schemathesis 3.39.16__py3-none-any.whl → 4.0.0__py3-none-any.whl

schemathesis/engine/core.py

@@ -0,0 +1,169 @@
+from __future__ import annotations
+
+import threading
+from dataclasses import dataclass
+from typing import Sequence
+
+from schemathesis.auths import unregister as unregister_auth
+from schemathesis.core import SpecificationFeature
+from schemathesis.engine import Status, events, phases
+from schemathesis.schemas import BaseSchema
+
+from .context import EngineContext
+from .events import EventGenerator
+from .phases import Phase, PhaseName, PhaseSkipReason
+
+
+@dataclass
+class Engine:
+    schema: BaseSchema
+
+    def execute(self) -> EventStream:
+        """Execute all test phases."""
+        # Unregister auth if explicitly provided
+        if self.schema.config.auth.is_defined:
+            unregister_auth()
+
+        ctx = EngineContext(schema=self.schema, stop_event=threading.Event())
+        plan = self._create_execution_plan()
+        return EventStream(plan.execute(ctx), ctx.control.stop_event)
+
+    def _create_execution_plan(self) -> ExecutionPlan:
+        """Create execution plan based on configuration."""
+        phases = [
+            self.get_phase_config(PhaseName.PROBING, is_supported=True, requires_links=False),
+            self.get_phase_config(
+                PhaseName.EXAMPLES,
+                is_supported=self.schema.specification.supports_feature(SpecificationFeature.EXAMPLES),
+                requires_links=False,
+            ),
+            self.get_phase_config(
+                PhaseName.COVERAGE,
+                is_supported=self.schema.specification.supports_feature(SpecificationFeature.COVERAGE),
+                requires_links=False,
+            ),
+            self.get_phase_config(PhaseName.FUZZING, is_supported=True, requires_links=False),
+            self.get_phase_config(
+                PhaseName.STATEFUL_TESTING,
+                is_supported=self.schema.specification.supports_feature(SpecificationFeature.STATEFUL_TESTING),
+                requires_links=True,
+            ),
+        ]
+        return ExecutionPlan(phases)
+
+    def get_phase_config(
+        self,
+        phase_name: PhaseName,
+        *,
+        is_supported: bool = True,
+        requires_links: bool = False,
+    ) -> Phase:
+        """Helper to determine phase configuration with proper skip reasons."""
+        # Check if feature is supported by the schema
+        if not is_supported:
+            return Phase(
+                name=phase_name,
+                is_supported=False,
+                is_enabled=False,
+                skip_reason=PhaseSkipReason.NOT_SUPPORTED,
+            )
+
+        phase = phase_name.value.lower()
+        if (
+            phase in ("examples", "coverage", "fuzzing", "stateful")
+            and not self.schema.config.phases.get_by_name(name=phase).enabled
+        ):
+            return Phase(
+                name=phase_name,
+                is_supported=True,
+                is_enabled=False,
+                skip_reason=PhaseSkipReason.DISABLED,
+            )
+
+        if requires_links and self.schema.statistic.links.total == 0:
+            return Phase(
+                name=phase_name,
+                is_supported=True,
+                is_enabled=False,
+                skip_reason=PhaseSkipReason.NOT_APPLICABLE,
+            )
+
+        # Phase can be executed
+        return Phase(
+            name=phase_name,
+            is_supported=True,
+            is_enabled=True,
+            skip_reason=None,
+        )
+
+
+@dataclass
+class ExecutionPlan:
+    """Manages test execution phases."""
+
+    phases: Sequence[Phase]
+
+    def execute(self, engine: EngineContext) -> EventGenerator:
+        """Execute all phases in sequence."""
