Shortcuts on this page
r m x toggle line displays
j k next/prev highlighted chunk
0 (zero) top of page
1 (one) first highlighted chunk
1from __future__ import annotations
3from collections import defaultdict
4from collections.abc import Iterable
5from copy import copy
6from functools import lru_cache, partial
7from typing import TYPE_CHECKING, Any
9from pydantic_core import CoreSchema, PydanticCustomError, ValidationError, to_jsonable_python
10from pydantic_core import core_schema as cs
12from ._fields import PydanticMetadata
13from ._import_utils import import_cached_field_info
15if TYPE_CHECKING:
16 pass
18STRICT = {'strict'}
19FAIL_FAST = {'fail_fast'}
20LENGTH_CONSTRAINTS = {'min_length', 'max_length'}
21INEQUALITY = {'le', 'ge', 'lt', 'gt'}
22NUMERIC_CONSTRAINTS = {'multiple_of', *INEQUALITY}
23ALLOW_INF_NAN = {'allow_inf_nan'}
25STR_CONSTRAINTS = {
26 *LENGTH_CONSTRAINTS,
27 *STRICT,
28 'strip_whitespace',
29 'to_lower',
30 'to_upper',
31 'pattern',
32 'coerce_numbers_to_str',
33}
34BYTES_CONSTRAINTS = {*LENGTH_CONSTRAINTS, *STRICT}
36LIST_CONSTRAINTS = {*LENGTH_CONSTRAINTS, *STRICT, *FAIL_FAST}
37TUPLE_CONSTRAINTS = {*LENGTH_CONSTRAINTS, *STRICT, *FAIL_FAST}
38SET_CONSTRAINTS = {*LENGTH_CONSTRAINTS, *STRICT, *FAIL_FAST}
39DICT_CONSTRAINTS = {*LENGTH_CONSTRAINTS, *STRICT}
40GENERATOR_CONSTRAINTS = {*LENGTH_CONSTRAINTS, *STRICT}
41SEQUENCE_CONSTRAINTS = {*LENGTH_CONSTRAINTS, *FAIL_FAST}
43FLOAT_CONSTRAINTS = {*NUMERIC_CONSTRAINTS, *ALLOW_INF_NAN, *STRICT}
44DECIMAL_CONSTRAINTS = {'max_digits', 'decimal_places', *FLOAT_CONSTRAINTS}
45INT_CONSTRAINTS = {*NUMERIC_CONSTRAINTS, *ALLOW_INF_NAN, *STRICT}
46BOOL_CONSTRAINTS = STRICT
47UUID_CONSTRAINTS = STRICT
49DATE_TIME_CONSTRAINTS = {*NUMERIC_CONSTRAINTS, *STRICT}
50TIMEDELTA_CONSTRAINTS = {*NUMERIC_CONSTRAINTS, *STRICT}
51TIME_CONSTRAINTS = {*NUMERIC_CONSTRAINTS, *STRICT}
52LAX_OR_STRICT_CONSTRAINTS = STRICT
53ENUM_CONSTRAINTS = STRICT
54COMPLEX_CONSTRAINTS = STRICT
56UNION_CONSTRAINTS = {'union_mode'}
57URL_CONSTRAINTS = {
58 'max_length',
59 'allowed_schemes',
60 'host_required',
61 'default_host',
62 'default_port',
63 'default_path',
64}
66TEXT_SCHEMA_TYPES = ('str', 'bytes', 'url', 'multi-host-url')
67SEQUENCE_SCHEMA_TYPES = ('list', 'tuple', 'set', 'frozenset', 'generator', *TEXT_SCHEMA_TYPES)
68NUMERIC_SCHEMA_TYPES = ('float', 'int', 'date', 'time', 'timedelta', 'datetime')
70CONSTRAINTS_TO_ALLOWED_SCHEMAS: dict[str, set[str]] = defaultdict(set)
72constraint_schema_pairings: list[tuple[set[str], tuple[str, ...]]] = [
73 (STR_CONSTRAINTS, TEXT_SCHEMA_TYPES),
74 (BYTES_CONSTRAINTS, ('bytes',)),
75 (LIST_CONSTRAINTS, ('list',)),
76 (TUPLE_CONSTRAINTS, ('tuple',)),
77 (SET_CONSTRAINTS, ('set', 'frozenset')),
78 (DICT_CONSTRAINTS, ('dict',)),
79 (GENERATOR_CONSTRAINTS, ('generator',)),
80 (FLOAT_CONSTRAINTS, ('float',)),
81 (INT_CONSTRAINTS, ('int',)),
82 (DATE_TIME_CONSTRAINTS, ('date', 'time', 'datetime', 'timedelta')),
83 # TODO: this is a bit redundant, we could probably avoid some of these
