Source code for hypothesis._settings

# 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/.

"""The settings module configures runtime options for Hypothesis.

Either an explicit settings object can be used or the default object on
this module can be modified.
"""

import contextlib
import datetime
import inspect
import os
import warnings
from enum import Enum, EnumMeta, IntEnum, unique
from typing import (
    TYPE_CHECKING,
    Any,
    ClassVar,
    Collection,
    Dict,
    List,
    Optional,
    TypeVar,
    Union,
)

import attr

from hypothesis.errors import (
    HypothesisDeprecationWarning,
    InvalidArgument,
    InvalidState,
)
from hypothesis.internal.reflection import get_pretty_function_description
from hypothesis.internal.validation import check_type, try_convert
from hypothesis.utils.conventions import not_set
from hypothesis.utils.dynamicvariables import DynamicVariable

if TYPE_CHECKING:
    from hypothesis.database import ExampleDatabase

__all__ = ["settings"]

all_settings: Dict[str, "Setting"] = {}

T = TypeVar("T")


class settingsProperty:
    def __init__(self, name, show_default):
        self.name = name
        self.show_default = show_default

    def __get__(self, obj, type=None):
        if obj is None:
            return self
        else:
            try:
                result = obj.__dict__[self.name]
                # This is a gross hack, but it preserves the old behaviour that
                # you can change the storage directory and it will be reflected
                # in the default database.
                if self.name == "database" and result is not_set:
                    from hypothesis.database import ExampleDatabase

                    result = ExampleDatabase(not_set)
                return result
            except KeyError:
                raise AttributeError(self.name) from None

    def __set__(self, obj, value):
        obj.__dict__[self.name] = value

    def __delete__(self, obj):
        raise AttributeError(f"Cannot delete attribute {self.name}")

    @property
    def __doc__(self):
        description = all_settings[self.name].description
        default = (
            repr(getattr(settings.default, self.name))
            if self.show_default
            else "(dynamically calculated)"
        )
        return f"{description}\n\ndefault value: ``{default}``"


default_variable = DynamicVariable(None)


class settingsMeta(type):
    def __init__(cls, *args, **kwargs):
        super().__init__(*args, **kwargs)

    @property
    def default(cls):
        v = default_variable.value
        if v is not None:
            return v
        if hasattr(settings, "_current_profile"):
            settings.load_profile(settings._current_profile)
            assert default_variable.value is not None
        return default_variable.value

    def _assign_default_internal(cls, value):
        default_variable.value = value

    def __setattr__(cls, name, value):
        if name == "default":
            raise AttributeError(
                "Cannot assign to the property settings.default - "
                "consider using settings.load_profile instead."
            )
        elif not (isinstance(value, settingsProperty) or name.startswith("_")):
            raise AttributeError(
                f"Cannot assign hypothesis.settings.{name}={value!r} - the settings "
                "class is immutable.  You can change the global default "
                "settings with settings.load_profile, or use @settings(...) "
                "to decorate your test instead."
            )
        return super().__setattr__(name, value)


class settings(metaclass=settingsMeta):
    """A settings object configures options including verbosity, runtime controls,
    persistence, determinism, and more.

    Default values are picked up from the settings.default object and
    changes made there will be picked up in newly created settings.
    """

    __definitions_are_locked = False
    _profiles: ClassVar[Dict[str, "settings"]] = {}
    __module__ = "hypothesis"

    def __getattr__(self, name):
        if name in all_settings:
            return all_settings[name].default
        else:
            raise AttributeError(f"settings has no attribute {name}")

    def __init__(
        self,
        parent: Optional["settings"] = None,
        *,
        # This looks pretty strange, but there's good reason: we want Mypy to detect
        # bad calls downstream, but not to freak out about the `= not_set` part even
        # though it's not semantically valid to pass that as an argument value.
        # The intended use is "like **kwargs, but more tractable for tooling".
        max_examples: int = not_set,  # type: ignore
        derandomize: bool = not_set,  # type: ignore
        database: Optional["ExampleDatabase"] = not_set,  # type: ignore
        verbosity: "Verbosity" = not_set,  # type: ignore
        phases: Collection["Phase"] = not_set,  # type: ignore
        stateful_step_count: int = not_set,  # type: ignore
        report_multiple_bugs: bool = not_set,  # type: ignore
        suppress_health_check: Collection["HealthCheck"] = not_set,  # type: ignore
        deadline: Union[int, float, datetime.timedelta, None] = not_set,  # type: ignore
        print_blob: bool = not_set,  # type: ignore
        backend: str = not_set,  # type: ignore
    ) -> None:
        if parent is not None:
            check_type(settings, parent, "parent")
        if derandomize not in (not_set, False):
            if database not in (not_set, None):  # type: ignore
                raise InvalidArgument(
                    "derandomize=True implies database=None, so passing "
                    f"{database=} too is invalid."
                )
            database = None

