azure-ai-evaluation 1.3.0__py3-none-any.whl → 1.5.0__py3-none-any.whl

azure/ai/evaluation/_evaluators/_tool_call_accuracy/tool_call_accuracy.prompty

@@ -0,0 +1,71 @@
+---
+name: Tool Call Accuracy
+description: Evaluates Tool Call Accuracy for tool used by agent
+model:
+  api: chat
+  parameters:
+    temperature: 0.0
+    max_tokens: 800
+    top_p: 1.0
+    presence_penalty: 0
+    frequency_penalty: 0
+    response_format:
+      type: text
+
+inputs:
+  query:
+    type: List
+  tool_call:
+    type: Dict
+  tool_definition:
+    type: Dict
+
+---
+system:
+# Instruction
+## Goal
+### Your are an expert in evaluating the accuracy of a tool call considering relevance and potential usefulness including syntactic and semantic correctness of a proposed tool call from an intelligent system based on provided definition and data. Your goal will involve answering the questions below using the information provided.
+- **Definition**: You are given a definition of the communication trait that is being evaluated to help guide your Score.
+- **Data**: Your input data include CONVERSATION , TOOL CALL and TOOL DEFINITION.
+- **Tasks**: To complete your evaluation you will be asked to evaluate the Data in different ways.
+
+user:
+# Definition
+**Tool Call Accuracy** refers to the relevance and potential usefulness of a TOOL CALL in the context of an ongoing CONVERSATION and EXTRACTION of RIGHT PARAMETER VALUES from the CONVERSATION.It assesses how likely the TOOL CALL is to contribute meaningfully to the CONVERSATION and help address the user's needs. Focus on evaluating the potential value of the TOOL CALL within the specific context of the given CONVERSATION, without making assumptions beyond the provided information.
+  Consider the following factors in your evaluation:
+
+  1. Relevance: How well does the proposed tool call align with the current topic and flow of the conversation?
+  2. Parameter Appropriateness: Do the parameters used in the TOOL CALL match the TOOL DEFINITION and are the parameters relevant to the latest user's query?
+  3. Parameter Value Correctness: Are the parameters values used in the TOOL CALL present or inferred by CONVERSATION and relevant to the latest user's query?
+  4. Potential Value: Is the information this tool call might provide likely to be useful in advancing the conversation or addressing the user expressed or implied needs?
+  5. Context Appropriateness: Does the tool call make sense at this point in the conversation, given what has been discussed so far?
+
+
+# Ratings
+## [Tool Call Accuracy: 0] (Irrelevant)
+**Definition:**
+ 1. The TOOL CALL is not relevant and will not help resolve the user's need.
+ 2. TOOL CALL include parameters values that are not present or inferred from CONVERSATION.
+ 3. TOOL CALL has parameters that is not present in TOOL DEFINITION.
+
+## [Tool Call Accuracy: 1] (Relevant)
+**Definition:**
+ 1. The TOOL CALL is directly relevant and very likely to help resolve the user's need.
+ 2. TOOL CALL include parameters values that are present or inferred from CONVERSATION.
+ 3. TOOL CALL has parameters that is present in TOOL DEFINITION.
+
+# Data
+CONVERSATION : {{query}}
+TOOL CALL: {{tool_call}}
+TOOL DEFINITION: {{tool_definition}}
+
+
+# Tasks
+## Please provide your assessment Score for the previous CONVERSATION , TOOL CALL and TOOL DEFINITION based on the Definitions above. Your output should include the following information:
+- **ThoughtChain**: To improve the reasoning process, think step by step and include a step-by-step explanation of your thought process as you analyze the data based on the definitions. Keep it brief and start your ThoughtChain with "Let's think step by step:".
+- **Explanation**: a very short explanation of why you think the input Data should get that Score.
+- **Score**: based on your previous analysis, provide your Score. The Score you give MUST be a integer score (i.e., "0", "1") based on the levels of the definitions.
+
+
+## Please provide your answers between the tags: <S0>your chain of thoughts</S0>, <S1>your explanation</S1>, <S2>your Score</S2>.
+# Output