84 (STRICT, (*TEXT_SCHEMA_TYPES, *SEQUENCE_SCHEMA_TYPES, *NUMERIC_SCHEMA_TYPES, 'typed-dict', 'model')),
85 (UNION_CONSTRAINTS, ('union',)),
86 (URL_CONSTRAINTS, ('url', 'multi-host-url')),
87 (BOOL_CONSTRAINTS, ('bool',)),
88 (UUID_CONSTRAINTS, ('uuid',)),
89 (LAX_OR_STRICT_CONSTRAINTS, ('lax-or-strict',)),
90 (ENUM_CONSTRAINTS, ('enum',)),
91 (DECIMAL_CONSTRAINTS, ('decimal',)),
92 (COMPLEX_CONSTRAINTS, ('complex',)),
93]
95for constraints, schemas in constraint_schema_pairings:
96 for c in constraints:
97 CONSTRAINTS_TO_ALLOWED_SCHEMAS[c].update(schemas)
100def as_jsonable_value(v: Any) -> Any:
101 if type(v) not in (int, str, float, bytes, bool, type(None)):
102 return to_jsonable_python(v)
103 return v
106def expand_grouped_metadata(annotations: Iterable[Any]) -> Iterable[Any]:
107 """Expand the annotations.
109 Args:
110 annotations: An iterable of annotations.
112 Returns:
113 An iterable of expanded annotations.
115 Example:
116 ```python
117 from annotated_types import Ge, Len
119 from pydantic._internal._known_annotated_metadata import expand_grouped_metadata
121 print(list(expand_grouped_metadata([Ge(4), Len(5)])))
122 #> [Ge(ge=4), MinLen(min_length=5)]
123 ```
124 """
125 import annotated_types as at
127 FieldInfo = import_cached_field_info()
129 for annotation in annotations:
130 if isinstance(annotation, at.GroupedMetadata):
131 yield from annotation
132 elif isinstance(annotation, FieldInfo):
133 yield from annotation.metadata
134 # this is a bit problematic in that it results in duplicate metadata
135 # all of our "consumers" can handle it, but it is not ideal
136 # we probably should split up FieldInfo into:
137 # - annotated types metadata
138 # - individual metadata known only to Pydantic
139 annotation = copy(annotation)
140 annotation.metadata = []
141 yield annotation
142 else:
143 yield annotation
146@lru_cache
147def _get_at_to_constraint_map() -> dict[type, str]:
148 """Return a mapping of annotated types to constraints.
150 Normally, we would define a mapping like this in the module scope, but we can't do that
151 because we don't permit module level imports of `annotated_types`, in an attempt to speed up
152 the import time of `pydantic`. We still only want to have this dictionary defined in one place,
153 so we use this function to cache the result.
154 """
155 import annotated_types as at
157 return {
158 at.Gt: 'gt',
159 at.Ge: 'ge',
160 at.Lt: 'lt',
161 at.Le: 'le',
162 at.MultipleOf: 'multiple_of',
163 at.MinLen: 'min_length',
164 at.MaxLen: 'max_length',
165 }
168def apply_known_metadata(annotation: Any, schema: CoreSchema) -> CoreSchema | None: # noqa: C901
169 """Apply `annotation` to `schema` if it is an annotation we know about (Gt, Le, etc.).
170 Otherwise return `None`.
172 This does not handle all known annotations. If / when it does, it can always
173 return a CoreSchema and return the unmodified schema if the annotation should be ignored.
175 Assumes that GroupedMetadata has already been expanded via `expand_grouped_metadata`.
177 Args:
178 annotation: The annotation.
179 schema: The schema.