        defaults = parent or settings.default
        if defaults is not None:
            for setting in all_settings.values():
                value = locals()[setting.name]
                if value is not_set:
                    object.__setattr__(
                        self, setting.name, getattr(defaults, setting.name)
                    )
                else:
                    object.__setattr__(self, setting.name, setting.validator(value))

    def __call__(self, test: T) -> T:
        """Make the settings object (self) an attribute of the test.

        The settings are later discovered by looking them up on the test itself.
        """
        # Aliasing as Any avoids mypy errors (attr-defined) when accessing and
        # setting custom attributes on the decorated function or class.
        _test: Any = test

        # Using the alias here avoids a mypy error (return-value) later when
        # ``test`` is returned, because this check results in type refinement.
        if not callable(_test):
            raise InvalidArgument(
                "settings objects can be called as a decorator with @given, "
                f"but decorated {test=} is not callable."
            )
        if inspect.isclass(test):
            from hypothesis.stateful import RuleBasedStateMachine

            if issubclass(_test, RuleBasedStateMachine):
                attr_name = "_hypothesis_internal_settings_applied"
                if getattr(test, attr_name, False):
                    raise InvalidArgument(
                        "Applying the @settings decorator twice would "
                        "overwrite the first version; merge their arguments "
                        "instead."
                    )
                setattr(test, attr_name, True)
                _test.TestCase.settings = self
                return test  # type: ignore
            else:
                raise InvalidArgument(
                    "@settings(...) can only be used as a decorator on "
                    "functions, or on subclasses of RuleBasedStateMachine."
                )
        if hasattr(_test, "_hypothesis_internal_settings_applied"):
            # Can't use _hypothesis_internal_use_settings as an indicator that
            # @settings was applied, because @given also assigns that attribute.
            descr = get_pretty_function_description(test)
            raise InvalidArgument(
                f"{descr} has already been decorated with a settings object.\n"
                f"    Previous:  {_test._hypothesis_internal_use_settings!r}\n"
                f"    This:  {self!r}"
            )

        _test._hypothesis_internal_use_settings = self
        _test._hypothesis_internal_settings_applied = True
        return test

    @classmethod
    def _define_setting(
        cls,
        name,
        description,
        *,
        default,
        options=None,
        validator=None,
        show_default=True,
    ):
        """Add a new setting.

        - name is the name of the property that will be used to access the
          setting. This must be a valid python identifier.
        - description will appear in the property's docstring
        - default is the default value. This may be a zero argument
          function in which case it is evaluated and its result is stored
          the first time it is accessed on any given settings object.
        """
        if settings.__definitions_are_locked:
            raise InvalidState(
                "settings have been locked and may no longer be defined."
            )
        if options is not None:
            options = tuple(options)
            assert default in options

            def validator(value):
                if value not in options:
                    msg = f"Invalid {name}, {value!r}. Valid options: {options!r}"
                    raise InvalidArgument(msg)
                return value

        else:
            assert validator is not None

        all_settings[name] = Setting(
            name=name,
            description=description.strip(),
            default=default,
            validator=validator,
        )
        setattr(settings, name, settingsProperty(name, show_default))

    @classmethod
    def lock_further_definitions(cls):
        settings.__definitions_are_locked = True

    def __setattr__(self, name, value):
        raise AttributeError("settings objects are immutable")

    def __repr__(self):
        from hypothesis.internal.conjecture.data import AVAILABLE_PROVIDERS

        bits = sorted(
            f"{name}={getattr(self, name)!r}"
            for name in all_settings
            if (name != "backend" or len(AVAILABLE_PROVIDERS) > 1)  # experimental
        )
        return "settings({})".format(", ".join(bits))

    def show_changed(self):
        bits = []
        for name, setting in all_settings.items():
            value = getattr(self, name)
            if value != setting.default:
                bits.append(f"{name}={value!r}")
        return ", ".join(sorted(bits, key=len))

