1"""Bucket of reusable internal utilities.
2
3This should be reduced as much as possible with functions only used in one place, moved to that place.
4"""
5
6from __future__ import annotations as _annotations
7
8import dataclasses
9import keyword
10import sys
11import warnings
12import weakref
13from collections import OrderedDict, defaultdict, deque
14from collections.abc import Callable, Iterable, Mapping
15from collections.abc import Set as AbstractSet
16from copy import deepcopy
17from functools import cached_property
18from inspect import Parameter
19from itertools import zip_longest
20from types import BuiltinFunctionType, CodeType, FunctionType, GeneratorType, LambdaType, ModuleType
21from typing import TYPE_CHECKING, Any, Generic, TypeVar, overload
22
23from typing_extensions import TypeAlias, TypeGuard, deprecated
24
25from pydantic import PydanticDeprecatedSince211
26
27from . import _repr, _typing_extra
28from ._import_utils import import_cached_base_model
29
30if TYPE_CHECKING:
31 # TODO remove type error comments when we drop support for Python 3.9
32 MappingIntStrAny: TypeAlias = Mapping[int, Any] | Mapping[str, Any] # pyright: ignore[reportGeneralTypeIssues]
33 AbstractSetIntStr: TypeAlias = AbstractSet[int] | AbstractSet[str] # pyright: ignore[reportGeneralTypeIssues]
34 from ..main import BaseModel
35
36
37# these are types that are returned unchanged by deepcopy
38IMMUTABLE_NON_COLLECTIONS_TYPES: set[type[Any]] = {
39 int,
40 float,
41 complex,
42 str,
43 bool,
44 bytes,
45 type,
46 _typing_extra.NoneType,
47 FunctionType,
48 BuiltinFunctionType,
49 LambdaType,
50 weakref.ref,
51 CodeType,
52 # note: including ModuleType will differ from behaviour of deepcopy by not producing error.
53 # It might be not a good idea in general, but considering that this function used only internally
54 # against default values of fields, this will allow to actually have a field with module as default value
55 ModuleType,
56 NotImplemented.__class__,
57 Ellipsis.__class__,
58}
59
60# these are types that if empty, might be copied with simple copy() instead of deepcopy()
61BUILTIN_COLLECTIONS: set[type[Any]] = {
62 list,
63 set,
64 tuple,
65 frozenset,
66 dict,
67 OrderedDict,
68 defaultdict,
69 deque,
70}
71
72
73def can_be_positional(param: Parameter) -> bool:
74 """Return whether the parameter accepts a positional argument.
75
76 ```python {test="skip" lint="skip"}
77 def func(a, /, b, *, c):
78 pass
79
80 params = inspect.signature(func).parameters
81 can_be_positional(params['a'])
82 #> True
83 can_be_positional(params['b'])
84 #> True
85 can_be_positional(params['c'])
86 #> False
87 ```
88 """
89 return param.kind in (Parameter.POSITIONAL_ONLY, Parameter.POSITIONAL_OR_KEYWORD)
90
91
92def sequence_like(v: Any) -> bool:
93 return isinstance(v, (list, tuple, set, frozenset, GeneratorType, deque))
94
95
96def lenient_isinstance(o: Any, class_or_tuple: type[Any] | tuple[type[Any], ...] | None) -> bool: # pragma: no cover
97 try:
98 return isinstance(o, class_or_tuple) # type: ignore[arg-type]
99 except TypeError:
100 return False
101
102
103def lenient_issubclass(cls: Any, class_or_tuple: Any) -> bool: # pragma: no cover
104 try:
105 return isinstance(cls, type) and issubclass(cls, class_or_tuple)
106 except TypeError:
107 if isinstance(cls, _typing_extra.WithArgsTypes):
108 return False
109 raise # pragma: no cover
110
111
112def is_model_class(cls: Any) -> TypeGuard[type[BaseModel]]:
113 """Returns true if cls is a _proper_ subclass of BaseModel, and provides proper type-checking,
114 unlike raw calls to lenient_issubclass.