azure/ai/evaluation/_evaluators/_ungrounded_attributes/__init__.py

@@ -0,0 +1,5 @@
+from ._ungrounded_attributes import UngroundedAttributesEvaluator
+
+__all__ = [
+    "UngroundedAttributesEvaluator",
+]

azure/ai/evaluation/_evaluators/_ungrounded_attributes/_ungrounded_attributes.py

@@ -0,0 +1,103 @@
+# ---------------------------------------------------------
+# Copyright (c) Microsoft Corporation. All rights reserved.
+# ---------------------------------------------------------
+from typing_extensions import overload, override
+from typing import Dict, Union
+
+from azure.ai.evaluation._common._experimental import experimental
+from azure.ai.evaluation._common.constants import EvaluationMetrics
+from azure.ai.evaluation._evaluators._common import RaiServiceEvaluatorBase
+
+@experimental
+class UngroundedAttributesEvaluator(RaiServiceEvaluatorBase[Union[str, bool]]):
+    """
+    Evaluates ungrounded inference of human attributes for a given query, response, and context for a single-turn evaluation only, 
+    where query represents the user query and response represents the AI system response given the provided context. 
+ 
+    Ungrounded Attributes checks for whether a response is first, ungrounded, and checks if it contains information about protected class or 
+    emotional state of a person.
+
+
+    It identifies the following attributes:
+    
+    - emotional_state
+    - protected_class
+    - groundedness
+
+    :param credential: The credential for connecting to Azure AI project. Required
+    :type credential: ~azure.core.credentials.TokenCredential
+    :param azure_ai_project: The scope of the Azure AI project.
+        It contains subscription id, resource group, and project name.
+    :type azure_ai_project: ~azure.ai.evaluation.AzureAIProject
+    :param kwargs: Additional arguments to pass to the evaluator.
+    :type kwargs: Any
+
+    .. admonition:: Example:
+
+        .. literalinclude:: ../samples/evaluation_samples_evaluate.py
+            :start-after: [START ungrounded_attributes_evaluator]
+            :end-before: [END ungrounded_attributes_evaluator]
+            :language: python
+            :dedent: 8
+            :caption: Initialize and call a UngroundedAttributesEvaluator with a query, response and context.
+
+    .. note::
+
+        If this evaluator is supplied to the `evaluate` function, the metric
+        for the ungrounded attributes will be "ungrounded_attributes_label".
+    """
+
+    id = "ungrounded_attributes"
+    """Evaluator identifier, experimental and to be used only with evaluation in cloud."""
+
+    @override
+    def __init__(
+        self,
+        credential,
+        azure_ai_project,
+    ):
+        super().__init__(
+            eval_metric=EvaluationMetrics.UNGROUNDED_ATTRIBUTES,
+            azure_ai_project=azure_ai_project,
+            credential=credential,
+        )
+
+    @overload
+    def __call__(
+        self,
+        *,
+        query: str,
+        response: str,
+        context: str,
+    ) -> Dict[str, Union[str, float]]:
+        """Evaluate a given query/response pair and context for ungrounded attributes
+
+        :keyword query: The query to be evaluated.
+        :paramtype query: str
+        :keyword response: The response to be evaluated.
+        :paramtype response: str
+        :keyword context: The context to be used for evaluation.
+        :paramtype context: str
+        :return: The ungrounded attributes label.
+        :rtype: Dict[str, Union[str, bool]]
+        """
+
+    @override
+    def __call__(  # pylint: disable=docstring-missing-param
+        self,
+        *args,
+        **kwargs,
+    ):
+        """Evaluate a given query/response pair and context for ungrounded attributes
+
+        :keyword query: The query to be evaluated.
+        :paramtype query: str
+        :keyword response: The response to be evaluated.
+        :paramtype response: str
+        :keyword context: The context to be used for evaluation.
+        :paramtype context: str
+        :return: The ungrounded attributes label.
+        :rtype: Dict[str, Union[str, bool]]
+        """
+        
+        return super().__call__(*args, **kwargs)

azure/ai/evaluation/_evaluators/_xpia/xpia.py

@@ -43,6 +43,8 @@ class IndirectAttackEvaluator(RaiServiceEvaluatorBase[Union[str, bool]]):
     :param azure_ai_project: The scope of the Azure AI project. It contains subscription id, resource group, and project
         name.
     :type azure_ai_project: ~azure.ai.evaluation.AzureAIProject
+    :param threshold: The threshold for the IndirectAttack evaluator. Default is 0.
+    :type threshold: int
 
