rasa-pro 3.9.18__py3-none-any.whl → 3.10.4__py3-none-any.whl

rasa/e2e_test/assertions.py

@@ -0,0 +1,1181 @@
+from __future__ import annotations
+
+import dataclasses
+import json
+import re
+from dataclasses import dataclass
+from enum import Enum
+from functools import lru_cache
+from typing import (
+    Any,
+    Callable,
+    Dict,
+    List,
+    Optional,
+    Set,
+    TYPE_CHECKING,
+    Text,
+    Tuple,
+    Type,
+)
+
+import pandas as pd
+import structlog
+
+import rasa.shared.utils.common
+from rasa.core.constants import (
+    DOMAIN_GROUND_TRUTH_METADATA_KEY,
+    UTTER_SOURCE_METADATA_KEY,
+)
+from rasa.core.policies.enterprise_search_policy import (
+    SEARCH_QUERY_METADATA_KEY,
+    SEARCH_RESULTS_METADATA_KEY,
+)
+from rasa.dialogue_understanding.patterns.clarify import FLOW_PATTERN_CLARIFICATION
+from rasa.shared.core.constants import DEFAULT_SLOT_NAMES
+from rasa.shared.core.events import (
+    ActionExecuted,
+    BotUttered,
+    DefinePrevUserUtteredFeaturization,
+    DialogueStackUpdated,
+    Event,
+    FlowCancelled,
+    FlowCompleted,
+    FlowStarted,
+    SlotSet,
+)
+from rasa.shared.exceptions import RasaException
+from rasa.utils.common import update_mlflow_log_level
+from rasa.utils.json_utils import SetEncoder
+
+if TYPE_CHECKING:
+    from rasa.e2e_test.e2e_config import LLMJudgeConfig
+
+
+structlogger = structlog.get_logger()
+
+DEFAULT_THRESHOLD = 0.5
+ELIGIBLE_UTTER_SOURCE_METADATA = [
+    "EnterpriseSearchPolicy",
+    "ContextualResponseRephraser",
+    "IntentlessPolicy",
+]
+
+
+class AssertionType(Enum):
+    FLOW_STARTED = "flow_started"
+    FLOW_COMPLETED = "flow_completed"
+    FLOW_CANCELLED = "flow_cancelled"
+    PATTERN_CLARIFICATION_CONTAINS = "pattern_clarification_contains"
+    ACTION_EXECUTED = "action_executed"
+    SLOT_WAS_SET = "slot_was_set"
+    SLOT_WAS_NOT_SET = "slot_was_not_set"
+    BOT_UTTERED = "bot_uttered"
+    GENERATIVE_RESPONSE_IS_RELEVANT = "generative_response_is_relevant"
+    GENERATIVE_RESPONSE_IS_GROUNDED = "generative_response_is_grounded"
+
+
+@lru_cache(maxsize=1)
+def _get_all_assertion_subclasses() -> Dict[str, Type[Assertion]]:
+    return {
+        sub_class.type(): sub_class
+        for sub_class in rasa.shared.utils.common.all_subclasses(Assertion)
+    }
+
+
+class InvalidAssertionType(RasaException):
+    """Raised if an assertion type is invalid."""
+
+    def __init__(self, assertion_type: str) -> None:
+        """Creates a `InvalidAssertionType`.
+
+        Args:
+            assertion_type: The invalid assertion type.
+        """
+        super().__init__(f"Invalid assertion type '{assertion_type}'.")
+
+
+@dataclass
+class Assertion:
+    """Base class for storing assertions."""
+
+    @classmethod
+    def type(cls) -> str:
+        """Returns the type of the assertion."""
+        raise NotImplementedError
+
+    @staticmethod
+    def from_dict(assertion_dict: Dict[Text, Any]) -> Assertion:
+        """Creates an assertion from a dictionary."""
+        raise NotImplementedError
+
+    def as_dict(self) -> Dict[str, Any]:
+        """Return the `Assertion` as a dictionary.
+
+        Returns:
+            The `Assertion` as a dictionary.
+        """
+        data = dataclasses.asdict(self)
+        data["type"] = self.type()
+        return data
+
+    @staticmethod
+    def create_typed_assertion(data: Dict[str, Any]) -> Assertion:
+        """Creates a `Assertion` from a dictionary.
+
+        Args:
+            data: The dictionary to create the `Assertion` from.
+
+        Returns:
+            The created `Assertion`.
+        """
+        typ = next(iter(data.keys()))
+
+        subclass_mapping = _get_all_assertion_subclasses()
+
+        clazz = subclass_mapping.get(typ)
+
+        if clazz is None:
+            structlogger.warning("assertion.unknown_type", data=data)
+            raise InvalidAssertionType(typ)
+
+        try:
+            return clazz.from_dict(data)
+        except NotImplementedError:
+            structlogger.warning("assertion.unknown_type", data=data)
+            raise InvalidAssertionType(typ)
+
+    def run(
+        self,
+        turn_events: List[Event],
+        prior_events: List[Event],
+        assertion_order_error_message: str = "",
+        **kwargs: Any,
+    ) -> Tuple[Optional[AssertionFailure], Optional[Event]]:
+        """Run the assertion on the given events for that user turn.
