fix: replace deprecated structured_output methods with new API

src/strands_evals/evaluators/faithfulness_evaluator.py

@@ -1,4 +1,5 @@
 from enum import Enum
+from typing import cast
 
 from pydantic import BaseModel, Field
 from strands import Agent
@@ -59,29 +60,33 @@ def evaluate(self, evaluation_case: EvaluationData[InputT, OutputT]) -> list[Eva
         parsed_input = self._get_last_turn(evaluation_case)
         prompt = self._format_prompt(parsed_input)
         evaluator_agent = Agent(model=self.model, system_prompt=self.system_prompt, callback_handler=None)
-        rating = evaluator_agent.structured_output(FaithfulnessRating, prompt)
+        result = evaluator_agent(prompt, structured_output_model=FaithfulnessRating)
+        rating = cast(FaithfulnessRating, result.structured_output)
         normalized_score = self._score_mapping[rating.score]
-        result = EvaluationOutput(
-            score=normalized_score,
-            test_pass=normalized_score >= 0.5,
-            reason=rating.reasoning,
-            label=rating.score,
-        )
-        return [result]
+        return [
+            EvaluationOutput(
+                score=normalized_score,
+                test_pass=normalized_score >= 0.5,
+                reason=rating.reasoning,
+                label=rating.score,
+            )
+        ]
 
     async def evaluate_async(self, evaluation_case: EvaluationData[InputT, OutputT]) -> list[EvaluationOutput]:
         parsed_input = self._get_last_turn(evaluation_case)
         prompt = self._format_prompt(parsed_input)
         evaluator_agent = Agent(model=self.model, system_prompt=self.system_prompt, callback_handler=None)
-        rating = await evaluator_agent.structured_output_async(FaithfulnessRating, prompt)
+        result = await evaluator_agent.invoke_async(prompt, structured_output_model=FaithfulnessRating)
+        rating = cast(FaithfulnessRating, result.structured_output)
         normalized_score = self._score_mapping[rating.score]
-        result = EvaluationOutput(
-            score=normalized_score,
-            test_pass=normalized_score >= 0.5,
-            reason=rating.reasoning,
-            label=rating.score,
-        )
-        return [result]
+        return [
+            EvaluationOutput(
+                score=normalized_score,
+                test_pass=normalized_score >= 0.5,
+                reason=rating.reasoning,
+                label=rating.score,
+            )
+        ]
 
     def _get_last_turn(self, evaluation_case: EvaluationData[InputT, OutputT]) -> TraceLevelInput:
         """Extract the most recent turn from the conversation for evaluation."""

src/strands_evals/evaluators/goal_success_rate_evaluator.py

@@ -1,4 +1,5 @@
 from enum import Enum
+from typing import cast
 
 from pydantic import BaseModel, Field
 from strands import Agent
@@ -53,29 +54,33 @@ def evaluate(self, evaluation_case: EvaluationData[InputT, OutputT]) -> list[Eva
         session_input = self._parse_trajectory(evaluation_case)
         prompt = self._format_prompt(session_input)
         evaluator_agent = Agent(model=self.model, system_prompt=self.system_prompt, callback_handler=None)
-        rating = evaluator_agent.structured_output(GoalSuccessRating, prompt)
+        result = evaluator_agent(prompt, structured_output_model=GoalSuccessRating)
+        rating = cast(GoalSuccessRating, result.structured_output)
         normalized_score = self._score_mapping[rating.score]
-        result = EvaluationOutput(
-            score=normalized_score,
-            test_pass=normalized_score >= 1.0,
-            reason=rating.reasoning,
-            label=rating.score,
-        )
-        return [result]
+        return [
+            EvaluationOutput(
+                score=normalized_score,
+                test_pass=normalized_score >= 1.0,
+                reason=rating.reasoning,
+                label=rating.score,
+            )
+        ]
 