     .. admonition:: Example:
 

azure/ai/evaluation/_exceptions.py

@@ -62,15 +62,18 @@ class ErrorTarget(Enum):
     CODE_CLIENT = "CodeClient"
     RAI_CLIENT = "RAIClient"
     COHERENCE_EVALUATOR = "CoherenceEvaluator"
+    COMPLETENESS_EVALUATOR = "CompletenessEvaluator"
     CONTENT_SAFETY_CHAT_EVALUATOR = "ContentSafetyEvaluator"
     ECI_EVALUATOR = "ECIEvaluator"
     F1_EVALUATOR = "F1Evaluator"
     GROUNDEDNESS_EVALUATOR = "GroundednessEvaluator"
     PROTECTED_MATERIAL_EVALUATOR = "ProtectedMaterialEvaluator"
+    INTENT_RESOLUTION_EVALUATOR = "IntentResolutionEvaluator"
     RELEVANCE_EVALUATOR = "RelevanceEvaluator"
     SIMILARITY_EVALUATOR = "SimilarityEvaluator"
     FLUENCY_EVALUATOR = "FluencyEvaluator"
     RETRIEVAL_EVALUATOR = "RetrievalEvaluator"
+    TASK_ADHERENCE_EVALUATOR = "TaskAdherenceEvaluator"
     INDIRECT_ATTACK_EVALUATOR = "IndirectAttackEvaluator"
     INDIRECT_ATTACK_SIMULATOR = "IndirectAttackSimulator"
     ADVERSARIAL_SIMULATOR = "AdversarialSimulator"
@@ -80,6 +83,8 @@ class ErrorTarget(Enum):
     MODELS = "Models"
     UNKNOWN = "Unknown"
     CONVERSATION = "Conversation"
+    TOOL_CALL_ACCURACY_EVALUATOR = "ToolCallAccuracyEvaluator"
+    RED_TEAM = "RedTeam"
 
 
 class EvaluationException(AzureError):

azure/ai/evaluation/_legacy/__init__.py

@@ -0,0 +1,3 @@
+# ---------------------------------------------------------
+# Copyright (c) Microsoft Corporation. All rights reserved.
+# ---------------------------------------------------------

azure/ai/evaluation/_legacy/_adapters/__init__.py

@@ -0,0 +1,21 @@
+# ---------------------------------------------------------
+# Copyright (c) Microsoft Corporation. All rights reserved.
+# ---------------------------------------------------------
+
+# NOTE: This contains adapters that make the Promptflow dependency optional. In the first phase,
+#       Promptflow will still be installed as part of the azure-ai-evaluation dependencies. This
+#       will be removed in the future once the code migration is complete.
+
+from typing import Final
+
+
+_has_legacy = False
+try:
+    from promptflow.client import PFClient
+
+    _has_legacy = True
+except ImportError:
+    pass
+
+HAS_LEGACY_SDK: Final[bool] = _has_legacy
+MISSING_LEGACY_SDK: Final[bool] = not _has_legacy

azure/ai/evaluation/_legacy/_adapters/_configuration.py

@@ -0,0 +1,45 @@
+# ---------------------------------------------------------
+# Copyright (c) Microsoft Corporation. All rights reserved.
+# ---------------------------------------------------------
+
+from pathlib import Path
+from typing import Any, Dict, Final, Optional
+from typing_extensions import TypeAlias
+
+
+try:
+    from promptflow._sdk._configuration import Configuration as _Configuration
+except ImportError:
+    _global_config: Final[Dict[str, Any]] = {}
+
+    class _Configuration:
+        TRACE_DESTINATION: Final[str] = "trace.destination"
+        _instance = None
+
+        def __init__(self, *, override_config: Optional[Dict[str, Any]] = None) -> None:
+            self._config = override_config or {}
+
+        @classmethod
+        def get_instance(cls) -> "_Configuration":
+            """Use this to get instance to avoid multiple copies of same global config."""
+            if cls._instance is None:
+                cls._instance = Configuration(override_config=_global_config)
+            return cls._instance
+
+        def set_config(self, key: str, value: Any) -> None:
+            # Simulated config storage
+            self._config[key] = value
+
+        def get_config(self, key: str) -> Any:
+            # Simulated config storage
+            if key in self._config:
+                return self._config[key]
+            return _global_config.get(key, None)
+
+        def get_trace_destination(self, path: Optional[Path] = None) -> Optional[str]:
+            if path:
+                raise NotImplementedError("Setting trace destination with a path is not supported.")
+            return self._config.get("trace.destination", None)
+
+
+Configuration: TypeAlias = _Configuration

