Sindbad~EG File Manager
"""schema is a library for validating Python data structures, such as those
obtained from config-files, forms, external services or command-line
parsing, converted from JSON/YAML (or something else) to Python data-types."""
import re
try:
from contextlib import ExitStack
except ImportError:
from contextlib2 import ExitStack
__version__ = "0.7.1"
__all__ = [
"Schema",
"And",
"Or",
"Regex",
"Optional",
"Use",
"Forbidden",
"Const",
"Literal",
"SchemaError",
"SchemaWrongKeyError",
"SchemaMissingKeyError",
"SchemaForbiddenKeyError",
"SchemaUnexpectedTypeError",
"SchemaOnlyOneAllowedError",
]
class SchemaError(Exception):
"""Error during Schema validation."""
def __init__(self, autos, errors=None):
self.autos = autos if type(autos) is list else [autos]
self.errors = errors if type(errors) is list else [errors]
Exception.__init__(self, self.code)
@property
def code(self):
"""
Removes duplicates values in auto and error list.
parameters.
"""
def uniq(seq):
"""
Utility function that removes duplicate.
"""
seen = set()
seen_add = seen.add
# This way removes duplicates while preserving the order.
return [x for x in seq if x not in seen and not seen_add(x)]
data_set = uniq(i for i in self.autos if i is not None)
error_list = uniq(i for i in self.errors if i is not None)
if error_list:
return "\n".join(error_list)
return "\n".join(data_set)
class SchemaWrongKeyError(SchemaError):
"""Error Should be raised when an unexpected key is detected within the
data set being."""
pass
class SchemaMissingKeyError(SchemaError):
"""Error should be raised when a mandatory key is not found within the
data set being validated"""
pass
class SchemaOnlyOneAllowedError(SchemaError):
"""Error should be raised when an only_one Or key has multiple matching candidates"""
pass
class SchemaForbiddenKeyError(SchemaError):
"""Error should be raised when a forbidden key is found within the
data set being validated, and its value matches the value that was specified"""
pass
class SchemaUnexpectedTypeError(SchemaError):
"""Error should be raised when a type mismatch is detected within the
data set being validated."""
pass
class And(object):
"""
Utility function to combine validation directives in AND Boolean fashion.
"""
def __init__(self, *args, **kw):
self._args = args
if not set(kw).issubset({"error", "schema", "ignore_extra_keys"}):
diff = {"error", "schema", "ignore_extra_keys"}.difference(kw)
raise TypeError("Unknown keyword arguments %r" % list(diff))
self._error = kw.get("error")
self._ignore_extra_keys = kw.get("ignore_extra_keys", False)
# You can pass your inherited Schema class.
self._schema = kw.get("schema", Schema)
def __repr__(self):
return "%s(%s)" % (self.__class__.__name__, ", ".join(repr(a) for a in self._args))
@property
def args(self):
"""The provided parameters"""
return self._args
def validate(self, data):
"""
Validate data using defined sub schema/expressions ensuring all
values are valid.
:param data: to be validated with sub defined schemas.
:return: returns validated data
"""
for s in [self._schema(s, error=self._error, ignore_extra_keys=self._ignore_extra_keys) for s in self._args]:
data = s.validate(data)
return data
class Or(And):
"""Utility function to combine validation directives in a OR Boolean
fashion."""
def __init__(self, *args, **kwargs):
self.only_one = kwargs.pop("only_one", False)
self.match_count = 0
super(Or, self).__init__(*args, **kwargs)
def reset(self):
failed = self.match_count > 1 and self.only_one
self.match_count = 0
if failed:
raise SchemaOnlyOneAllowedError(["There are multiple keys present " + "from the %r condition" % self])
def validate(self, data):
"""
Validate data using sub defined schema/expressions ensuring at least
one value is valid.
:param data: data to be validated by provided schema.
:return: return validated data if not validation
"""
autos, errors = [], []
for s in [self._schema(s, error=self._error, ignore_extra_keys=self._ignore_extra_keys) for s in self._args]:
try:
validation = s.validate(data)
self.match_count += 1
if self.match_count > 1 and self.only_one:
break
return validation
except SchemaError as _x:
autos, errors = _x.autos, _x.errors
raise SchemaError(
["%r did not validate %r" % (self, data)] + autos,
[self._error.format(data) if self._error else None] + errors,
)
class Regex(object):
"""
Enables schema.py to validate string using regular expressions.