[docs] @staticmethod def register_profile( name: str, parent: Optional["settings"] = None, **kwargs: Any, ) -> None: """Registers a collection of values to be used as a settings profile. Settings profiles can be loaded by name - for example, you might create a 'fast' profile which runs fewer examples, keep the 'default' profile, and create a 'ci' profile that increases the number of examples and uses a different database to store failures. The arguments to this method are exactly as for :class:`~hypothesis.settings`: optional ``parent`` settings, and keyword arguments for each setting that will be set differently to parent (or settings.default, if parent is None). """ check_type(str, name, "name") settings._profiles[name] = settings(parent=parent, **kwargs)
[docs] @staticmethod def get_profile(name: str) -> "settings": """Return the profile with the given name.""" check_type(str, name, "name") try: return settings._profiles[name] except KeyError: raise InvalidArgument(f"Profile {name!r} is not registered") from None
[docs] @staticmethod def load_profile(name: str) -> None: """Loads in the settings defined in the profile provided. If the profile does not exist, InvalidArgument will be raised. Any setting not defined in the profile will be the library defined default for that setting. """ check_type(str, name, "name") settings._current_profile = name settings._assign_default_internal(settings.get_profile(name))
@contextlib.contextmanager def local_settings(s): with default_variable.with_value(s): yield s @attr.s() class Setting: name = attr.ib() description = attr.ib() default = attr.ib() validator = attr.ib() def _max_examples_validator(x): check_type(int, x, name="max_examples") if x < 1: raise InvalidArgument( f"max_examples={x!r} should be at least one. You can disable " "example generation with the `phases` setting instead." ) return x settings._define_setting( "max_examples", default=100, validator=_max_examples_validator, description=""" Once this many satisfying examples have been considered without finding any counter-example, Hypothesis will stop looking. Note that we might call your test function fewer times if we find a bug early or can tell that we've exhausted the search space; or more if we discard some examples due to use of .filter(), assume(), or a few other things that can prevent the test case from completing successfully. The default value is chosen to suit a workflow where the test will be part of a suite that is regularly executed locally or on a CI server, balancing total running time against the chance of missing a bug. If you are writing one-off tests, running tens of thousands of examples is quite reasonable as Hypothesis may miss uncommon bugs with default settings. For very complex code, we have observed Hypothesis finding novel bugs after *several million* examples while testing :pypi:`SymPy`. If you are running more than 100k examples for a test, consider using our :ref:`integration for coverage-guided fuzzing <fuzz_one_input>` - it really shines when given minutes or hours to run. """, ) settings._define_setting( "derandomize", default=False, options=(True, False), description=""" If True, seed Hypothesis' random number generator using a hash of the test function, so that every run will test the same set of examples until you update Hypothesis, Python, or the test function. This allows you to `check for regressions and look for bugs <https://blog.nelhage.com/post/two-kinds-of-testing/>`__ using :ref:`separate settings profiles <settings_profiles>` - for example running quick deterministic tests on every commit, and a longer non-deterministic nightly testing run. """, ) def _validate_database(db): from hypothesis.database import ExampleDatabase if db is None or isinstance(db, ExampleDatabase): return db raise InvalidArgument( "Arguments to the database setting must be None or an instance of " f"ExampleDatabase. Try passing database=ExampleDatabase({db!r}), or " "construct and use one of the specific subclasses in " "hypothesis.database" ) settings._define_setting( "database", default=not_set, show_default=False, description=""" An instance of :class:`~hypothesis.database.ExampleDatabase` that will be used to save examples to and load previous examples from. May be ``None`` in which case no storage will be used. See the :doc:`example database documentation <database>` for a list of built-in example database implementations, and how to define custom implementations. """, validator=_validate_database, )
[docs] @unique class Phase(IntEnum): explicit = 0 #: controls whether explicit examples are run. reuse = 1 #: controls whether previous examples will be reused. generate = 2 #: controls whether new examples will be generated. target = 3 #: controls whether examples will be mutated for targeting. shrink = 4 #: controls whether examples will be shrunk. explain = 5 #: controls whether Hypothesis attempts to explain test failures. def __repr__(self): return f"Phase.{self.name}"