azure/ai/evaluation/_legacy/_adapters/_constants.py

@@ -0,0 +1,10 @@
+# ---------------------------------------------------------
+# Copyright (c) Microsoft Corporation. All rights reserved.
+# ---------------------------------------------------------
+
+from typing import Final
+
+
+PF_FLOW_ENTRY_IN_TMP: Final[str] = "PF_FLOW_ENTRY_IN_TMP"
+PF_FLOW_META_LOAD_IN_SUBPROCESS: Final[str] = "PF_FLOW_META_LOAD_IN_SUBPROCESS"
+LINE_NUMBER: Final[str] = "line_number"

azure/ai/evaluation/_legacy/_adapters/_errors.py

@@ -0,0 +1,29 @@
+# ---------------------------------------------------------
+# Copyright (c) Microsoft Corporation. All rights reserved.
+# ---------------------------------------------------------
+
+from typing import Any
+from typing_extensions import TypeAlias
+
+
+try:
+    from promptflow.core._errors import MissingRequiredPackage as _MissingRequiredPackage
+except ImportError:
+    from azure.ai.evaluation._exceptions import ErrorBlame, ErrorCategory, ErrorTarget, EvaluationException
+
+    class _MissingRequiredPackage(EvaluationException):
+        """Raised when a required package is missing.
+
+        :param message: A message describing the error. This is the error message the user will see.
+        :type message: str
+        """
+
+        def __init__(self, message: str, **kwargs: Any):
+            kwargs.setdefault("category", ErrorCategory.MISSING_PACKAGE)
+            kwargs.setdefault("blame", ErrorBlame.SYSTEM_ERROR)
+            kwargs.setdefault("target", ErrorTarget.EVALUATE)
+            kwargs.setdefault("internal_message", "Missing required package.")
+            super().__init__(message=message, **kwargs)
+
+
+MissingRequiredPackage: TypeAlias = _MissingRequiredPackage

azure/ai/evaluation/_legacy/_adapters/_flows.py

@@ -0,0 +1,28 @@
+# ---------------------------------------------------------
+# Copyright (c) Microsoft Corporation. All rights reserved.
+# ---------------------------------------------------------
+
+from typing_extensions import TypeAlias
+
+
+try:
+    from promptflow._sdk.entities._flows import AsyncPrompty as _AsyncPrompty
+    from promptflow._sdk.entities._flows import FlexFlow as _FlexFlow
+    from promptflow._sdk.entities._flows.dag import Flow as _Flow
+except ImportError:
+    from azure.ai.evaluation._legacy.prompty import AsyncPrompty as _AsyncPrompty
+
+    class _FlexFlow:
+        pass
+
+    _FlexFlow.__name__ = "FlexFlow"
+
+    class _Flow:
+        name: str
+
+    _Flow.__name__ = "Flow"
+
+
+AsyncPrompty: TypeAlias = _AsyncPrompty
+FlexFlow: TypeAlias = _FlexFlow
+Flow: TypeAlias = _Flow

azure/ai/evaluation/_legacy/_adapters/_service.py

@@ -0,0 +1,16 @@
+# ---------------------------------------------------------
+# Copyright (c) Microsoft Corporation. All rights reserved.
+# ---------------------------------------------------------
+
+from typing import Callable, Final
+
+
+try:
+    from promptflow._cli._pf._service import stop_service as _stop_service
+except ImportError:
+
+    def _stop_service() -> None:
+        pass
+
+
+stop_service: Final[Callable[[], None]] = _stop_service