181 Returns:
182 An updated schema with annotation if it is an annotation we know about, `None` otherwise.
184 Raises:
185 RuntimeError: If a constraint can't be applied to a specific schema type.
186 ValueError: If an unknown constraint is encountered.
187 """
188 import annotated_types as at
190 from ._validators import NUMERIC_VALIDATOR_LOOKUP, forbid_inf_nan_check
192 schema = schema.copy()
193 schema_update, other_metadata = collect_known_metadata([annotation])
194 schema_type = schema['type']
196 chain_schema_constraints: set[str] = {
197 'pattern',
198 'strip_whitespace',
199 'to_lower',
200 'to_upper',
201 'coerce_numbers_to_str',
202 }
203 chain_schema_steps: list[CoreSchema] = []
205 for constraint, value in schema_update.items():
206 if constraint not in CONSTRAINTS_TO_ALLOWED_SCHEMAS:
207 raise ValueError(f'Unknown constraint {constraint}')
208 allowed_schemas = CONSTRAINTS_TO_ALLOWED_SCHEMAS[constraint]
210 # if it becomes necessary to handle more than one constraint
211 # in this recursive case with function-after or function-wrap, we should refactor
212 # this is a bit challenging because we sometimes want to apply constraints to the inner schema,
213 # whereas other times we want to wrap the existing schema with a new one that enforces a new constraint.
214 if schema_type in {'function-before', 'function-wrap', 'function-after'} and constraint == 'strict':
215 schema['schema'] = apply_known_metadata(annotation, schema['schema']) # type: ignore # schema is function schema
216 return schema
218 # if we're allowed to apply constraint directly to the schema, like le to int, do that
219 if schema_type in allowed_schemas:
220 if constraint == 'union_mode' and schema_type == 'union':
221 schema['mode'] = value # type: ignore # schema is UnionSchema
222 else:
223 schema[constraint] = value
224 continue
226 # else, apply a function after validator to the schema to enforce the corresponding constraint
227 if constraint in chain_schema_constraints:
229 def _apply_constraint_with_incompatibility_info(
230 value: Any, handler: cs.ValidatorFunctionWrapHandler
231 ) -> Any:
232 try:
233 x = handler(value)
234 except ValidationError as ve:
235 # if the error is about the type, it's likely that the constraint is incompatible the type of the field
236 # for example, the following invalid schema wouldn't be caught during schema build, but rather at this point
237 # with a cryptic 'string_type' error coming from the string validator,
238 # that we'd rather express as a constraint incompatibility error (TypeError)
239 # Annotated[list[int], Field(pattern='abc')]
240 if 'type' in ve.errors()[0]['type']:
241 raise TypeError(
242 f"Unable to apply constraint '{constraint}' to supplied value {value} for schema of type '{schema_type}'" # noqa: B023
243 )
244 raise ve
245 return x
247 chain_schema_steps.append(
248 cs.no_info_wrap_validator_function(
249 _apply_constraint_with_incompatibility_info, cs.str_schema(**{constraint: value})
250 )
251 )
252 elif constraint in NUMERIC_VALIDATOR_LOOKUP:
253 if constraint in LENGTH_CONSTRAINTS:
254 inner_schema = schema
255 while inner_schema['type'] in {'function-before', 'function-wrap', 'function-after'}:
256 inner_schema = inner_schema['schema'] # type: ignore
257 inner_schema_type = inner_schema['type']
258 if inner_schema_type == 'list' or (
259 inner_schema_type == 'json-or-python' and inner_schema['json_schema']['type'] == 'list' # type: ignore
260 ):
261 js_constraint_key = 'minItems' if constraint == 'min_length' else 'maxItems'
262 else:
263 js_constraint_key = 'minLength' if constraint == 'min_length' else 'maxLength'
264 else:
265 js_constraint_key = constraint
267 schema = cs.no_info_after_validator_function(
268 partial(NUMERIC_VALIDATOR_LOOKUP[constraint], **{constraint: value}), schema
269 )
270 metadata = schema.get('metadata', {})
271 if (existing_json_schema_updates := metadata.get('pydantic_js_updates')) is not None:
272 metadata['pydantic_js_updates'] = {
273 **existing_json_schema_updates,
274 **{js_constraint_key: as_jsonable_value(value)},
275 }
276 else:
277 metadata['pydantic_js_updates'] = {js_constraint_key: as_jsonable_value(value)}
278 schema['metadata'] = metadata
279 elif constraint == 'allow_inf_nan' and value is False:
280 schema = cs.no_info_after_validator_function(
281 forbid_inf_nan_check,
282 schema,
283 )
284 else:
285 # It's rare that we'd get here, but it's possible if we add a new constraint and forget to handle it
286 # Most constraint errors are caught at runtime during attempted application
287 raise RuntimeError(f"Unable to apply constraint '{constraint}' to schema of type '{schema_type}'")
289 for annotation in other_metadata:
290 if (annotation_type := type(annotation)) in (at_to_constraint_map := _get_at_to_constraint_map()):
291 constraint = at_to_constraint_map[annotation_type]
292 validator = NUMERIC_VALIDATOR_LOOKUP.get(constraint)
293 if validator is None:
294 raise ValueError(f'Unknown constraint {constraint}')
295 schema = cs.no_info_after_validator_function(
296 partial(validator, {constraint: getattr(annotation, constraint)}), schema
297 )
298 continue
299 elif isinstance(annotation, (at.Predicate, at.Not)):
300 predicate_name = f'{annotation.func.__qualname__!r} ' if hasattr(annotation.func, '__qualname__') else ''
302 # Note: B023 is ignored because even though we iterate over `other_metadata`, it is guaranteed
303 # to be of length 1. `apply_known_metadata()` is called from `GenerateSchema`, where annotations
304 # were already expanded via `expand_grouped_metadata()`. Confusing, but this falls into the annotations
305 # refactor.