"""
# Map all flags bits to a more readable description
NAMES = [
"re.ASCII",
"re.DEBUG",
"re.VERBOSE",
"re.UNICODE",
"re.DOTALL",
"re.MULTILINE",
"re.LOCALE",
"re.IGNORECASE",
"re.TEMPLATE",
]
def __init__(self, pattern_str, flags=0, error=None):
self._pattern_str = pattern_str
flags_list = [Regex.NAMES[i] for i, f in enumerate("{0:09b}".format(flags)) if f != "0"] # Name for each bit
if flags_list:
self._flags_names = ", flags=" + "|".join(flags_list)
else:
self._flags_names = ""
self._pattern = re.compile(pattern_str, flags=flags)
self._error = error
def __repr__(self):
return "%s(%r%s)" % (self.__class__.__name__, self._pattern_str, self._flags_names)
@property
def pattern_str(self):
"""The pattern for the represented regular expression"""
return self._pattern_str
def validate(self, data):
"""
Validated data using defined regex.
:param data: data to be validated
:return: return validated data.
"""
e = self._error
try:
if self._pattern.search(data):
return data
else:
raise SchemaError("%r does not match %r" % (self, data), e)
except TypeError:
raise SchemaError("%r is not string nor buffer" % data, e)
class Use(object):
"""
For more general use cases, you can use the Use class to transform
the data while it is being validate.
"""
def __init__(self, callable_, error=None):
if not callable(callable_):
raise TypeError("Expected a callable, not %r" % callable_)
self._callable = callable_
self._error = error
def __repr__(self):
return "%s(%r)" % (self.__class__.__name__, self._callable)
def validate(self, data):
try:
return self._callable(data)
except SchemaError as x:
raise SchemaError([None] + x.autos, [self._error.format(data) if self._error else None] + x.errors)
except BaseException as x:
f = _callable_str(self._callable)
raise SchemaError("%s(%r) raised %r" % (f, data, x), self._error.format(data) if self._error else None)
COMPARABLE, CALLABLE, VALIDATOR, TYPE, DICT, ITERABLE = range(6)
def _priority(s):
"""Return priority for a given object."""
if type(s) in (list, tuple, set, frozenset):
return ITERABLE
if type(s) is dict:
return DICT
if issubclass(type(s), type):
return TYPE
if isinstance(s, Literal):
return COMPARABLE
if hasattr(s, "validate"):
return VALIDATOR
if callable(s):
return CALLABLE
else:
return COMPARABLE
class Schema(object):
"""
Entry point of the library, use this class to instantiate validation
schema for the data that will be validated.
"""
def __init__(self, schema, error=None, ignore_extra_keys=False, name=None, description=None, as_reference=False):
self._schema = schema
self._error = error
self._ignore_extra_keys = ignore_extra_keys
self._name = name
self._description = description
# Ask json_schema to create a definition for this schema and use it as part of another
self.as_reference = as_reference
if as_reference and name is None:
raise ValueError("Schema used as reference should have a name")
def __repr__(self):
return "%s(%r)" % (self.__class__.__name__, self._schema)
@property
def schema(self):
return self._schema
@property
def description(self):
return self._description
@property
def name(self):
return self._name
@property
def ignore_extra_keys(self):
return self._ignore_extra_keys
@staticmethod
def _dict_key_priority(s):
"""Return priority for a given key object."""
if isinstance(s, Hook):
return _priority(s._schema) - 0.5
if isinstance(s, Optional):
return _priority(s._schema) + 0.5
return _priority(s)
@staticmethod
def _is_optional_type(s):
"""Return True if the given key is optional (does not have to be found)"""
return any(isinstance(s, optional_type) for optional_type in [Optional, Hook])
def is_valid(self, data):
"""Return whether the given data has passed all the validations
that were specified in the given schema.
"""
try:
self.validate(data)
except SchemaError:
return False
else:
return True
def _prepend_schema_name(self, message):
"""
If a custom schema name has been defined, prepends it to the error
message that gets raised when a schema error occurs.
