# This file is part of Hypothesis, which may be found at # https://github.com/HypothesisWorks/hypothesis/ # # Copyright the Hypothesis Authors. # Individual contributors are listed in AUTHORS.rst and the git log. # # This Source Code Form is subject to the terms of the Mozilla Public License, # v. 2.0. If a copy of the MPL was not distributed with this file, You can # obtain one at https://mozilla.org/MPL/2.0/. # Since Hypothesis doesn't have a hard dependency on attrs, be careful to only import # this file when attrs is in sys.modules. from collections.abc import Collection, Generator, Iterable, Sequence from functools import reduce from itertools import chain from types import EllipsisType from typing import Any, TypeVar import attr # attr/validators.pyi does not expose types for these, even though they exist # in source. from attr.validators import ( # type: ignore _AndValidator, _InstanceOfValidator, _InValidator, _OptionalValidator, ) from attrs import Attribute, AttrsInstance, Factory from hypothesis import strategies as st from hypothesis.errors import ResolutionFailed from hypothesis.internal.compat import get_type_hints from hypothesis.strategies._internal.core import BuildsStrategy from hypothesis.strategies._internal.strategies import SearchStrategy from hypothesis.strategies._internal.types import is_a_type, type_sorting_key from hypothesis.utils.conventions import infer T = TypeVar("T") def get_attribute_by_alias( fields: Iterable[Attribute], alias: str, *, target: type[AttrsInstance] | None = None, ) -> Attribute: """ Get an attrs attribute by its alias, rather than its name (compare getattr(fields, name)). ``target`` is used only to provide a nicer error message, and can be safely omitted. """ # attrs supports defining an alias for a field, which is the name used when # defining __init__. The init args are what we pull from when determining # what parameters we need to supply to the class, so it's what we need to # match against as well, rather than the class-level attribute name. matched_fields = [f for f in fields if f.alias == alias] if not matched_fields: raise TypeError( f"Unexpected keyword argument {alias} for attrs class" f"{f' {target}' if target else ''}. Expected one of " f"{[f.name for f in fields]}" ) # alias is used as an arg in __init__, so it is guaranteed to be unique, if # it exists. assert len(matched_fields) == 1 return matched_fields[0] def from_attrs( target: type[AttrsInstance], args: tuple[SearchStrategy[Any], ...], kwargs: dict[str, SearchStrategy[Any] | EllipsisType], to_infer: Iterable[str], ) -> SearchStrategy: """An internal version of builds(), specialised for Attrs classes.""" attributes: tuple[Attribute, ...] = attr.fields(target) kwargs = {k: v for k, v in kwargs.items() if v is not infer} for name in to_infer: attrib = get_attribute_by_alias(attributes, name, target=target) kwargs[name] = from_attrs_attribute(attrib, target) # We might make this strategy more efficient if we added a layer here that # retries drawing if validation fails, for improved composition. # The treatment of timezones in datetimes() provides a precedent. return BuildsStrategy(target, args, kwargs) def from_attrs_attribute( attrib: Attribute, target: type[AttrsInstance] ) -> SearchStrategy: """Infer a strategy from the metadata on an attr.Attribute object.""" # Try inferring from the default argument. Note that this will only help if # the user passed `...` to builds() for this attribute, but in that case # we use it as the minimal example. default: SearchStrategy = st.nothing() # attr/__init__.pyi uses overloads to declare Factory as a function, not a # class. This is a fib - at runtime and always, it is a class. if isinstance(attrib.default, Factory): # type: ignore assert attrib.default is not None if not attrib.default.takes_self: default = st.builds(attrib.default.factory) elif attrib.default is not attr.NOTHING: default = st.just(attrib.default) # Try inferring None, exact values, or type from attrs provided validators. # updated to none() on seeing an OptionalValidator null: SearchStrategy = st.nothing() # list of in_ validator collections to sample from in_collections = [] # type constraints to pass to types_to_strategy() validator_types = set() if attrib.validator is not None: validator = attrib.validator if isinstance(validator, _OptionalValidator): null = st.none() validator = validator.validator if isinstance(validator, _AndValidator): vs = validator._validators else: vs = [validator] for v in vs: if isinstance(v, _InValidator): if isinstance(v.options, str): in_collections.append(list(all_substrings(v.options))) else: in_collections.append(v.options) elif isinstance(v, _InstanceOfValidator): validator_types.add(v.type) # This is the important line. We compose the final strategy from various # parts. The default value, if any, is the minimal shrink, followed by # None (again, if allowed). We then prefer to sample from values passed # to an in_ validator if available, but infer from a type otherwise. # Pick one because (sampled_from((1, 2)) | from_type(int)) would usually # fail validation by generating e.g. zero! if in_collections: sample = st.sampled_from(list(ordered_intersection(in_collections))) strat = default | null | sample else: strat = default | null | types_to_strategy(attrib, validator_types) # Better to give a meaningful error here than an opaque "could not draw" # when we try to get a value but have lost track of where this was created. if strat.is_empty: raise ResolutionFailed( "Cannot infer a strategy from the default, validator, type, or " f"converter for attribute={attrib!r} of class={target!r}" ) return strat def types_to_strategy(attrib: Attribute, types: Collection[Any]) -> SearchStrategy: """Find all the type metadata for this attribute, reconcile it, and infer a strategy from the mess.""" # If we know types from the validator(s), that's sufficient. if len(types) == 1: (typ,) = types if isinstance(typ, tuple): return st.one_of(*map(st.from_type, typ)) return st.from_type(typ) elif types: # We have a list of tuples of types, and want to find a type # (or tuple of types) that is a subclass of all of of them. type_tuples = [k if isinstance(k, tuple) else (k,) for k in types] # Flatten the list, filter types that would fail validation, and # sort so that ordering is stable between runs and shrinks well. allowed = [ t for t in set(sum(type_tuples, ())) if all(issubclass(t, tup) for tup in type_tuples) ] allowed.sort(key=type_sorting_key) return st.one_of([st.from_type(t) for t in allowed]) # Otherwise, try the `type` attribute as a fallback, and finally try # the type hints on a converter (desperate!) before giving up. if is_a_type(getattr(attrib, "type", None)): assert attrib.type is not None # The convoluted test is because variable annotations may be stored # in string form; attrs doesn't evaluate them and we don't handle them. # See PEP 526, PEP 563, and Hypothesis issue #1004 for details. return st.from_type(attrib.type) converter = getattr(attrib, "converter", None) if isinstance(converter, type): return st.from_type(converter) elif callable(converter): hints = get_type_hints(converter) if "return" in hints: return st.from_type(hints["return"]) return st.nothing() def ordered_intersection(in_: Sequence[Iterable[T]]) -> Generator[T, None, None]: """Set union of n sequences, ordered for reproducibility across runs.""" intersection = reduce(set.intersection, in_, set(in_[0])) for x in chain.from_iterable(in_): if x in intersection: yield x intersection.remove(x) def all_substrings(s: str) -> Generator[str, None, None]: """Generate all substrings of `s`, in order of length then occurrence. Includes the empty string (first), and any duplicates that are present. >>> list(all_substrings('010')) ['', '0', '1', '0', '01', '10', '010'] """ yield s[:0] for n, _ in enumerate(s): for i in range(len(s) - n): yield s[i : i + n + 1]