+
+        Args:
+            turn_events: The events to run the assertion on.
+            prior_events: All events prior to the current turn.
+            assertion_order_error_message: The error message to append if the assertion
+                order is enabled.
+            kwargs: Additional keyword arguments.
+
+        Returns:
+            A tuple of the assertion failure and the matching event if the assertion
+            passes, otherwise `None`.
+        """
+        raise NotImplementedError
+
+    def _generate_assertion_failure(
+        self,
+        error_message: str,
+        prior_events: List[Event],
+        turn_events: List[Event],
+        line: Optional[int] = None,
+    ) -> Tuple[AssertionFailure, None]:
+        return AssertionFailure(
+            assertion=self,
+            error_message=error_message,
+            actual_events_transcript=create_actual_events_transcript(
+                prior_events, turn_events
+            ),
+            error_line=line,
+        ), None
+
+
+@dataclass
+class FlowStartedAssertion(Assertion):
+    """Class for storing the flow started assertion."""
+
+    flow_id: str
+    line: Optional[int] = None
+
+    @classmethod
+    def type(cls) -> str:
+        return AssertionType.FLOW_STARTED.value
+
+    @staticmethod
+    def from_dict(assertion_dict: Dict[Text, Any]) -> FlowStartedAssertion:
+        return FlowStartedAssertion(
+            flow_id=assertion_dict.get(AssertionType.FLOW_STARTED.value),
+            line=assertion_dict.lc.line + 1 if hasattr(assertion_dict, "lc") else None,
+        )
+
+    def run(
+        self,
+        turn_events: List[Event],
+        prior_events: List[Event],
+        assertion_order_error_message: str = "",
+        **kwargs: Any,
+    ) -> Tuple[Optional[AssertionFailure], Optional[Event]]:
+        """Run the flow started assertion on the given events for that user turn."""
+        try:
+            matching_event = next(
+                event
+                for event in turn_events
+                if isinstance(event, FlowStarted) and event.flow_id == self.flow_id
+            )
+        except StopIteration:
+            error_message = f"Flow with id '{self.flow_id}' did not start."
+            error_message += assertion_order_error_message
+
+            return self._generate_assertion_failure(
+                error_message, prior_events, turn_events, self.line
+            )
+
+        return None, matching_event
+
+    def __hash__(self) -> int:
+        return hash(json.dumps(self.as_dict()))
+
+
+@dataclass
+class FlowCompletedAssertion(Assertion):
+    """Class for storing the flow completed assertion."""
+
+    flow_id: str
+    flow_step_id: Optional[str] = None
+    line: Optional[int] = None
+
+    @classmethod
+    def type(cls) -> str:
+        return AssertionType.FLOW_COMPLETED.value
+
+    @staticmethod
+    def from_dict(assertion_dict: Dict[Text, Any]) -> FlowCompletedAssertion:
+        line = assertion_dict.lc.line + 1 if hasattr(assertion_dict, "lc") else None
+        assertion_dict = assertion_dict.get(AssertionType.FLOW_COMPLETED.value, {})
+
+        return FlowCompletedAssertion(
+            flow_id=assertion_dict.get("flow_id"),
+            flow_step_id=assertion_dict.get("flow_step_id"),
+            line=line,
+        )
+
+    def run(
+        self,
+        turn_events: List[Event],
+        prior_events: List[Event],
+        assertion_order_error_message: str = "",
+        **kwargs: Any,
+    ) -> Tuple[Optional[AssertionFailure], Optional[Event]]:
+        """Run the flow completed assertion on the given events for that user turn."""
+        try:
+            matching_event = next(
+                event
+                for event in turn_events
+                if isinstance(event, FlowCompleted) and event.flow_id == self.flow_id
+            )
+        except StopIteration:
+            error_message = f"Flow with id '{self.flow_id}' did not complete."
+            error_message += assertion_order_error_message
+
+            return self._generate_assertion_failure(
+                error_message, prior_events, turn_events, self.line
+            )
+
+        if (
+            self.flow_step_id is not None
+            and matching_event.step_id != self.flow_step_id
+        ):
+            error_message = (
+                f"Flow with id '{self.flow_id}' did not complete "
+                f"at expected step id '{self.flow_step_id}'. The actual "
+                f"step id was '{matching_event.step_id}'."