"""
if self._name:
message = "{0!r} {1!s}".format(self._name, message)
return message
def validate(self, data):
Schema = self.__class__
s = self._schema
e = self._error
i = self._ignore_extra_keys
if isinstance(s, Literal):
s = s.schema
flavor = _priority(s)
if flavor == ITERABLE:
data = Schema(type(s), error=e).validate(data)
o = Or(*s, error=e, schema=Schema, ignore_extra_keys=i)
return type(data)(o.validate(d) for d in data)
if flavor == DICT:
exitstack = ExitStack()
data = Schema(dict, error=e).validate(data)
new = type(data)() # new - is a dict of the validated values
coverage = set() # matched schema keys
# for each key and value find a schema entry matching them, if any
sorted_skeys = sorted(s, key=self._dict_key_priority)
for skey in sorted_skeys:
if hasattr(skey, "reset"):
exitstack.callback(skey.reset)
with exitstack:
# Evaluate dictionaries last
data_items = sorted(data.items(), key=lambda value: isinstance(value[1], dict))
for key, value in data_items:
for skey in sorted_skeys:
svalue = s[skey]
try:
nkey = Schema(skey, error=e).validate(key)
except SchemaError:
pass
else:
if isinstance(skey, Hook):
# As the content of the value makes little sense for
# keys with a hook, we reverse its meaning:
# we will only call the handler if the value does match
# In the case of the forbidden key hook,
# we will raise the SchemaErrorForbiddenKey exception
# on match, allowing for excluding a key only if its
# value has a certain type, and allowing Forbidden to
# work well in combination with Optional.
try:
nvalue = Schema(svalue, error=e).validate(value)
except SchemaError:
continue
skey.handler(nkey, data, e)
else:
try:
nvalue = Schema(svalue, error=e, ignore_extra_keys=i).validate(value)
except SchemaError as x:
k = "Key '%s' error:" % nkey
message = self._prepend_schema_name(k)
raise SchemaError([message] + x.autos, [e] + x.errors)
else:
new[nkey] = nvalue
coverage.add(skey)
break
required = set(k for k in s if not self._is_optional_type(k))
if not required.issubset(coverage):
missing_keys = required - coverage
s_missing_keys = ", ".join(repr(k) for k in sorted(missing_keys, key=repr))
message = "Missing key%s: %s" % (_plural_s(missing_keys), s_missing_keys)
message = self._prepend_schema_name(message)
raise SchemaMissingKeyError(message, e)
if not self._ignore_extra_keys and (len(new) != len(data)):
wrong_keys = set(data.keys()) - set(new.keys())
s_wrong_keys = ", ".join(repr(k) for k in sorted(wrong_keys, key=repr))
message = "Wrong key%s %s in %r" % (_plural_s(wrong_keys), s_wrong_keys, data)
message = self._prepend_schema_name(message)
raise SchemaWrongKeyError(message, e.format(data) if e else None)
# Apply default-having optionals that haven't been used:
defaults = set(k for k in s if type(k) is Optional and hasattr(k, "default")) - coverage
for default in defaults:
new[default.key] = default.default() if callable(default.default) else default.default
return new
if flavor == TYPE:
if isinstance(data, s) and not (isinstance(data, bool) and s == int):
return data
else:
message = "%r should be instance of %r" % (data, s.__name__)
message = self._prepend_schema_name(message)
raise SchemaUnexpectedTypeError(message, e.format(data) if e else None)
if flavor == VALIDATOR:
try:
return s.validate(data)
except SchemaError as x:
raise SchemaError([None] + x.autos, [e] + x.errors)
except BaseException as x:
message = "%r.validate(%r) raised %r" % (s, data, x)
message = self._prepend_schema_name(message)
raise SchemaError(message, self._error.format(data) if self._error else None)
if flavor == CALLABLE:
f = _callable_str(s)
try:
if s(data):
return data
except SchemaError as x:
raise SchemaError([None] + x.autos, [e] + x.errors)
except BaseException as x:
message = "%s(%r) raised %r" % (f, data, x)
message = self._prepend_schema_name(message)
raise SchemaError(message, self._error.format(data) if self._error else None)
message = "%s(%r) should evaluate to True" % (f, data)
message = self._prepend_schema_name(message)
raise SchemaError(message, e)
if s == data:
return data
else:
message = "%r does not match %r" % (s, data)
message = self._prepend_schema_name(message)
raise SchemaError(message, e.format(data) if e else None)
def json_schema(self, schema_id, use_refs=False):
"""Generate a draft-07 JSON schema dict representing the Schema.