+        yield events.EngineStarted()
+        try:
+            if engine.is_interrupted:
+                yield from self._finish(engine)
+                return
+            if engine.is_interrupted:
+                yield from self._finish(engine)  # type: ignore[unreachable]
+                return
+
+            # Run main phases
+            for phase in self.phases:
+                if engine.has_reached_the_failure_limit:
+                    phase.skip_reason = PhaseSkipReason.FAILURE_LIMIT_REACHED
+                yield events.PhaseStarted(phase=phase)
+                if phase.should_execute(engine):
+                    yield from phases.execute(engine, phase)
+                else:
+                    if engine.has_reached_the_failure_limit:
+                        phase.skip_reason = PhaseSkipReason.FAILURE_LIMIT_REACHED
+                    yield events.PhaseFinished(phase=phase, status=Status.SKIP, payload=None)
+                if engine.is_interrupted:
+                    break  # type: ignore[unreachable]
+
+        except KeyboardInterrupt:
+            engine.stop()
+            yield events.Interrupted(phase=None)
+
+        # Always finish
+        yield from self._finish(engine)
+
+    def _finish(self, ctx: EngineContext) -> EventGenerator:
+        """Finish the test run."""
+        yield events.EngineFinished(running_time=ctx.running_time)
+
+
+@dataclass
+class EventStream:
+    """Schemathesis event stream.
+
+    Provides an API to control the execution flow.
+    """
+
+    generator: EventGenerator
+    stop_event: threading.Event
+
+    def __next__(self) -> events.EngineEvent:
+        return next(self.generator)
+
+    def __iter__(self) -> EventGenerator:
+        return self.generator
+
+    def stop(self) -> None:
+        """Stop the event stream.
+
+        Its next value will be the last one (Finished).
+        """
+        self.stop_event.set()
+
+    def finish(self) -> events.EngineEvent:
+        """Stop the event stream & return the last event."""
+        self.stop()
+        return next(self)

schemathesis/engine/errors.py

@@ -0,0 +1,464 @@
+"""Handling of recoverable errors in Schemathesis Engine.
+
+This module provides utilities for analyzing, classifying, and formatting exceptions
+that occur during test execution via Schemathesis Engine.
+"""
+
+from __future__ import annotations
+
+import enum
+import re
+from dataclasses import dataclass
+from functools import cached_property
+from typing import TYPE_CHECKING, Callable, Iterator, Sequence, cast
+
+from schemathesis import errors
+from schemathesis.core.errors import (
+    RECURSIVE_REFERENCE_ERROR_MESSAGE,
+    InvalidTransition,
+    SerializationNotPossible,
+    format_exception,
+    get_request_error_extras,
+    get_request_error_message,
+    split_traceback,
+)
+
+if TYPE_CHECKING:
+    import hypothesis.errors
+    import requests
+    from requests.exceptions import ChunkedEncodingError
+
+__all__ = ["EngineErrorInfo", "DeadlineExceeded", "UnsupportedRecursiveReference", "UnexpectedError"]
+
+
+class DeadlineExceeded(errors.SchemathesisError):
+    """Test took too long to run."""
+
+    @classmethod
+    def from_exc(cls, exc: hypothesis.errors.DeadlineExceeded) -> DeadlineExceeded:
+        runtime = exc.runtime.total_seconds() * 1000
+        deadline = exc.deadline.total_seconds() * 1000
+        return cls(
+            f"Test running time is too slow! It took {runtime:.2f}ms, which exceeds the deadline of {deadline:.2f}ms.\n"
+        )
+
+
+class UnsupportedRecursiveReference(errors.SchemathesisError):
+    """Recursive reference is impossible to resolve due to current limitations."""
+
+    def __init__(self) -> None:
+        super().__init__(RECURSIVE_REFERENCE_ERROR_MESSAGE)
+
+
+class UnexpectedError(errors.SchemathesisError):
+    """An unexpected error during the engine execution.
+
+    Used primarily to not let Hypothesis consider the test as flaky or detect multiple failures as we handle it
+    on our side.
+    """
+
+
+class EngineErrorInfo:
+    """Extended information about errors that happen during engine execution.
+
+    It serves as a caching wrapper around exceptions to avoid repeated computations.