+            )
+            error_message += assertion_order_error_message
+            return self._generate_assertion_failure(
+                error_message, prior_events, turn_events, self.line
+            )
+
+        return None, matching_event
+
+    def __hash__(self) -> int:
+        return hash(json.dumps(self.as_dict()))
+
+
+@dataclass
+class FlowCancelledAssertion(Assertion):
+    """Class for storing the flow cancelled assertion."""
+
+    flow_id: str
+    flow_step_id: Optional[str] = None
+    line: Optional[int] = None
+
+    @classmethod
+    def type(cls) -> str:
+        return AssertionType.FLOW_CANCELLED.value
+
+    @staticmethod
+    def from_dict(assertion_dict: Dict[Text, Any]) -> FlowCancelledAssertion:
+        line = assertion_dict.lc.line + 1 if hasattr(assertion_dict, "lc") else None
+        assertion_dict = assertion_dict.get(AssertionType.FLOW_CANCELLED.value, {})
+
+        return FlowCancelledAssertion(
+            flow_id=assertion_dict.get("flow_id"),
+            flow_step_id=assertion_dict.get("flow_step_id"),
+            line=line,
+        )
+
+    def run(
+        self,
+        turn_events: List[Event],
+        prior_events: List[Event],
+        assertion_order_error_message: str = "",
+        **kwargs: Any,
+    ) -> Tuple[Optional[AssertionFailure], Optional[Event]]:
+        """Run the flow cancelled assertion on the given events for that user turn."""
+        try:
+            matching_event = next(
+                event
+                for event in turn_events
+                if isinstance(event, FlowCancelled) and event.flow_id == self.flow_id
+            )
+        except StopIteration:
+            error_message = f"Flow with id '{self.flow_id}' was not cancelled."
+            error_message += assertion_order_error_message
+
+            return self._generate_assertion_failure(
+                error_message, prior_events, turn_events, self.line
+            )
+
+        if (
+            self.flow_step_id is not None
+            and matching_event.step_id != self.flow_step_id
+        ):
+            error_message = (
+                f"Flow with id '{self.flow_id}' was not cancelled "
+                f"at expected step id '{self.flow_step_id}'. The actual "
+                f"step id was '{matching_event.step_id}'."
+            )
+            error_message += assertion_order_error_message
+
+            return self._generate_assertion_failure(
+                error_message, prior_events, turn_events, self.line
+            )
+
+        return None, matching_event
+
+    def __hash__(self) -> int:
+        return hash(json.dumps(self.as_dict()))
+
+
+@dataclass
+class PatternClarificationContainsAssertion(Assertion):
+    """Class for storing the pattern clarification contains assertion."""
+
+    flow_names: Set[str]
+    line: Optional[int] = None
+
+    @classmethod
+    def type(cls) -> str:
+        return AssertionType.PATTERN_CLARIFICATION_CONTAINS.value
+
+    @staticmethod
+    def from_dict(
+        assertion_dict: Dict[Text, Any],
+    ) -> PatternClarificationContainsAssertion:
+        return PatternClarificationContainsAssertion(
+            flow_names=set(
+                assertion_dict.get(
+                    AssertionType.PATTERN_CLARIFICATION_CONTAINS.value, []
+                )
+            ),
+            line=assertion_dict.lc.line + 1 if hasattr(assertion_dict, "lc") else None,
+        )
+
+    def run(
+        self,
+        turn_events: List[Event],
+        prior_events: List[Event],
+        assertion_order_error_message: str = "",
+        **kwargs: Any,
+    ) -> Tuple[Optional[AssertionFailure], Optional[Event]]:
+        """Run the flow completed assertion on the given events for that user turn."""
+        try:
+            matching_event = next(
+                event
+                for event in turn_events
+                if isinstance(event, FlowStarted)
+                and event.flow_id == FLOW_PATTERN_CLARIFICATION
+            )
+        except StopIteration:
+            error_message = f"'{FLOW_PATTERN_CLARIFICATION}' pattern did not trigger."
+            error_message += assertion_order_error_message
+
+            return self._generate_assertion_failure(
+                error_message, prior_events, turn_events, self.line
+            )
+
+        actual_flow_names = set(matching_event.metadata.get("names", set()))
+        if actual_flow_names != self.flow_names:
+            error_message = (
+                f"'{FLOW_PATTERN_CLARIFICATION}' pattern did not contain "
+                f"the expected options. Expected options: {self.flow_names}. "
+            )
+            error_message += assertion_order_error_message
+
+            return self._generate_assertion_failure(
+                error_message, prior_events, turn_events, self.line
+            )
+
+        return None, matching_event
+
+    def __hash__(self) -> int:
+        return hash(json.dumps(self.as_dict(), cls=SetEncoder))
+
+
+@dataclass
+class ActionExecutedAssertion(Assertion):
+    """Class for storing the action executed assertion."""