115 """
116 BaseModel = import_cached_base_model()
117
118 return lenient_issubclass(cls, BaseModel) and cls is not BaseModel
119
120
121def is_valid_identifier(identifier: str) -> bool:
122 """Checks that a string is a valid identifier and not a Python keyword.
123 :param identifier: The identifier to test.
124 :return: True if the identifier is valid.
125 """
126 return identifier.isidentifier() and not keyword.iskeyword(identifier)
127
128
129KeyType = TypeVar('KeyType')
130
131
132def deep_update(mapping: dict[KeyType, Any], *updating_mappings: dict[KeyType, Any]) -> dict[KeyType, Any]:
133 updated_mapping = mapping.copy()
134 for updating_mapping in updating_mappings:
135 for k, v in updating_mapping.items():
136 if k in updated_mapping and isinstance(updated_mapping[k], dict) and isinstance(v, dict):
137 updated_mapping[k] = deep_update(updated_mapping[k], v)
138 else:
139 updated_mapping[k] = v
140 return updated_mapping
141
142
143def update_not_none(mapping: dict[Any, Any], **update: Any) -> None:
144 mapping.update({k: v for k, v in update.items() if v is not None})
145
146
147T = TypeVar('T')
148
149
150def unique_list(
151 input_list: list[T] | tuple[T, ...],
152 *,
153 name_factory: Callable[[T], str] = str,
154) -> list[T]:
155 """Make a list unique while maintaining order.
156 We update the list if another one with the same name is set
157 (e.g. model validator overridden in subclass).
158 """
159 result: list[T] = []
160 result_names: list[str] = []
161 for v in input_list:
162 v_name = name_factory(v)
163 if v_name not in result_names:
164 result_names.append(v_name)
165 result.append(v)
166 else:
167 result[result_names.index(v_name)] = v
168
169 return result
170
171
172class ValueItems(_repr.Representation):
173 """Class for more convenient calculation of excluded or included fields on values."""
174
175 __slots__ = ('_items', '_type')
176
177 def __init__(self, value: Any, items: AbstractSetIntStr | MappingIntStrAny) -> None:
178 items = self._coerce_items(items)
179
180 if isinstance(value, (list, tuple)):
181 items = self._normalize_indexes(items, len(value)) # type: ignore
182
183 self._items: MappingIntStrAny = items # type: ignore
184
185 def is_excluded(self, item: Any) -> bool:
186 """Check if item is fully excluded.
187
188 :param item: key or index of a value
189 """
190 return self.is_true(self._items.get(item))
191
192 def is_included(self, item: Any) -> bool:
193 """Check if value is contained in self._items.
194
195 :param item: key or index of value
196 """
197 return item in self._items
198
199 def for_element(self, e: int | str) -> AbstractSetIntStr | MappingIntStrAny | None:
200 """:param e: key or index of element on value
201 :return: raw values for element if self._items is dict and contain needed element
202 """
203 item = self._items.get(e) # type: ignore
204 return item if not self.is_true(item) else None
205
206 def _normalize_indexes(self, items: MappingIntStrAny, v_length: int) -> dict[int | str, Any]:
207 """:param items: dict or set of indexes which will be normalized
208 :param v_length: length of sequence indexes of which will be
209
210 >>> self._normalize_indexes({0: True, -2: True, -1: True}, 4)
211 {0: True, 2: True, 3: True}
212 >>> self._normalize_indexes({'__all__': True}, 4)
213 {0: True, 1: True, 2: True, 3: True}
214 """
215 normalized_items: dict[int | str, Any] = {}
216 all_items = None
217 for i, v in items.items():
218 if not (isinstance(v, Mapping) or isinstance(v, AbstractSet) or self.is_true(v)):
219 raise TypeError(f'Unexpected type of exclude value for index "{i}" {v.__class__}')
220 if i == '__all__':
221 all_items = self._coerce_value(v)
222 continue
223 if not isinstance(i, int):
224 raise TypeError(
225 'Excluding fields from a sequence of sub-models or dicts must be performed index-wise: '
226 'expected integer keys or keyword "__all__"'
227 )
228 normalized_i = v_length + i if i < 0 else i
229 normalized_items[normalized_i] = self.merge(v, normalized_items.get(normalized_i))
230
231 if not all_items:
232 return normalized_items
233 if self.is_true(all_items):
234 for i in range(v_length):
235 normalized_items.setdefault(i, ...)