+    """
+
+    def __init__(self, error: Exception, code_sample: str | None = None) -> None:
+        self._error = error
+        self._code_sample = code_sample
+
+    def __str__(self) -> str:
+        return self._error_repr
+
+    @cached_property
+    def _kind(self) -> RuntimeErrorKind:
+        """Error kind."""
+        return _classify(error=self._error)
+
+    @property
+    def title(self) -> str:
+        """A general error description."""
+        import requests
+
+        if isinstance(self._error, InvalidTransition):
+            return "Invalid Link Definition"
+
+        if isinstance(self._error, requests.RequestException):
+            return "Network Error"
+
+        if self._kind in (
+            RuntimeErrorKind.HYPOTHESIS_HEALTH_CHECK_DATA_TOO_LARGE,
+            RuntimeErrorKind.HYPOTHESIS_HEALTH_CHECK_FILTER_TOO_MUCH,
+            RuntimeErrorKind.HYPOTHESIS_HEALTH_CHECK_TOO_SLOW,
+            RuntimeErrorKind.HYPOTHESIS_HEALTH_CHECK_LARGE_BASE_EXAMPLE,
+        ):
+            return "Failed Health Check"
+
+        if self._kind in (
+            RuntimeErrorKind.SCHEMA_INVALID_REGULAR_EXPRESSION,
+            RuntimeErrorKind.SCHEMA_GENERIC,
+            RuntimeErrorKind.HYPOTHESIS_UNSATISFIABLE,
+        ):
+            return "Schema Error"
+
+        return {
+            RuntimeErrorKind.SCHEMA_UNSUPPORTED: "Unsupported Schema",
+            RuntimeErrorKind.SCHEMA_NO_LINKS_FOUND: "Missing Open API links",
+            RuntimeErrorKind.SCHEMA_INVALID_STATE_MACHINE: "Invalid OpenAPI Links Definition",
+            RuntimeErrorKind.HYPOTHESIS_UNSUPPORTED_GRAPHQL_SCALAR: "Unknown GraphQL Scalar",
+            RuntimeErrorKind.SERIALIZATION_UNBOUNDED_PREFIX: "XML serialization error",
+            RuntimeErrorKind.SERIALIZATION_NOT_POSSIBLE: "Serialization not possible",
+        }.get(self._kind, "Runtime Error")
+
+    @property
+    def message(self) -> str:
+        """Detailed error description."""
+        import hypothesis.errors
+        import requests
+
+        if isinstance(self._error, requests.RequestException):
+            return get_request_error_message(self._error)
+
+        if self._kind == RuntimeErrorKind.SCHEMA_UNSUPPORTED:
+            return str(self._error).strip()
+
+        if self._kind == RuntimeErrorKind.HYPOTHESIS_UNSUPPORTED_GRAPHQL_SCALAR and isinstance(
+            self._error, hypothesis.errors.InvalidArgument
+        ):
+            scalar_name = scalar_name_from_error(self._error)
+            return f"Scalar type '{scalar_name}' is not recognized"
+
+        if self._kind == RuntimeErrorKind.HYPOTHESIS_HEALTH_CHECK_DATA_TOO_LARGE:
+            return HEALTH_CHECK_MESSAGE_DATA_TOO_LARGE
+        if self._kind == RuntimeErrorKind.HYPOTHESIS_HEALTH_CHECK_FILTER_TOO_MUCH:
+            return HEALTH_CHECK_MESSAGE_FILTER_TOO_MUCH
+        if self._kind == RuntimeErrorKind.HYPOTHESIS_HEALTH_CHECK_TOO_SLOW:
+            return HEALTH_CHECK_MESSAGE_TOO_SLOW
+        if self._kind == RuntimeErrorKind.HYPOTHESIS_HEALTH_CHECK_LARGE_BASE_EXAMPLE:
+            return HEALTH_CHECK_MESSAGE_LARGE_BASE_EXAMPLE
+
+        if self._kind == RuntimeErrorKind.HYPOTHESIS_UNSATISFIABLE:
+            return f"{self._error}. Possible reasons:"
+
+        if self._kind in (
+            RuntimeErrorKind.SCHEMA_INVALID_REGULAR_EXPRESSION,
+            RuntimeErrorKind.SCHEMA_GENERIC,
+        ):
+            return self._error.message  # type: ignore
+
+        return str(self._error)
+
+    @cached_property
+    def extras(self) -> list[str]:
+        """Additional context about the error."""