+
+    action_name: str
+    line: Optional[int] = None
+
+    @classmethod
+    def type(cls) -> str:
+        return AssertionType.ACTION_EXECUTED.value
+
+    @staticmethod
+    def from_dict(assertion_dict: Dict[Text, Any]) -> ActionExecutedAssertion:
+        return ActionExecutedAssertion(
+            action_name=assertion_dict.get(AssertionType.ACTION_EXECUTED.value),
+            line=assertion_dict.lc.line + 1 if hasattr(assertion_dict, "lc") else None,
+        )
+
+    def run(
+        self,
+        turn_events: List[Event],
+        prior_events: List[Event],
+        assertion_order_error_message: str = "",
+        **kwargs: Any,
+    ) -> Tuple[Optional[AssertionFailure], Optional[Event]]:
+        """Run the action executed assertion on the given events for that user turn."""
+        try:
+            matching_event = next(
+                event
+                for event in turn_events
+                if isinstance(event, ActionExecuted)
+                and event.action_name == self.action_name
+            )
+        except StopIteration:
+            error_message = f"Action '{self.action_name}' did not execute."
+            error_message += assertion_order_error_message
+
+            return self._generate_assertion_failure(
+                error_message, prior_events, turn_events, self.line
+            )
+
+        return None, matching_event
+
+    def __hash__(self) -> int:
+        return hash(json.dumps(self.as_dict()))
+
+
+@dataclass
+class AssertedSlot:
+    """Class for storing information asserted about slots."""
+
+    name: str
+    value: Any
+    line: Optional[int] = None
+
+    @staticmethod
+    def from_dict(slot_dict: Dict[Text, Any]) -> AssertedSlot:
+        return AssertedSlot(
+            name=slot_dict.get("name"),
+            value=slot_dict.get("value", "value key is undefined"),
+            line=slot_dict.lc.line + 1 if hasattr(slot_dict, "lc") else None,
+        )
+
+
+@dataclass
+class SlotWasSetAssertion(Assertion):
+    """Class for storing the slot was set assertion."""
+
+    slots: List[AssertedSlot]
+
+    @classmethod
+    def type(cls) -> str:
+        return AssertionType.SLOT_WAS_SET.value
+
+    @staticmethod
+    def from_dict(assertion_dict: Dict[Text, Any]) -> SlotWasSetAssertion:
+        return SlotWasSetAssertion(
+            slots=[
+                AssertedSlot.from_dict(slot)
+                for slot in assertion_dict.get(AssertionType.SLOT_WAS_SET.value, [])
+            ],
+        )
+
+    def run(
+        self,
+        turn_events: List[Event],
+        prior_events: List[Event],
+        assertion_order_error_message: str = "",
+        **kwargs: Any,
+    ) -> Tuple[Optional[AssertionFailure], Optional[Event]]:
+        """Run the slot_was_set assertion on the given events for that user turn."""
+        matching_event = None
+
+        for slot in self.slots:
+            matching_events = [
+                event
+                for event in turn_events
+                if isinstance(event, SlotSet) and event.key == slot.name
+            ]
+            if not matching_events:
+                error_message = f"Slot '{slot.name}' was not set."
+                error_message += assertion_order_error_message
+
+                return self._generate_assertion_failure(
+                    error_message, prior_events, turn_events, slot.line
+                )
+
+            if slot.value == "value key is undefined":
+                matching_event = matching_events[0]
+                structlogger.debug(
+                    "slot_was_set_assertion.run",
+                    last_event_seen=matching_event,
+                    event_info="Slot value is not asserted and we have "
+                    "multiple events for the same slot. "
+                    "We will mark the first event as last event seen.",
+                )
+                continue
+
+            try:
+                matching_event = next(
+                    event for event in matching_events if event.value == slot.value
+                )
+            except StopIteration:
+                error_message = (
+                    f"Slot '{slot.name}' was set to a different value "
+                    f"'{matching_events[-1].value}' than the "
+                    f"expected '{slot.value}' value."
+                )
+                error_message += assertion_order_error_message
+
+                return self._generate_assertion_failure(
+                    error_message, prior_events, turn_events, slot.line
+                )
+
+        return None, matching_event
+
+    def __hash__(self) -> int:
+        return hash(json.dumps(self.as_dict()))
+
+
+@dataclass
+class SlotWasNotSetAssertion(Assertion):
+    """Class for storing the slot was not set assertion."""
+
+    slots: List[AssertedSlot]
+
+    @classmethod
+    def type(cls) -> str:
+        return AssertionType.SLOT_WAS_NOT_SET.value
+
+    @staticmethod
+    def from_dict(assertion_dict: Dict[Text, Any]) -> SlotWasNotSetAssertion:
+        return SlotWasNotSetAssertion(
+            slots=[
+                AssertedSlot.from_dict(slot)
+                for slot in assertion_dict.get(AssertionType.SLOT_WAS_NOT_SET.value, [])
+            ]
+        )
+
+    def run(
+        self,
+        turn_events: List[Event],
+        prior_events: List[Event],
+        assertion_order_error_message: str = "",
+        **kwargs: Any,
+    ) -> Tuple[Optional[AssertionFailure], Optional[Event]]:
+        """Run the slot_was_not_set assertion on the given events for that user turn."""