     async def evaluate_async(self, evaluation_case: EvaluationData[InputT, OutputT]) -> list[EvaluationOutput]:
         session_input = self._parse_trajectory(evaluation_case)
         prompt = self._format_prompt(session_input)
         evaluator_agent = Agent(model=self.model, system_prompt=self.system_prompt, callback_handler=None)
-        rating = await evaluator_agent.structured_output_async(GoalSuccessRating, prompt)
+        result = await evaluator_agent.invoke_async(prompt, structured_output_model=GoalSuccessRating)
+        rating = cast(GoalSuccessRating, result.structured_output)
         normalized_score = self._score_mapping[rating.score]
-        result = EvaluationOutput(
-            score=normalized_score,
-            test_pass=normalized_score >= 1.0,
-            reason=rating.reasoning,
-            label=rating.score,
-        )
-        return [result]
+        return [
+            EvaluationOutput(
+                score=normalized_score,
+                test_pass=normalized_score >= 1.0,
+                reason=rating.reasoning,
+                label=rating.score,
+            )
+        ]
 
     def _format_prompt(self, session_input: SessionLevelInput) -> str:
         """Format evaluation prompt from session-level input."""

src/strands_evals/evaluators/harmfulness_evaluator.py

@@ -1,4 +1,5 @@
 from enum import Enum
+from typing import cast
 
 from pydantic import BaseModel, Field
 from strands import Agent
@@ -52,29 +53,33 @@ def evaluate(self, evaluation_case: EvaluationData[InputT, OutputT]) -> list[Eva
         parsed_input = self._get_last_turn(evaluation_case)
         prompt = self._format_prompt(parsed_input)
         evaluator_agent = Agent(model=self.model, system_prompt=self.system_prompt, callback_handler=None)
-        rating = evaluator_agent.structured_output(HarmfulnessRating, prompt)
+        result = evaluator_agent(prompt, structured_output_model=HarmfulnessRating)
+        rating = cast(HarmfulnessRating, result.structured_output)
         normalized_score = self._score_mapping[rating.score]
-        result = EvaluationOutput(
-            score=normalized_score,
-            test_pass=normalized_score == 1.0,
-            reason=rating.reasoning,
-            label=rating.score,
-        )
-        return [result]
+        return [
+            EvaluationOutput(
+                score=normalized_score,
+                test_pass=normalized_score == 1.0,
+                reason=rating.reasoning,
+                label=rating.score,
+            )
+        ]
 
     async def evaluate_async(self, evaluation_case: EvaluationData[InputT, OutputT]) -> list[EvaluationOutput]:
         parsed_input = self._get_last_turn(evaluation_case)
         prompt = self._format_prompt(parsed_input)
         evaluator_agent = Agent(model=self.model, system_prompt=self.system_prompt, callback_handler=None)
-        rating = await evaluator_agent.structured_output_async(HarmfulnessRating, prompt)
+        result = await evaluator_agent.invoke_async(prompt, structured_output_model=HarmfulnessRating)
+        rating = cast(HarmfulnessRating, result.structured_output)
         normalized_score = self._score_mapping[rating.score]
-        result = EvaluationOutput(
-            score=normalized_score,
-            test_pass=normalized_score == 1.0,
-            reason=rating.reasoning,
-            label=rating.score,
-        )
-        return [result]
+        return [
+            EvaluationOutput(
+                score=normalized_score,
+                test_pass=normalized_score == 1.0,
+                reason=rating.reasoning,
+                label=rating.score,
+            )
+        ]
 
     def _get_last_turn(self, evaluation_case: EvaluationData[InputT, OutputT]) -> TraceLevelInput:
         """Extract the most recent turn from the conversation for evaluation."""

src/strands_evals/evaluators/helpfulness_evaluator.py

@@ -1,4 +1,5 @@
 from enum import Enum
+from typing import cast
 
 from pydantic import BaseModel, Field
 from strands import Agent
@@ -65,29 +66,33 @@ def evaluate(self, evaluation_case: EvaluationData[InputT, OutputT]) -> list[Eva
         parsed_input = self._get_last_turn(evaluation_case)
         prompt = self._format_prompt(parsed_input)
         evaluator_agent = Agent(model=self.model, system_prompt=self.system_prompt, callback_handler=None)
-        rating = evaluator_agent.structured_output(HelpfulnessRating, prompt)
+        result = evaluator_agent(prompt, structured_output_model=HelpfulnessRating)
+        rating = cast(HelpfulnessRating, result.structured_output)
         normalized_score = self._score_mapping[rating.score]
-        result = EvaluationOutput(
-            score=normalized_score,
-            test_pass=normalized_score >= 0.5,
-            reason=rating.reasoning,
-            label=rating.score,
-        )
-        return [result]
+        return [
+            EvaluationOutput(
+                score=normalized_score,
+                test_pass=normalized_score >= 0.5,
+                reason=rating.reasoning,
+                label=rating.score,
+            )
+        ]
 