+        import requests
+
+        if isinstance(self._error, requests.RequestException):
+            return get_request_error_extras(self._error)
+
+        if self._kind == RuntimeErrorKind.HYPOTHESIS_UNSATISFIABLE:
+            return [
+                "- Contradictory schema constraints, such as a minimum value exceeding the maximum.",
+                "- Invalid schema definitions for headers or cookies, for example allowing for non-ASCII characters.",
+                "- Excessive schema complexity, which hinders parameter generation.",
+            ]
+
+        return []
+
+    @cached_property
+    def _error_repr(self) -> str:
+        return format_exception(self._error, with_traceback=False)
+
+    @property
+    def has_useful_traceback(self) -> bool:
+        return self._kind not in (
+            RuntimeErrorKind.SCHEMA_INVALID_REGULAR_EXPRESSION,
+            RuntimeErrorKind.SCHEMA_INVALID_STATE_MACHINE,
+            RuntimeErrorKind.SCHEMA_UNSUPPORTED,
+            RuntimeErrorKind.SCHEMA_GENERIC,
+            RuntimeErrorKind.SCHEMA_NO_LINKS_FOUND,
+            RuntimeErrorKind.SERIALIZATION_NOT_POSSIBLE,
+            RuntimeErrorKind.HYPOTHESIS_UNSUPPORTED_GRAPHQL_SCALAR,
+            RuntimeErrorKind.HYPOTHESIS_UNSATISFIABLE,
+            RuntimeErrorKind.HYPOTHESIS_HEALTH_CHECK_LARGE_BASE_EXAMPLE,
+            RuntimeErrorKind.HYPOTHESIS_HEALTH_CHECK_TOO_SLOW,
+            RuntimeErrorKind.HYPOTHESIS_HEALTH_CHECK_DATA_TOO_LARGE,
+            RuntimeErrorKind.HYPOTHESIS_HEALTH_CHECK_FILTER_TOO_MUCH,
+            RuntimeErrorKind.NETWORK_OTHER,
+        )
+
+    @cached_property
+    def traceback(self) -> str:
+        return format_exception(self._error, with_traceback=True)
+
+    def format(self, *, bold: Callable[[str], str] = str, indent: str = "    ") -> str:
+        """Format error message with optional styling and traceback."""
+        message = []
+
+        title = self.title
+        if title:
+            message.append(f"{title}\n")
+
+        # Main message
+        body = self.message or str(self._error)
+        message.append(body)
+
+        # Extras
+        if self.extras:
+            extras = self.extras
+        elif self.has_useful_traceback:
+            extras = split_traceback(self.traceback)
+        else:
+            extras = []
+
+        if extras:
+            message.append("")  # Empty line before extras
+            message.extend(f"{indent}{extra}" for extra in extras)
+
+        if self._code_sample is not None:
+            message.append(f"\nReproduce with: \n\n    {self._code_sample}")
+
+        # Suggestion
+        suggestion = get_runtime_error_suggestion(self._kind, bold=bold)
+        if suggestion is not None:
+            message.append(f"\nTip: {suggestion}")
+
+        return "\n".join(message)
+
+
+def scalar_name_from_error(exception: hypothesis.errors.InvalidArgument) -> str:
+    # This one is always available as the format is checked upfront
+    match = re.search(r"Scalar '(\w+)' is not supported", str(exception))
+    match = cast(re.Match, match)
+    return match.group(1)
+
+
+def extract_health_check_error(error: hypothesis.errors.FailedHealthCheck) -> hypothesis.HealthCheck | None:
+    from hypothesis import HealthCheck
+
+    match = re.search(r"add HealthCheck\.(\w+) to the suppress_health_check ", str(error))
+    if match:
+        return {
+            "data_too_large": HealthCheck.data_too_large,
+            "filter_too_much": HealthCheck.filter_too_much,
+            "too_slow": HealthCheck.too_slow,
+            "large_base_example": HealthCheck.large_base_example,
+        }.get(match.group(1))
+    return None
+
+
+def get_runtime_error_suggestion(error_type: RuntimeErrorKind, bold: Callable[[str], str] = str) -> str | None:
+    """Get a user-friendly suggestion for handling the error."""