+        matching_event = None
+
+        for slot in self.slots:
+            matching_events = [
+                event
+                for event in turn_events
+                if isinstance(event, SlotSet) and event.key == slot.name
+            ]
+            if not matching_events:
+                continue
+
+            # take the most recent event in the list of matching events
+            # since that is the final value in the tracker for that user turn
+            matching_event = matching_events[-1]
+
+            if (
+                slot.value == "value key is undefined"
+                and matching_event.value is not None
+            ):
+                error_message = (
+                    f"Slot '{slot.name}' was set to '{matching_event.value}' but "
+                    f"it should not have been set."
+                )
+                error_message += assertion_order_error_message
+
+                return self._generate_assertion_failure(
+                    error_message, prior_events, turn_events, slot.line
+                )
+
+            if matching_event.value == slot.value:
+                error_message = (
+                    f"Slot '{slot.name}' was set to '{slot.value}' "
+                    f"but it should not have been set."
+                )
+                error_message += assertion_order_error_message
+
+                return self._generate_assertion_failure(
+                    error_message, prior_events, turn_events, slot.line
+                )
+
+        return None, matching_event
+
+    def __hash__(self) -> int:
+        return hash(json.dumps(self.as_dict()))
+
+
+@dataclass
+class AssertedButton:
+    """Class for storing information asserted about buttons."""
+
+    title: str
+    payload: Optional[str] = None
+
+    @staticmethod
+    def from_dict(button_dict: Dict[Text, Any]) -> AssertedButton:
+        return AssertedButton(
+            title=button_dict.get("title"),
+            payload=button_dict.get("payload"),
+        )
+
+
+@dataclass
+class BotUtteredAssertion(Assertion):
+    """Class for storing the bot uttered assertion."""
+
+    utter_name: Optional[str] = None
+    text_matches: Optional[str] = None
+    buttons: Optional[List[AssertedButton]] = None
+    line: Optional[int] = None
+
+    @classmethod
+    def type(cls) -> str:
+        return AssertionType.BOT_UTTERED.value
+
+    @staticmethod
+    def from_dict(assertion_dict: Dict[Text, Any]) -> BotUtteredAssertion:
+        utter_name, text_matches, buttons = (
+            BotUtteredAssertion._extract_assertion_properties(assertion_dict)
+        )
+
+        if BotUtteredAssertion._assertion_is_empty(utter_name, text_matches, buttons):
+            raise RasaException(
+                "A 'bot_uttered' assertion is empty, it should contain at least one "
+                "of the allowed properties: 'utter_name', 'text_matches', 'buttons'."
+            )
+
+        return BotUtteredAssertion(
+            utter_name=utter_name,
+            text_matches=text_matches,
+            buttons=buttons,
+            line=assertion_dict.lc.line + 1 if hasattr(assertion_dict, "lc") else None,
+        )
+
+    @staticmethod
+    def _extract_assertion_properties(
+        assertion_dict: Dict[Text, Any],
+    ) -> Tuple[Optional[str], Optional[str], List[AssertedButton]]:
+        """Extracts the assertion properties from a dictionary."""
+        assertion_dict = assertion_dict.get(AssertionType.BOT_UTTERED.value, {})
+        utter_name = assertion_dict.get("utter_name")
+        text_matches = assertion_dict.get("text_matches")
+        buttons = [
+            AssertedButton.from_dict(button)
+            for button in assertion_dict.get("buttons", [])
+        ]
+
+        return utter_name, text_matches, buttons
+
+    @staticmethod
+    def _assertion_is_empty(
+        utter_name: Optional[str],
+        text_matches: Optional[str],
+        buttons: List[AssertedButton],
+    ) -> bool:
+        """Validate if the bot uttered assertion is empty."""
+        if not utter_name and not text_matches and not buttons:
+            return True
+
+        return False
+
+    def run(
+        self,
+        turn_events: List[Event],
+        prior_events: List[Event],
+        assertion_order_error_message: str = "",
+        **kwargs: Any,
+    ) -> Tuple[Optional[AssertionFailure], Optional[Event]]:
+        """Run the bot_uttered assertion on the given events for that user turn."""
+        matching_event = None
+
+        if self.utter_name is not None:
+            try:
+                matching_event = next(
+                    event
+                    for event in turn_events
+                    if isinstance(event, BotUttered)
+                    and event.metadata.get("utter_action") == self.utter_name
+                )
+            except StopIteration:
+                error_message = f"Bot did not utter '{self.utter_name}' response."