azure/ai/evaluation/_legacy/_adapters/client.py

@@ -0,0 +1,51 @@
+# ---------------------------------------------------------
+# Copyright (c) Microsoft Corporation. All rights reserved.
+# ---------------------------------------------------------
+
+from os import PathLike
+from typing import Any, Callable, Dict, Optional, Union
+from typing_extensions import TypeAlias
+
+import pandas as pd
+
+from ._errors import MissingRequiredPackage
+from ._configuration import Configuration
+from .entities import Run
+
+
+try:
+    from promptflow.client import PFClient as _PFClient
+except ImportError:
+
+    class _PFClient:
+        def __init__(self, **kwargs):
+            self._config = Configuration(override_config=kwargs.pop("config", None))
+
+        def run(
+            self,
+            flow: Union[str, PathLike, Callable],
+            *,
+            data: Union[str, PathLike],
+            run: Optional[Union[str, Run]] = None,
+            column_mapping: Optional[dict] = None,
+            variant: Optional[str] = None,
+            connections: Optional[dict] = None,
+            environment_variables: Optional[dict] = None,
+            name: Optional[str] = None,
+            display_name: Optional[str] = None,
+            tags: Optional[Dict[str, str]] = None,
+            resume_from: Optional[Union[str, Run]] = None,
+            code: Optional[Union[str, PathLike]] = None,
+            init: Optional[dict] = None,
+            **kwargs,
+        ) -> Run:
+            raise MissingRequiredPackage("Please install 'promptflow' package to use PFClient")
+
+        def get_details(self, run: Union[str, Run], max_results: int = 100, all_results: bool = False) -> pd.DataFrame:
+            return pd.DataFrame()
+
+        def get_metrics(self, run: Union[str, Run]) -> Dict[str, Any]:
+            return {}
+
+
+PFClient: TypeAlias = _PFClient

azure/ai/evaluation/_legacy/_adapters/entities.py

@@ -0,0 +1,26 @@
+# ---------------------------------------------------------
+# Copyright (c) Microsoft Corporation. All rights reserved.
+# ---------------------------------------------------------
+
+from typing_extensions import TypeAlias
+
+
+try:
+    from promptflow._sdk.entities import Run as _Run
+except ImportError:
+    from typing_extensions import Protocol
+    from typing import Any, Dict, Optional
+    from datetime import datetime
+    from pathlib import Path
+
+    class _Run(Protocol):
+        name: str
+        status: str
+        _properties: Dict[str, Any]
+        _created_on: datetime
+        _end_time: Optional[datetime]
+        _experiment_name: Optional[str]
+        _output_path: Path
+
+
+Run: TypeAlias = _Run

azure/ai/evaluation/_legacy/_adapters/tracing.py

@@ -0,0 +1,28 @@
+# ---------------------------------------------------------
+# Copyright (c) Microsoft Corporation. All rights reserved.
+# ---------------------------------------------------------
+
+from typing import Callable, Dict, Final, Optional
+from typing_extensions import TypeAlias
+
+
+try:
+    from promptflow.tracing import ThreadPoolExecutorWithContext as _ThreadPoolExecutorWithContext
+    from promptflow.tracing._integrations._openai_injector import (
+        inject_openai_api as _inject,
+        recover_openai_api as _recover,
+    )
+    from promptflow.tracing import _start_trace
+except ImportError:
+    from concurrent.futures import ThreadPoolExecutor as _ThreadPoolExecutorWithContext
+    from azure.ai.evaluation._legacy._batch_engine._openai_injector import (
+        inject_openai_api as _inject,
+        recover_openai_api as _recover,
+    )
+    from azure.ai.evaluation._legacy._batch_engine._trace import start_trace as _start_trace
+
+
+ThreadPoolExecutorWithContext: TypeAlias = _ThreadPoolExecutorWithContext
+inject_openai_api: Final[Callable[[], None]] = _inject
+recover_openai_api: Final[Callable[[], None]] = _recover
+start_trace: Final = _start_trace