236 return normalized_items
237 for i in range(v_length):
238 normalized_item = normalized_items.setdefault(i, {})
239 if not self.is_true(normalized_item):
240 normalized_items[i] = self.merge(all_items, normalized_item)
241 return normalized_items
242
243 @classmethod
244 def merge(cls, base: Any, override: Any, intersect: bool = False) -> Any:
245 """Merge a `base` item with an `override` item.
246
247 Both `base` and `override` are converted to dictionaries if possible.
248 Sets are converted to dictionaries with the sets entries as keys and
249 Ellipsis as values.
250
251 Each key-value pair existing in `base` is merged with `override`,
252 while the rest of the key-value pairs are updated recursively with this function.
253
254 Merging takes place based on the "union" of keys if `intersect` is
255 set to `False` (default) and on the intersection of keys if
256 `intersect` is set to `True`.
257 """
258 override = cls._coerce_value(override)
259 base = cls._coerce_value(base)
260 if override is None:
261 return base
262 if cls.is_true(base) or base is None:
263 return override
264 if cls.is_true(override):
265 return base if intersect else override
266
267 # intersection or union of keys while preserving ordering:
268 if intersect:
269 merge_keys = [k for k in base if k in override] + [k for k in override if k in base]
270 else:
271 merge_keys = list(base) + [k for k in override if k not in base]
272
273 merged: dict[int | str, Any] = {}
274 for k in merge_keys:
275 merged_item = cls.merge(base.get(k), override.get(k), intersect=intersect)
276 if merged_item is not None:
277 merged[k] = merged_item
278
279 return merged
280
281 @staticmethod
282 def _coerce_items(items: AbstractSetIntStr | MappingIntStrAny) -> MappingIntStrAny:
283 if isinstance(items, Mapping):
284 pass
285 elif isinstance(items, AbstractSet):
286 items = dict.fromkeys(items, ...) # type: ignore
287 else:
288 class_name = getattr(items, '__class__', '???')
289 raise TypeError(f'Unexpected type of exclude value {class_name}')
290 return items # type: ignore
291
292 @classmethod
293 def _coerce_value(cls, value: Any) -> Any:
294 if value is None or cls.is_true(value):
295 return value
296 return cls._coerce_items(value)
297
298 @staticmethod
299 def is_true(v: Any) -> bool:
300 return v is True or v is ...
301
302 def __repr_args__(self) -> _repr.ReprArgs:
303 return [(None, self._items)]
304
305
306if TYPE_CHECKING:
307
308 def LazyClassAttribute(name: str, get_value: Callable[[], T]) -> T: ...
309
310else:
311
312 class LazyClassAttribute:
313 """A descriptor exposing an attribute only accessible on a class (hidden from instances).
314
315 The attribute is lazily computed and cached during the first access.
316 """
317
318 def __init__(self, name: str, get_value: Callable[[], Any]) -> None:
319 self.name = name
320 self.get_value = get_value
321
322 @cached_property
323 def value(self) -> Any:
324 return self.get_value()
325
326 def __get__(self, instance: Any, owner: type[Any]) -> None:
327 if instance is None:
328 return self.value
329 raise AttributeError(f'{self.name!r} attribute of {owner.__name__!r} is class-only')
330
331
332Obj = TypeVar('Obj')
333
334
335def smart_deepcopy(obj: Obj) -> Obj:
336 """Return type as is for immutable built-in types
337 Use obj.copy() for built-in empty collections
338 Use copy.deepcopy() for non-empty collections and unknown objects.
339 """
340 obj_type = obj.__class__
341 if obj_type in IMMUTABLE_NON_COLLECTIONS_TYPES:
342 return obj # fastest case: obj is immutable and not collection therefore will not be copied anyway
343 try:
344 if not obj and obj_type in BUILTIN_COLLECTIONS:
345 # faster way for empty collections, no need to copy its members
346 return obj if obj_type is tuple else obj.copy() # tuple doesn't have copy method # type: ignore
347 except (TypeError, ValueError, RuntimeError):
348 # do we really dare to catch ALL errors? Seems a bit risky
349 pass
350
351 return deepcopy(obj) # slowest way when we actually might need a deepcopy
352
353
354_SENTINEL = object()
355
356
357def all_identical(left: Iterable[Any], right: Iterable[Any]) -> bool:
358 """Check that the items of `left` are the same objects as those in `right`.
359
360 >>> a, b = object(), object()
361 >>> all_identical([a, b, a], [a, b, a])
362 True
363 >>> all_identical([a, b, [a]], [a, b, [a]]) # new list object, while "equal" is not "identical"
364 False
365 """
366 for left_item, right_item in zip_longest(left, right, fillvalue=_SENTINEL):
367 if left_item is not right_item:
368 return False
369 return True
370
371
372def get_first_not_none(a: Any, b: Any) -> Any:
373 """Return the first argument if it is not `None`, otherwise return the second argument."""
374 return a if a is not None else b
375
376
377@dataclasses.dataclass(frozen=True)
378class SafeGetItemProxy:
379 """Wrapper redirecting `__getitem__` to `get` with a sentinel value as default
380
381 This makes is safe to use in `operator.itemgetter` when some keys may be missing
382 """
383
384 # Define __slots__manually for performances
385 # @dataclasses.dataclass() only support slots=True in python>=3.10
386 __slots__ = ('wrapped',)
387
388 wrapped: Mapping[str, Any]
389
390 def __getitem__(self, key: str, /) -> Any:
391 return self.wrapped.get(key, _SENTINEL)
392
393 # required to pass the object to operator.itemgetter() instances due to a quirk of typeshed
394 # https://github.com/python/mypy/issues/13713
395 # https://github.com/python/typeshed/pull/8785
396 # Since this is typing-only, hide it in a typing.TYPE_CHECKING block
397 if TYPE_CHECKING:
398
399 def __contains__(self, key: str, /) -> bool:
400 return self.wrapped.__contains__(key)
401
402
403_ModelT = TypeVar('_ModelT', bound='BaseModel')
404_RT = TypeVar('_RT')
405
406
407class deprecated_instance_property(Generic[_ModelT, _RT]):
408 """A decorator exposing the decorated class method as a property, with a warning on instance access.
409
410 This decorator takes a class method defined on the `BaseModel` class and transforms it into
411 an attribute. The attribute can be accessed on both the class and instances of the class. If accessed
412 via an instance, a deprecation warning is emitted stating that instance access will be removed in V3.
413 """
414
415 def __init__(self, fget: Callable[[type[_ModelT]], _RT], /) -> None:
416 # Note: fget should be a classmethod:
417 self.fget = fget
418
419 @overload
420 def __get__(self, instance: None, objtype: type[_ModelT]) -> _RT: ...
421 @overload
422 @deprecated(
423 'Accessing this attribute on the instance is deprecated, and will be removed in Pydantic V3. '
424 'Instead, you should access this attribute from the model class.',
425 category=None,
426 )
427 def __get__(self, instance: _ModelT, objtype: type[_ModelT]) -> _RT: ...
428 def __get__(self, instance: _ModelT | None, objtype: type[_ModelT]) -> _RT:
429 if instance is not None:
430 # fmt: off
431 attr_name = (
432 self.fget.__name__
433 if sys.version_info >= (3, 10)
434 else self.fget.__func__.__name__ # pyright: ignore[reportFunctionMemberAccess]
435 )
436 # fmt: on
437 warnings.warn(
438 f'Accessing the {attr_name!r} attribute on the instance is deprecated. '
439 'Instead, you should access this attribute from the model class.',
440 category=PydanticDeprecatedSince211,
441 stacklevel=2,
442 )
443 return self.fget.__get__(instance, objtype)()