+                error_message += assertion_order_error_message
+
+                return self._generate_assertion_failure(
+                    error_message, prior_events, turn_events, self.line
+                )
+
+        if self.text_matches is not None:
+            pattern = re.compile(self.text_matches)
+            try:
+                matching_event = next(
+                    event
+                    for event in turn_events
+                    if isinstance(event, BotUttered) and pattern.search(event.text)
+                )
+            except StopIteration:
+                error_message = (
+                    f"Bot did not utter any response which "
+                    f"matches the provided text pattern "
+                    f"'{self.text_matches}'."
+                )
+                error_message += assertion_order_error_message
+
+                return self._generate_assertion_failure(
+                    error_message, prior_events, turn_events, self.line
+                )
+
+        if self.buttons:
+            try:
+                matching_event = next(
+                    event
+                    for event in turn_events
+                    if isinstance(event, BotUttered) and self._buttons_match(event)
+                )
+            except StopIteration:
+                error_message = (
+                    "Bot did not utter any response with the expected buttons."
+                )
+                error_message += assertion_order_error_message
+                return self._generate_assertion_failure(
+                    error_message, prior_events, turn_events, self.line
+                )
+
+        return None, matching_event
+
+    def _buttons_match(self, event: BotUttered) -> bool:
+        """Check if the bot response contains the expected buttons."""
+        # a button is a dictionary with keys 'title' and 'payload'
+        actual_buttons = event.data.get("buttons", [])
+        if not actual_buttons:
+            return False
+
+        return all(
+            self._button_matches(actual_button, expected_button)
+            for actual_button, expected_button in zip(actual_buttons, self.buttons)
+        )
+
+    @staticmethod
+    def _button_matches(
+        actual_button: Dict[str, Any], expected_button: AssertedButton
+    ) -> bool:
+        """Check if the actual button matches the expected button."""
+        return (
+            actual_button.get("title") == expected_button.title
+            and actual_button.get("payload") == expected_button.payload
+        )
+
+    def __hash__(self) -> int:
+        return hash(json.dumps(self.as_dict()))
+
+
+@dataclass
+class GenerativeResponseMixin(Assertion):
+    """Mixin class for storing generative response assertions."""
+
+    threshold: float = DEFAULT_THRESHOLD
+    utter_name: Optional[str] = None
+    line: Optional[int] = None
+    metric_adjective: Optional[str] = None
+    metric_name: Optional[str] = None
+    mlflow_metric: Callable = print
+
+    @classmethod
+    def type(cls) -> str:
+        return ""
+
+    def _get_ground_truth(self, matching_event: BotUttered) -> str:
+        raise NotImplementedError
+
+    def as_dict(self) -> Dict[str, Any]:
+        data = super().as_dict()
+        data.pop("metric_name")
+        data.pop("metric_adjective")
+        data.pop("mlflow_metric")
+
+        return data
+
+    def _run_llm_evaluation(
+        self,
+        matching_event: BotUttered,
+        step_text: str,
+        llm_judge_config: "LLMJudgeConfig",
+        assertion_order_error_message: str,
+        prior_events: List[Event],
+        turn_events: List[Event],
+    ) -> Tuple[Optional[AssertionFailure], Optional[Event]]:
+        """Run the LLM evaluation on the given event."""
+        import mlflow
+
+        # we need to configure the log level for mlflow
+        # after a local import to avoid unnecessary logs
+        update_mlflow_log_level()
+
+        # extract user question from event if available
+        user_question_from_event = matching_event.metadata.get(
+            SEARCH_QUERY_METADATA_KEY
+        )
+        user_question = (
+            user_question_from_event if user_question_from_event else step_text
+        )
+
+        ground_truth = self._get_ground_truth(matching_event)
+
+        eval_data = pd.DataFrame(
+            {
+                "inputs": [user_question],
+                "ground_truth": [ground_truth],
+                "predictions": [matching_event.text],
+            }
+        )
+
+        model_uri = llm_judge_config.get_model_uri()
+
+        structlogger.debug(
+            f"generative_response_is_{self.metric_adjective}_assertion.run_llm_evaluation",
+            model_uri=model_uri,
+        )
+
+        with mlflow.start_run():
+            results = mlflow.evaluate(
+                data=eval_data,
+                targets="ground_truth",
+                predictions="predictions",
+                model_type="question-answering",
+                evaluators="default",
+                extra_metrics=[
+                    self.mlflow_metric(model_uri),
+                ],
+            )
+
+        # Evaluation result for each data record is available in `results.tables`.
