate(self, value, context, template=None):
@MEMOIZE
def check_cache_or_validate_value(cache_value_prefix,
value_to_validate):
"""Check if validation result stored in cache or validate value.
The function checks that value was validated and validation
result stored in cache. If not then it executes validation and
stores the result of validation in cache.
If caching is disabled it requests for validation each time.
:param cache_value_prefix: cache prefix that used to distinguish
value in heat cache. So the cache key
would be the following:
cache_value_prefix + value_to_validate.
:param value_to_validate: value that need to be validated
:return: True if value is valid otherwise False
"""
try:
self.validate_with_client(context.clients, value_to_validate)
except self.expected_exceptions as e:
self._error_message = str(e)
return False
else:
return True
class_name = reflection.get_class_name(self, fully_qualified=False)
cache_value_prefix = "{0}:{1}".format(class_name,
six.text_type(context.tenant_id))
validation_result = check_cache_or_validate_value(
cache_value_prefix, value)
# if validation failed we should not store it in cache
# cause validation will be fixed soon (by admin or other guy)
# and we don't need to require user wait for expiration time
if not validation_result:
check_cache_or_validate_value.invalidate(cache_value_prefix,
value)
return validation_result
def validate_with_client(self, client, resource_id):
if self.resource_client_name and self.resource_getter_name:
getattr(client.client_plugin(self.resource_client_name),
self.resource_getter_name)(resource_id)
else:
raise exception.InvalidSchemaError(
message=_('Client name and resource getter name must be '
'specified.'))