class HealthCheckMeta(EnumMeta): def __iter__(self): deprecated = (HealthCheck.return_value, HealthCheck.not_a_test_method) return iter(x for x in super().__iter__() if x not in deprecated)
[docs] @unique class HealthCheck(Enum, metaclass=HealthCheckMeta): """Arguments for :attr:`~hypothesis.settings.suppress_health_check`. Each member of this enum is a type of health check to suppress. """ def __repr__(self): return f"{self.__class__.__name__}.{self.name}" @classmethod def all(cls) -> List["HealthCheck"]: # Skipping of deprecated attributes is handled in HealthCheckMeta.__iter__ note_deprecation( "`Healthcheck.all()` is deprecated; use `list(HealthCheck)` instead.", since="2023-04-16", has_codemod=True, stacklevel=1, ) return list(HealthCheck) data_too_large = 1 """Checks if too many examples are aborted for being too large. This is measured by the number of random choices that Hypothesis makes in order to generate something, not the size of the generated object. For example, choosing a 100MB object from a predefined list would take only a few bits, while generating 10KB of JSON from scratch might trigger this health check. """ filter_too_much = 2 """Check for when the test is filtering out too many examples, either through use of :func:`~hypothesis.assume()` or :ref:`filter() <filtering>`, or occasionally for Hypothesis internal reasons.""" too_slow = 3 """Check for when your data generation is extremely slow and likely to hurt testing.""" return_value = 5 """Deprecated; we always error if a test returns a non-None value.""" large_base_example = 7 """Checks if the natural example to shrink towards is very large.""" not_a_test_method = 8 """Deprecated; we always error if :func:`@given <hypothesis.given>` is applied to a method defined by :class:`python:unittest.TestCase` (i.e. not a test).""" function_scoped_fixture = 9 """Checks if :func:`@given <hypothesis.given>` has been applied to a test with a pytest function-scoped fixture. Function-scoped fixtures run once for the whole function, not once per example, and this is usually not what you want. Because of this limitation, tests that need to set up or reset state for every example need to do so manually within the test itself, typically using an appropriate context manager. Suppress this health check only in the rare case that you are using a function-scoped fixture that does not need to be reset between individual examples, but for some reason you cannot use a wider fixture scope (e.g. session scope, module scope, class scope). This check requires the :ref:`Hypothesis pytest plugin<pytest-plugin>`, which is enabled by default when running Hypothesis inside pytest.""" differing_executors = 10 """Checks if :func:`@given <hypothesis.given>` has been applied to a test which is executed by different :ref:`executors<custom-function-execution>`. If your test function is defined as a method on a class, that class will be your executor, and subclasses executing an inherited test is a common way for things to go wrong. The correct fix is often to bring the executor instance under the control of hypothesis by explicit parametrization over, or sampling from, subclasses, or to refactor so that :func:`@given <hypothesis.given>` is specified on leaf subclasses."""
@unique class Verbosity(IntEnum): quiet = 0 normal = 1 verbose = 2 debug = 3 def __repr__(self): return f"Verbosity.{self.name}" settings._define_setting( "verbosity", options=tuple(Verbosity), default=Verbosity.normal, description="Control the verbosity level of Hypothesis messages", ) def _validate_phases(phases): phases = tuple(phases) for a in phases: if not isinstance(a, Phase): raise InvalidArgument(f"{a!r} is not a valid phase") return tuple(p for p in list(Phase) if p in phases) settings._define_setting( "phases", default=tuple(Phase), description=( "Control which phases should be run. " "See :ref:`the full documentation for more details <phases>`" ), validator=_validate_phases, ) def _validate_stateful_step_count(x): check_type(int, x, name="stateful_step_count") if x < 1: raise InvalidArgument(f"stateful_step_count={x!r} must be at least one.") return x settings._define_setting( name="stateful_step_count", default=50, validator=_validate_stateful_step_count, description=""" Number of steps to run a stateful program for before giving up on it breaking. """, ) settings._define_setting( name="report_multiple_bugs", default=True, options=(True, False), description=""" Because Hypothesis runs the test many times, it can sometimes find multiple bugs in a single run. Reporting all of them at once is usually very useful, but replacing the exceptions can occasionally clash with debuggers. If disabled, only the exception with the smallest minimal example is raised. """, ) def validate_health_check_suppressions(suppressions): suppressions = try_convert(list, suppressions, "suppress_health_check") for s in suppressions: if not isinstance(s, HealthCheck): raise InvalidArgument( f"Non-HealthCheck value {s!r} of type {type(s).