+        eval_table = results.tables["eval_results_table"]
+        score = eval_table.iloc[0][f"{self.metric_name}/v1/score"]
+        justification = eval_table.iloc[0][f"{self.metric_name}/v1/justification"]
+
+        # convert 1-5 score to 0-1 float
+        score = score * 20 / 100 if score is not None else 0
+
+        structlogger.debug(
+            f"generative_response_is_{self.metric_adjective}_assertion.run_results",
+            matching_event=repr(matching_event),
+            score=score,
+            justification=justification,
+        )
+
+        if score < self.threshold:
+            error_message = (
+                f"Generative response '{matching_event.text}' "
+                f"given to the user input '{user_question}' "
+                f"was not {self.metric_adjective}. "
+                f"Expected score to be above '{self.threshold}' threshold, "
+                f"but was '{score}'. The explanation for this score is: "
+                f"{justification}."
+            )
+            error_message += assertion_order_error_message
+
+            return self._generate_assertion_failure(
+                error_message, prior_events, turn_events, self.line
+            )
+
+        return None, matching_event
+
+    def _run_assertion_with_utter_name(
+        self,
+        matching_events: List[BotUttered],
+        step_text: str,
+        llm_judge_config: "LLMJudgeConfig",
+        assertion_order_error_message: str,
+        prior_events: List[Event],
+        turn_events: List[Event],
+    ) -> Tuple[Optional[AssertionFailure], Optional[Event]]:
+        """Assert metric for the given utter name."""
+        try:
+            matching_event = next(
+                event
+                for event in matching_events
+                if event.metadata.get("utter_action") == self.utter_name
+            )
+        except StopIteration:
+            error_message = f"Bot did not utter '{self.utter_name}' response."
+            error_message += assertion_order_error_message
+
+            return self._generate_assertion_failure(
+                error_message, prior_events, turn_events, self.line
+            )
+
+        return self._run_llm_evaluation(
+            matching_event,
+            step_text,
+            llm_judge_config,
+            assertion_order_error_message,
+            prior_events,
+            turn_events,
+        )
+
+    def _run_assertion_for_multiple_generative_responses(
+        self,
+        matching_events: List[BotUttered],
+        step_text: str,
+        llm_judge_config: "LLMJudgeConfig",
+        assertion_order_error_message: str,
+        prior_events: List[Event],
+        turn_events: List[Event],
+    ) -> Tuple[Optional[AssertionFailure], Optional[Event]]:
+        """Run LLM evaluation for multiple bot utterances."""
+        structlogger.debug(
+            f"generative_response_is_{self.metric_adjective}_assertion.run",
+            event_info="Multiple generative responses found, "
+            "we will evaluate each of the responses.",
+        )
+
+        passing_events = set()
+        for event in matching_events:
+            failure, event_result = self._run_llm_evaluation(
+                event,
+                step_text,
+                llm_judge_config,
+                assertion_order_error_message,
+                prior_events,
+                turn_events,
+            )
+            if event_result is not None:
+                passing_events.add(event_result)
+        else:
+            if not passing_events:
+                error_message = (
+                    f"None of the generative responses issued by either the "
+                    f"Enterprise Search Policy, IntentlessPolicy or the "
+                    f"Contextual Response Rephraser were {self.metric_adjective}."
+                )
+                error_message += assertion_order_error_message
+
+                return self._generate_assertion_failure(
+                    error_message, prior_events, turn_events, self.line
+                )
+
+        return None, list(passing_events)[-1]
+
+    def run(
+        self,
+        turn_events: List[Event],
+        prior_events: List[Event],
+        assertion_order_error_message: str = "",
+        llm_judge_config: Optional["LLMJudgeConfig"] = None,
+        step_text: Optional[str] = None,
+        **kwargs: Any,
+    ) -> Tuple[Optional[AssertionFailure], Optional[Event]]:
+        """Run the LLM evaluation on the given events for that user turn."""
+        matching_events: List[BotUttered] = _find_matching_generative_events(
+            turn_events
+        )
+
+        if not matching_events:
+            error_message = (
+                "No generative response issued by either the Enterprise Search Policy, "
+                "IntentlessPolicy or the Contextual Response Rephraser was found, "
+                "but one was expected."
+            )
+            error_message += assertion_order_error_message
+
+            return self._generate_assertion_failure(
+                error_message, prior_events, turn_events, self.line
+            )
+
+        if self.utter_name is not None:
+            return self._run_assertion_with_utter_name(
+                matching_events,
+                step_text,
+                llm_judge_config,
+                assertion_order_error_message,
+                prior_events,
+                turn_events,
+            )
+
+        if len(matching_events) > 1:
+            return self._run_assertion_for_multiple_generative_responses(
+                matching_events,
+                step_text,
+                llm_judge_config,
+                assertion_order_error_message,
+                prior_events,
+                turn_events,
+            )
+
+        matching_event = matching_events[0]
+
+        return self._run_llm_evaluation(
+            matching_event,
+            step_text,
+            llm_judge_config,
+            assertion_order_error_message,
+            prior_events,
+            turn_events,
+        )
+
+
+@dataclass
+class GenerativeResponseIsRelevantAssertion(GenerativeResponseMixin):
+    """Class for storing the generative response is relevant assertion."""