306 if isinstance(annotation, at.Predicate):
308 def val_func(v: Any) -> Any:
309 predicate_satisfied = annotation.func(v) # noqa: B023
310 if not predicate_satisfied:
311 raise PydanticCustomError(
312 'predicate_failed',
313 f'Predicate {predicate_name}failed', # pyright: ignore[reportArgumentType] # noqa: B023
314 )
315 return v
317 else:
319 def val_func(v: Any) -> Any:
320 predicate_satisfied = annotation.func(v) # noqa: B023
321 if predicate_satisfied:
322 raise PydanticCustomError(
323 'not_operation_failed',
324 f'Not of {predicate_name}failed', # pyright: ignore[reportArgumentType] # noqa: B023
325 )
326 return v
328 schema = cs.no_info_after_validator_function(val_func, schema)
329 else:
330 # ignore any other unknown metadata
331 return None
333 if chain_schema_steps:
334 chain_schema_steps = [schema] + chain_schema_steps
335 return cs.chain_schema(chain_schema_steps)
337 return schema
340def collect_known_metadata(annotations: Iterable[Any]) -> tuple[dict[str, Any], list[Any]]:
341 """Split `annotations` into known metadata and unknown annotations.
343 Args:
344 annotations: An iterable of annotations.
346 Returns:
347 A tuple contains a dict of known metadata and a list of unknown annotations.
349 Example:
350 ```python
351 from annotated_types import Gt, Len
353 from pydantic._internal._known_annotated_metadata import collect_known_metadata
355 print(collect_known_metadata([Gt(1), Len(42), ...]))
356 #> ({'gt': 1, 'min_length': 42}, [Ellipsis])
357 ```
358 """
359 annotations = expand_grouped_metadata(annotations)
361 res: dict[str, Any] = {}
362 remaining: list[Any] = []
364 for annotation in annotations:
365 # isinstance(annotation, PydanticMetadata) also covers ._fields:_PydanticGeneralMetadata
366 if isinstance(annotation, PydanticMetadata):
367 res.update(annotation.__dict__)
368 # we don't use dataclasses.asdict because that recursively calls asdict on the field values
369 elif (annotation_type := type(annotation)) in (at_to_constraint_map := _get_at_to_constraint_map()):
370 constraint = at_to_constraint_map[annotation_type]
371 res[constraint] = getattr(annotation, constraint)
372 elif isinstance(annotation, type) and issubclass(annotation, PydanticMetadata):
373 # also support PydanticMetadata classes being used without initialisation,
374 # e.g. `Annotated[int, Strict]` as well as `Annotated[int, Strict()]`
375 res.update({k: v for k, v in vars(annotation).items() if not k.startswith('_')})
376 else:
377 remaining.append(annotation)
378 # Nones can sneak in but pydantic-core will reject them
379 # it'd be nice to clean things up so we don't put in None (we probably don't _need_ to, it was just easier)
380 # but this is simple enough to kick that can down the road
381 res = {k: v for k, v in res.items() if v is not None}
382 return res, remaining
385def check_metadata(metadata: dict[str, Any], allowed: Iterable[str], source_type: Any) -> None:
386 """A small utility function to validate that the given metadata can be applied to the target.
387 More than saving lines of code, this gives us a consistent error message for all of our internal implementations.
389 Args:
390 metadata: A dict of metadata.
391 allowed: An iterable of allowed metadata.
392 source_type: The source type.
394 Raises:
395 TypeError: If there is metadatas that can't be applied on source type.
396 """
397 unknown = metadata.keys() - set(allowed)
398 if unknown:
399 raise TypeError(
400 f'The following constraints cannot be applied to {source_type!r}: {", ".join([f"{k!r}" for k in unknown])}'
401 )