This method can only be called when the Schema's value is a dict.
This method must be called with a schema_id.
:param schema_id: The value of the $id on the main schema
:param use_refs: Enable reusing object references in the resulting JSON schema.
Schemas with references are harder to read by humans, but are a lot smaller when there
is a lot of reuse
"""
seen = dict() # For use_refs
definitions_by_name = {}
def _json_schema(schema, is_main_schema=True, description=None, allow_reference=True):
Schema = self.__class__
def _create_or_use_ref(return_dict):
"""If not already seen, return the provided part of the schema unchanged.
If already seen, give an id to the already seen dict and return a reference to the previous part
of the schema instead.
"""
if not use_refs or is_main_schema:
return return_schema
hashed = hash(repr(sorted(return_dict.items())))
if hashed not in seen:
seen[hashed] = return_dict
return return_dict
else:
id_str = "#" + str(hashed)
seen[hashed]["$id"] = id_str
return {"$ref": id_str}
def _get_type_name(python_type):
"""Return the JSON schema name for a Python type"""
if python_type == str:
return "string"
elif python_type == int:
return "integer"
elif python_type == float:
return "number"
elif python_type == bool:
return "boolean"
elif python_type == list:
return "array"
elif python_type == dict:
return "object"
return "string"
def _to_json_type(value):
"""Attempt to convert a constant value (for "const" and "default") to a JSON serializable value"""
if value is None or type(value) in (str, int, float, bool, list, dict):
return value
if type(value) in (tuple, set, frozenset):
return list(value)
if isinstance(value, Literal):
return value.schema
return str(value)
def _to_schema(s, ignore_extra_keys):
if not isinstance(s, Schema):
return Schema(s, ignore_extra_keys=ignore_extra_keys)
return s
s = schema.schema
i = schema.ignore_extra_keys
flavor = _priority(s)
return_schema = {}
is_a_ref = allow_reference and schema.as_reference
if schema.description and not is_a_ref:
return_schema["description"] = schema.description
if description and not is_a_ref:
return_schema["description"] = description
if flavor != DICT and is_main_schema:
raise ValueError("The main schema must be a dict.")
if flavor == TYPE:
# Handle type
return_schema["type"] = _get_type_name(s)
elif flavor == ITERABLE:
# Handle arrays or dict schema
return_schema["type"] = "array"
if len(s) == 1:
return_schema["items"] = _json_schema(_to_schema(s[0], i), is_main_schema=False)
elif len(s) > 1:
return_schema["items"] = _json_schema(Schema(Or(*s)), is_main_schema=False)
elif isinstance(s, Or):
# Handle Or values
# Check if we can use an enum
if all(priority == COMPARABLE for priority in [_priority(value) for value in s.args]):
or_values = [str(s) if isinstance(s, Literal) else s for s in s.args]
# All values are simple, can use enum or const
if len(or_values) == 1:
return_schema["const"] = _to_json_type(or_values[0])
return return_schema
return_schema["enum"] = or_values
else:
# No enum, let's go with recursive calls
any_of_values = []
for or_key in s.args:
new_value = _json_schema(_to_schema(or_key, i), is_main_schema=False)
if new_value != {} and new_value not in any_of_values:
any_of_values.append(new_value)
if len(any_of_values) == 1:
# Only one representable condition remains, do not put under oneOf
return_schema.update(any_of_values[0])
else:
return_schema["anyOf"] = any_of_values
elif isinstance(s, And):
# Handle And values
all_of_values = []
for and_key in s.args:
new_value = _json_schema(_to_schema(and_key, i), is_main_schema=False)
if new_value != {} and new_value not in all_of_values:
all_of_values.append(new_value)
if len(all_of_values) == 1:
# Only one representable condition remains, do not put under allOf
return_schema.update(all_of_values[0])
else:
return_schema["allOf"] = all_of_values
elif flavor == COMPARABLE:
return_schema["const"] = _to_json_type(s)
elif flavor == VALIDATOR and type(s) == Regex:
return_schema["type"] = "string"
return_schema["pattern"] = s.pattern_str
else:
if flavor != DICT:
# If not handled, do not check
return return_schema
# Schema is a dict
# Check if we have to create a common definition and use as reference
if is_a_ref:
# Generate sub schema if not already done
if schema.name not in definitions_by_name:
definitions_by_name[schema.name] = {} # Avoid infinite loop
definitions_by_name[schema.name] = _json_schema(
schema, is_main_schema=False, allow_reference=False
)
return_schema["$ref"] = "#/definitions/" + schema.name
else:
required_keys = []
expanded_schema = {}
for key in s:
if isinstance(key, Hook):
continue
def _get_key_description(key):
"""Get the description associated to a key (as specified in a Literal object). Return None if not a Literal"""
if isinstance(key, Optional):
return _get_key_description(key.schema)
if isinstance(key, Literal):
return key.description
return None
def _get_key_name(key):
"""Get the name of a key (as specified in a Literal object). Return the key unchanged if not a Literal"""
if isinstance(key, Optional):
return _get_key_name(key.schema)
if isinstance(key, Literal):
return key.schema
return key
sub_schema = _to_schema(s[key], ignore_extra_keys=i)
key_name = _get_key_name(key)
if isinstance(key_name, str):
if not isinstance(key, Optional):
required_keys.append(key_name)
expanded_schema[key_name] = _json_schema(
sub_schema, is_main_schema=False, description=_get_key_description(key)
)
if isinstance(key, Optional) and hasattr(key, "default"):
expanded_schema[key_name]["default"] = _to_json_type(key.default)
elif isinstance(key_name, Or):
# JSON schema does not support having a key named one name or another, so we just add both options
# This is less strict because we cannot enforce that one or the other is required
for or_key in key_name.args:
expanded_schema[_get_key_name(or_key)] = _json_schema(
sub_schema, is_main_schema=False, description=_get_key_description(or_key)
)
return_schema.update(
{
"type": "object",
"properties": expanded_schema,
"required": required_keys,
"additionalProperties": i,
}
)
if is_main_schema:
return_schema.update({"$id": schema_id, "$schema": "http://json-schema.org/draft-07/schema#"})
if self._name:
return_schema["title"] = self._name
if definitions_by_name:
return_schema["definitions"] = {}
for definition_name, definition in definitions_by_name.items():
return_schema["definitions"][definition_name] = definition
return _create_or_use_ref(return_schema)
return _json_schema(self, True)
class Optional(Schema):
"""Marker for an optional part of the validation Schema."""
_MARKER = object()
def __init__(self, *args, **kwargs):
default = kwargs.pop("default", self._MARKER)
super(Optional, self).__init__(*args, **kwargs)
if default is not self._MARKER:
# See if I can come up with a static key to use for myself:
if _priority(self._schema) != COMPARABLE:
raise TypeError(
"Optional keys with defaults must have simple, "
"predictable values, like literal strings or ints. "
'"%r" is too complex.' % (self._schema,)
)
self.default = default
self.key = str(self._schema)
def __hash__(self):
return hash(self._schema)
def __eq__(self, other):
return (
self.__class__ is other.__class__
and getattr(self, "default", self._MARKER) == getattr(other, "default", self._MARKER)
and self._schema == other._schema
)
def reset(self):
if hasattr(self._schema, "reset"):
self._schema.reset()
class Hook(Schema):
def __init__(self, *args, **kwargs):
self.handler = kwargs.pop("handler", lambda *args: None)
super(Hook, self).__init__(*args, **kwargs)
self.key = self._schema
class Forbidden(Hook):
def __init__(self, *args, **kwargs):
kwargs["handler"] = self._default_function
super(Forbidden, self).__init__(*args, **kwargs)
@staticmethod
def _default_function(nkey, data, error):
raise SchemaForbiddenKeyError("Forbidden key encountered: %r in %r" % (nkey, data), error)
class Literal(object):
def __init__(self, value, description=None):
self._schema = value
self._description = description
def __str__(self):
return self._schema
def __repr__(self):
return 'Literal("' + self.schema + '", description="' + (self.description or "") + '")'
@property
def description(self):
return self._description
@property
def schema(self):
return self._schema
class Const(Schema):
def validate(self, data):
super(Const, self).validate(data)
return data
def _callable_str(callable_):
if hasattr(callable_, "__name__"):
return callable_.__name__
return str(callable_)
def _plural_s(sized):
return "s" if len(sized) > 1 else ""
Sindbad File Manager Version 1.0, Coded By Sindbad EG ~ The Terrorists