+
+    def _get_ground_truth(self, matching_event: BotUttered) -> str:
+        return ""
+
+    @classmethod
+    def type(cls) -> str:
+        return AssertionType.GENERATIVE_RESPONSE_IS_RELEVANT.value
+
+    @staticmethod
+    def from_dict(
+        assertion_dict: Dict[Text, Any],
+    ) -> GenerativeResponseIsRelevantAssertion:
+        import mlflow
+
+        assertion_dict = assertion_dict.get(
+            AssertionType.GENERATIVE_RESPONSE_IS_RELEVANT.value, {}
+        )
+        return GenerativeResponseIsRelevantAssertion(
+            threshold=assertion_dict.get("threshold", DEFAULT_THRESHOLD),
+            utter_name=assertion_dict.get("utter_name"),
+            line=assertion_dict.lc.line + 1 if hasattr(assertion_dict, "lc") else None,
+            metric_name="answer_relevance",
+            metric_adjective="relevant",
+            mlflow_metric=mlflow.metrics.genai.answer_relevance,
+        )
+
+    def __hash__(self) -> int:
+        return hash(json.dumps(self.as_dict()))
+
+
+@dataclass
+class GenerativeResponseIsGroundedAssertion(GenerativeResponseMixin):
+    """Class for storing the generative response is grounded assertion."""
+
+    ground_truth: Optional[str] = None
+
+    @classmethod
+    def type(cls) -> str:
+        return AssertionType.GENERATIVE_RESPONSE_IS_GROUNDED.value
+
+    @staticmethod
+    def from_dict(
+        assertion_dict: Dict[Text, Any],
+    ) -> GenerativeResponseIsGroundedAssertion:
+        import mlflow
+
+        assertion_dict = assertion_dict.get(
+            AssertionType.GENERATIVE_RESPONSE_IS_GROUNDED.value, {}
+        )
+        return GenerativeResponseIsGroundedAssertion(
+            threshold=assertion_dict.get("threshold", DEFAULT_THRESHOLD),
+            utter_name=assertion_dict.get("utter_name"),
+            ground_truth=assertion_dict.get("ground_truth"),
+            line=assertion_dict.lc.line + 1 if hasattr(assertion_dict, "lc") else None,
+            metric_name="answer_correctness",
+            metric_adjective="grounded",
+            mlflow_metric=mlflow.metrics.genai.answer_correctness,
+        )
+
+    def __hash__(self) -> int:
+        return hash(json.dumps(self.as_dict()))
+
+    def _get_ground_truth(self, matching_event: BotUttered) -> str:
+        # extract ground truth from event if available or use the provided ground truth
+        ground_truth_event_metadata = matching_event.metadata.get(
+            SEARCH_RESULTS_METADATA_KEY, ""
+        ) or matching_event.metadata.get(DOMAIN_GROUND_TRUTH_METADATA_KEY, "")
+
+        if isinstance(ground_truth_event_metadata, list):
+            ground_truth_event_metadata = "\n".join(ground_truth_event_metadata)
+
+        ground_truth = (
+            self.ground_truth
+            if self.ground_truth is not None
+            else ground_truth_event_metadata
+        )
+
+        return ground_truth
+
+
+@dataclass
+class AssertionFailure:
+    """Class for storing the assertion failure."""
+
+    assertion: Assertion
+    error_message: Text
+    actual_events_transcript: List[Text]
+    error_line: Optional[int] = None
+
+    def as_dict(self) -> Dict[Text, Any]:
+        """Returns the assertion failure as a dictionary."""
+        return {
+            "assertion": self.assertion.as_dict(),
+            "error_message": self.error_message,
+            "actual_events_transcript": self.actual_events_transcript,
+        }
+
+
+def create_actual_events_transcript(
+    prior_events: List[Event], turn_events: List[Event]
+) -> List[Text]:
+    """Create the actual events transcript for the assertion failure."""
+    all_events = prior_events + turn_events
+
+    event_transcript = []
+
+    for event in all_events:
+        if isinstance(event, SlotSet) and event.key in DEFAULT_SLOT_NAMES:
+            continue
+        if isinstance(event, DefinePrevUserUtteredFeaturization):
+            continue
+        if isinstance(event, DialogueStackUpdated):
+            continue
+
+        event_transcript.append(repr(event))
+
+    return event_transcript
+
+
+def _find_matching_generative_events(turn_events: List[Event]) -> List[BotUttered]:
+    """Find the matching events for the generative response assertions."""
+    return [
+        event
+        for event in turn_events
+        if isinstance(event, BotUttered)
+        and event.metadata.get(UTTER_SOURCE_METADATA_KEY)
+        in ELIGIBLE_UTTER_SOURCE_METADATA
+    ]