__name__} " "is invalid in suppress_health_check." ) if s in (HealthCheck.return_value, HealthCheck.not_a_test_method): note_deprecation( f"The {s.name} health check is deprecated, because this is always an error.", since="2023-03-15", has_codemod=False, stacklevel=2, ) return suppressions settings._define_setting( "suppress_health_check", default=(), description="""A list of :class:`~hypothesis.HealthCheck` items to disable.""", validator=validate_health_check_suppressions, ) class duration(datetime.timedelta): """A timedelta specifically measured in milliseconds.""" def __repr__(self): ms = self.total_seconds() * 1000 return f"timedelta(milliseconds={int(ms) if ms == int(ms) else ms!r})" def _validate_deadline(x): if x is None: return x invalid_deadline_error = InvalidArgument( f"deadline={x!r} (type {type(x).__name__}) must be a timedelta object, " "an integer or float number of milliseconds, or None to disable the " "per-test-case deadline." ) if isinstance(x, (int, float)): if isinstance(x, bool): raise invalid_deadline_error try: x = duration(milliseconds=x) except OverflowError: raise InvalidArgument( f"deadline={x!r} is invalid, because it is too large to represent " "as a timedelta. Use deadline=None to disable deadlines." ) from None if isinstance(x, datetime.timedelta): if x <= datetime.timedelta(0): raise InvalidArgument( f"deadline={x!r} is invalid, because it is impossible to meet a " "deadline <= 0. Use deadline=None to disable deadlines." ) return duration(seconds=x.total_seconds()) raise invalid_deadline_error settings._define_setting( "deadline", default=duration(milliseconds=200), validator=_validate_deadline, description=""" If set, a duration (as timedelta, or integer or float number of milliseconds) that each individual example (i.e. each time your test function is called, not the whole decorated test) within a test is not allowed to exceed. Tests which take longer than that may be converted into errors (but will not necessarily be if close to the deadline, to allow some variability in test run time). Set this to ``None`` to disable this behaviour entirely. """, ) def is_in_ci() -> bool: # GitHub Actions, Travis CI and AppVeyor have "CI" # Azure Pipelines has "TF_BUILD" return "CI" in os.environ or "TF_BUILD" in os.environ settings._define_setting( "print_blob", default=is_in_ci(), show_default=False, options=(True, False), description=""" If set to ``True``, Hypothesis will print code for failing examples that can be used with :func:`@reproduce_failure <hypothesis.reproduce_failure>` to reproduce the failing example. The default is ``True`` if the ``CI`` or ``TF_BUILD`` env vars are set, ``False`` otherwise. """, ) def _backend_validator(value): from hypothesis.internal.conjecture.data import AVAILABLE_PROVIDERS if value not in AVAILABLE_PROVIDERS: if value == "crosshair": # pragma: no cover install = '`pip install "hypothesis[crosshair]"` and try again.' raise InvalidArgument(f"backend={value!r} is not available. {install}") raise InvalidArgument( f"backend={value!r} is not available - maybe you need to install a plugin?" f"\n Installed backends: {sorted(AVAILABLE_PROVIDERS)!r}" ) return value settings._define_setting( "backend", default="hypothesis", show_default=False, validator=_backend_validator, description=""" EXPERIMENTAL AND UNSTABLE - see :ref:`alternative-backends`. The importable name of a backend which Hypothesis should use to generate primitive types. We aim to support heuristic-random, solver-based, and fuzzing-based backends. """, ) settings.lock_further_definitions() def note_deprecation( message: str, *, since: str, has_codemod: bool, stacklevel: int = 0 ) -> None: if since != "RELEASEDAY": date = datetime.date.fromisoformat(since) assert datetime.date(2021, 1, 1) <= date if has_codemod: message += ( "\n The `hypothesis codemod` command-line tool can automatically " "refactor your code to fix this warning." ) warnings.warn(HypothesisDeprecationWarning(message), stacklevel=2 + stacklevel) settings.register_profile("default", settings()) settings.load_profile("default") assert settings.default is not None # Check that the kwonly args to settings.__init__ is the same as the set of # defined settings - in case we've added or remove something from one but # not the other. assert set(all_settings) == { p.name for p in inspect.signature(settings.__init__).parameters.values() if p.kind == inspect.Parameter.KEYWORD_ONLY }