+
+    def _format_health_check_suggestion(label: str) -> str:
+        return f"Bypass this health check using {bold(f'`--suppress-health-check={label}`')}."
+
+    return {
+        RuntimeErrorKind.CONNECTION_SSL: f"Bypass SSL verification with {bold('`--tls-verify=false`')}.",
+        RuntimeErrorKind.HYPOTHESIS_UNSATISFIABLE: "Examine the schema for inconsistencies and consider simplifying it.",
+        RuntimeErrorKind.SCHEMA_NO_LINKS_FOUND: "Review your endpoint filters to include linked operations",
+        RuntimeErrorKind.SCHEMA_INVALID_REGULAR_EXPRESSION: "Ensure your regex is compatible with Python's syntax.\n"
+        "For guidance, visit: https://docs.python.org/3/library/re.html",
+        RuntimeErrorKind.HYPOTHESIS_UNSUPPORTED_GRAPHQL_SCALAR: "Define a custom strategy for it.\n"
+        "For guidance, visit: https://schemathesis.readthedocs.io/en/stable/guides/graphql-custom-scalars/",
+        RuntimeErrorKind.HYPOTHESIS_HEALTH_CHECK_DATA_TOO_LARGE: _format_health_check_suggestion("data_too_large"),
+        RuntimeErrorKind.HYPOTHESIS_HEALTH_CHECK_FILTER_TOO_MUCH: _format_health_check_suggestion("filter_too_much"),
+        RuntimeErrorKind.HYPOTHESIS_HEALTH_CHECK_TOO_SLOW: _format_health_check_suggestion("too_slow"),
+        RuntimeErrorKind.HYPOTHESIS_HEALTH_CHECK_LARGE_BASE_EXAMPLE: _format_health_check_suggestion(
+            "large_base_example"
+        ),
+    }.get(error_type)
+
+
+HEALTH_CHECK_MESSAGE_DATA_TOO_LARGE = """There's a notable occurrence of examples surpassing the maximum size limit.
+Typically, generating excessively large examples can compromise the quality of test outcomes.
+
+Consider revising the schema to more accurately represent typical use cases
+or applying constraints to reduce the data size."""
+HEALTH_CHECK_MESSAGE_FILTER_TOO_MUCH = """A significant number of generated examples are being filtered out, indicating
+that the schema's constraints may be too complex.
+
+This level of filtration can slow down testing and affect the distribution
+of generated data. Review and simplify the schema constraints where
+possible to mitigate this issue."""
+HEALTH_CHECK_MESSAGE_TOO_SLOW = "Data generation is extremely slow. Consider reducing the complexity of the schema."
+HEALTH_CHECK_MESSAGE_LARGE_BASE_EXAMPLE = """A health check has identified that the smallest example derived from the schema
+is excessively large, potentially leading to inefficient test execution.
+
+This is commonly due to schemas that specify large-scale data structures by
+default, such as an array with an extensive number of elements.
+
+Consider revising the schema to more accurately represent typical use cases
+or applying constraints to reduce the data size."""
+
+
+@enum.unique
+class RuntimeErrorKind(str, enum.Enum):
+    """Classification of runtime errors."""