azure/ai/evaluation/_legacy/_adapters/types.py

@@ -0,0 +1,15 @@
+# ---------------------------------------------------------
+# Copyright (c) Microsoft Corporation. All rights reserved.
+# ---------------------------------------------------------
+
+from typing import Any
+
+
+class AttrDict(dict):
+    """A dictionary that allows attribute access to its keys."""
+
+    def __getattr__(self, key: str) -> Any:
+        return self[key]
+
+    def __setattr__(self, key: str, value: Any) -> None:
+        self[key] = value

azure/ai/evaluation/_legacy/_adapters/utils.py

@@ -0,0 +1,31 @@
+# ---------------------------------------------------------
+# Copyright (c) Microsoft Corporation. All rights reserved.
+# ---------------------------------------------------------
+
+from typing import Final, Optional
+from typing_extensions import TypeAlias
+
+
+try:
+    from promptflow._utils.user_agent_utils import ClientUserAgentUtil as _ClientUserAgentUtil
+    from promptflow._utils.async_utils import async_run_allowing_running_loop as _async_run_allowing_running_loop
+    from promptflow._cli._utils import get_workspace_triad_from_local as _get_workspace_triad_from_local
+except ImportError:
+    from azure.ai.evaluation._legacy._batch_engine._utils_deprecated import (
+        async_run_allowing_running_loop as _async_run_allowing_running_loop,
+    )
+    from azure.ai.evaluation._evaluate._utils import AzureMLWorkspace
+
+    class _ClientUserAgentUtil:
+        @staticmethod
+        def append_user_agent(user_agent: Optional[str]):
+            # TODO ralphe: implement?
+            pass
+
+    def _get_workspace_triad_from_local() -> AzureMLWorkspace:
+        return AzureMLWorkspace("", "", "")
+
+
+ClientUserAgentUtil: TypeAlias = _ClientUserAgentUtil
+async_run_allowing_running_loop: Final = _async_run_allowing_running_loop
+get_workspace_triad_from_local: Final = _get_workspace_triad_from_local

azure/ai/evaluation/_legacy/_batch_engine/__init__.py

@@ -0,0 +1,9 @@
+# ---------------------------------------------------------
+# Copyright (c) Microsoft Corporation. All rights reserved.
+# ---------------------------------------------------------
+
+# NOTE: This is a direct port of the bare minimum needed for BatchEngine functionality from
+#       the original Promptflow code. The goal here is expediency, not elegance. As such
+#       parts of this code may be a little "quirky", seem incomplete in places, or contain
+#       longer TODOs comments than usual. In a future code update, large swaths of this code
+#       will be refactored or deleted outright.

azure/ai/evaluation/_legacy/_batch_engine/_config.py

@@ -0,0 +1,45 @@
+# ---------------------------------------------------------
+# Copyright (c) Microsoft Corporation. All rights reserved.
+# ---------------------------------------------------------
+
+from dataclasses import dataclass
+from logging import Logger
+
+from ..._constants import PF_BATCH_TIMEOUT_SEC_DEFAULT
+
+
+@dataclass
+class BatchEngineConfig:
+    """Context for a batch of evaluations. This will contain the configuration,
+    logging, and other needed information."""
+
+    logger: Logger
+    """The logger to use for logging messages."""
+
+    batch_timeout_seconds: int = PF_BATCH_TIMEOUT_SEC_DEFAULT
+    """The maximum amount of time to wait for all evaluations in the batch to complete."""
+
+    run_timeout_seconds: int = 600
+    """The maximum amount of time to wait for an evaluation to run against a single entry
+    in the data input to complete."""
+
+    max_concurrency: int = 10
+    """The maximum number of evaluations to run concurrently."""
+
+    use_async: bool = True
+    """Whether to use asynchronous evaluation."""
+
+    default_num_results: int = 100
+    """The default number of results to return if you don't ask for all results."""
+
+    def __post_init__(self):
+        if self.logger is None:
+            raise ValueError("logger cannot be None")
+        if self.batch_timeout_seconds <= 0:
+            raise ValueError("batch_timeout_seconds must be greater than 0")
+        if self.run_timeout_seconds <= 0:
+            raise ValueError("run_timeout_seconds must be greater than 0")
+        if self.max_concurrency <= 0:
+            raise ValueError("max_concurrency must be greater than 0")
+        if self.default_num_results <= 0:
+            raise ValueError("default_num_results must be greater than 0")