+
+    # Connection related issues
+    CONNECTION_SSL = "connection_ssl"
+    CONNECTION_OTHER = "connection_other"
+    NETWORK_OTHER = "network_other"
+
+    # Hypothesis issues
+    HYPOTHESIS_UNSATISFIABLE = "hypothesis_unsatisfiable"
+    HYPOTHESIS_UNSUPPORTED_GRAPHQL_SCALAR = "hypothesis_unsupported_graphql_scalar"
+    HYPOTHESIS_HEALTH_CHECK_DATA_TOO_LARGE = "hypothesis_health_check_data_too_large"
+    HYPOTHESIS_HEALTH_CHECK_FILTER_TOO_MUCH = "hypothesis_health_check_filter_too_much"
+    HYPOTHESIS_HEALTH_CHECK_TOO_SLOW = "hypothesis_health_check_too_slow"
+    HYPOTHESIS_HEALTH_CHECK_LARGE_BASE_EXAMPLE = "hypothesis_health_check_large_base_example"
+
+    SCHEMA_INVALID_REGULAR_EXPRESSION = "schema_invalid_regular_expression"
+    SCHEMA_INVALID_STATE_MACHINE = "schema_invalid_state_machine"
+    SCHEMA_NO_LINKS_FOUND = "schema_no_links_found"
+    SCHEMA_UNSUPPORTED = "schema_unsupported"
+    SCHEMA_GENERIC = "schema_generic"
+
+    SERIALIZATION_NOT_POSSIBLE = "serialization_not_possible"
+    SERIALIZATION_UNBOUNDED_PREFIX = "serialization_unbounded_prefix"
+
+    UNCLASSIFIED = "unclassified"
+
+
+def _classify(*, error: Exception) -> RuntimeErrorKind:
+    """Classify an error."""
+    import hypothesis.errors
+    import requests
+    from hypothesis import HealthCheck
+
+    # Network-related errors
+    if isinstance(error, requests.RequestException):
+        if isinstance(error, requests.exceptions.SSLError):
+            return RuntimeErrorKind.CONNECTION_SSL
+        if isinstance(error, requests.exceptions.ConnectionError):
+            return RuntimeErrorKind.CONNECTION_OTHER
+        return RuntimeErrorKind.NETWORK_OTHER
+
+    # Hypothesis errors
+    if (
+        isinstance(error, hypothesis.errors.InvalidArgument)
+        and str(error).endswith("larger than Hypothesis is designed to handle")
+        or "can never generate an example, because min_size is larger than Hypothesis supports" in str(error)
+    ):
+        return RuntimeErrorKind.HYPOTHESIS_HEALTH_CHECK_LARGE_BASE_EXAMPLE
+    if isinstance(error, hypothesis.errors.Unsatisfiable):
+        return RuntimeErrorKind.HYPOTHESIS_UNSATISFIABLE
+    if isinstance(error, hypothesis.errors.FailedHealthCheck):
+        health_check = extract_health_check_error(error)
+        if health_check is not None:
+            return {
+                HealthCheck.data_too_large: RuntimeErrorKind.HYPOTHESIS_HEALTH_CHECK_DATA_TOO_LARGE,
+                HealthCheck.filter_too_much: RuntimeErrorKind.HYPOTHESIS_HEALTH_CHECK_FILTER_TOO_MUCH,
+                HealthCheck.too_slow: RuntimeErrorKind.HYPOTHESIS_HEALTH_CHECK_TOO_SLOW,
+                HealthCheck.large_base_example: RuntimeErrorKind.HYPOTHESIS_HEALTH_CHECK_LARGE_BASE_EXAMPLE,
+            }[health_check]
+        return RuntimeErrorKind.UNCLASSIFIED
+    if isinstance(error, hypothesis.errors.InvalidArgument) and str(error).startswith("Scalar "):
+        # Comes from `hypothesis-graphql`
+        return RuntimeErrorKind.HYPOTHESIS_UNSUPPORTED_GRAPHQL_SCALAR
+
+    # Schema errors
+    if isinstance(error, errors.InvalidSchema):
+        if isinstance(error, errors.InvalidRegexPattern):
+            return RuntimeErrorKind.SCHEMA_INVALID_REGULAR_EXPRESSION
+        return RuntimeErrorKind.SCHEMA_GENERIC
+    if isinstance(error, errors.InvalidStateMachine):
+        return RuntimeErrorKind.SCHEMA_INVALID_STATE_MACHINE
+    if isinstance(error, errors.NoLinksFound):
+        return RuntimeErrorKind.SCHEMA_NO_LINKS_FOUND
+    if isinstance(error, UnsupportedRecursiveReference):
+        # Recursive references are not supported right now
+        return RuntimeErrorKind.SCHEMA_UNSUPPORTED
+    if isinstance(error, errors.SerializationError):
+        if isinstance(error, errors.UnboundPrefix):
+            return RuntimeErrorKind.SERIALIZATION_UNBOUNDED_PREFIX
+        return RuntimeErrorKind.SERIALIZATION_NOT_POSSIBLE
+    return RuntimeErrorKind.UNCLASSIFIED
+
+
+def deduplicate_errors(errors: Sequence[Exception]) -> Iterator[Exception]:
+    """Deduplicate a list of errors."""
+    seen = set()
+    serialization_media_types = set()
+
+    for error in errors:
+        # Collect media types
+        if isinstance(error, SerializationNotPossible):
+            for media_type in error.media_types:
+                serialization_media_types.add(media_type)
+            continue
+
+        message = canonicalize_error_message(error)
+        if message not in seen:
+            seen.add(message)
+            yield error
+
+    if serialization_media_types:
+        yield SerializationNotPossible.from_media_types(*sorted(serialization_media_types))
+
+
+MEMORY_ADDRESS_RE = re.compile("0x[0-9a-fA-F]+")
+URL_IN_ERROR_MESSAGE_RE = re.compile(r"Max retries exceeded with url: .*? \(Caused by")
+
+
+def canonicalize_error_message(error: Exception, with_traceback: bool = True) -> str:
+    """Canonicalize error messages by removing dynamic components."""
+    message = format_exception(error, with_traceback=with_traceback)
+    # Replace memory addresses
+    message = MEMORY_ADDRESS_RE.sub("0xbaaaaaaaaaad", message)
+    # Remove URL information
+    return URL_IN_ERROR_MESSAGE_RE.sub("", message)
+
+
+def clear_hypothesis_notes(exc: Exception) -> None:
+    notes = getattr(exc, "__notes__", [])
+    if any("while generating" in note for note in notes):
+        notes.clear()
+
+
+def is_unrecoverable_network_error(exc: Exception) -> bool:
+    from http.client import RemoteDisconnected
+
+    from urllib3.exceptions import ProtocolError
+
+    def has_connection_reset(inner: BaseException) -> bool:
+        exc_str = str(inner)
+        if any(pattern in exc_str for pattern in ["Connection reset by peer", "[Errno 104]", "ECONNRESET"]):
+            return True
+
+        if inner.__context__ is not None:
+            return has_connection_reset(inner.__context__)
+
+        return False
+
+    if isinstance(exc.__context__, ProtocolError):
+        if len(exc.__context__.args) == 2 and isinstance(exc.__context__.args[1], RemoteDisconnected):
+            return True
+        if len(exc.__context__.args) == 1 and exc.__context__.args[0] == "Response ended prematurely":
+            return True
+
+    return has_connection_reset(exc)
+
+
+@dataclass()
+class UnrecoverableNetworkError:
+    error: requests.ConnectionError | ChunkedEncodingError
+    code_sample: str
+
+    __slots__ = ("error", "code_sample")
+
+    def __init__(self, error: requests.ConnectionError | ChunkedEncodingError, code_sample: str) -> None:
+        self.error = error
+        self.code_sample = code_sample
+
+
+@dataclass
+class TestingState:
+    unrecoverable_network_error: UnrecoverableNetworkError | None
+
+    __slots__ = ("unrecoverable_network_error",)
+
+    def __init__(self) -> None:
+        self.unrecoverable_network_error = None