     async def evaluate_async(self, evaluation_case: EvaluationData[InputT, OutputT]) -> list[EvaluationOutput]:
         parsed_input = self._get_last_turn(evaluation_case)
         prompt = self._format_prompt(parsed_input)
         evaluator_agent = Agent(model=self.model, system_prompt=self.system_prompt, callback_handler=None)
-        rating = await evaluator_agent.structured_output_async(HelpfulnessRating, prompt)
+        result = await evaluator_agent.invoke_async(prompt, structured_output_model=HelpfulnessRating)
+        rating = cast(HelpfulnessRating, result.structured_output)
         normalized_score = self._score_mapping[rating.score]
-        result = EvaluationOutput(
-            score=normalized_score,
-            test_pass=normalized_score >= 0.5,
-            reason=rating.reasoning,
-            label=rating.score,
-        )
-        return [result]
+        return [
+            EvaluationOutput(
+                score=normalized_score,
+                test_pass=normalized_score >= 0.5,
+                reason=rating.reasoning,
+                label=rating.score,
+            )
+        ]
 
     def _get_last_turn(self, evaluation_case: EvaluationData[InputT, OutputT]) -> TraceLevelInput:
         """Extract the most recent turn from the conversation for evaluation."""

src/strands_evals/evaluators/interactions_evaluator.py

@@ -1,3 +1,5 @@
+from typing import cast
+
 from strands import Agent
 from strands.agent.conversation_manager import SlidingWindowConversationManager
 from strands.models.model import Model
@@ -198,8 +200,8 @@ def evaluate(self, evaluation_case: EvaluationData[InputT, OutputT]) -> list[Eva
         for i in range(num_interactions):
             is_last = i == num_interactions - 1
             evaluation_prompt = self._compose_prompt(evaluation_case, i, is_last)
-            result = evaluator_agent.structured_output(EvaluationOutput, evaluation_prompt)
-            results.append(result)
+            result = evaluator_agent(evaluation_prompt, structured_output_model=EvaluationOutput)
+            results.append(cast(EvaluationOutput, result.structured_output))
 
         return results
 
@@ -238,7 +240,7 @@ async def evaluate_async(self, evaluation_case: EvaluationData[InputT, OutputT])
         for i in range(num_interactions):
             is_last = i == num_interactions - 1
             evaluation_prompt = self._compose_prompt(evaluation_case, i, is_last)
-            result = await evaluator_agent.structured_output_async(EvaluationOutput, evaluation_prompt)
-            results.append(result)
+            result = await evaluator_agent.invoke_async(evaluation_prompt, structured_output_model=EvaluationOutput)
+            results.append(cast(EvaluationOutput, result.structured_output))
 
         return results

src/strands_evals/evaluators/output_evaluator.py

@@ -1,3 +1,5 @@
+from typing import cast
+
 from strands import Agent
 from strands.models.model import Model
 from typing_extensions import TypeVar, Union
@@ -51,8 +53,8 @@ def evaluate(self, evaluation_case: EvaluationData[InputT, OutputT]) -> list[Eva
         evaluation_prompt = compose_test_prompt(
             evaluation_case=evaluation_case, rubric=self.rubric, include_inputs=self.include_inputs
         )
-        result = evaluator_agent.structured_output(EvaluationOutput, evaluation_prompt)
-        return [result]
+        result = evaluator_agent(evaluation_prompt, structured_output_model=EvaluationOutput)
+        return [cast(EvaluationOutput, result.structured_output)]
 
     async def evaluate_async(self, evaluation_case: EvaluationData[InputT, OutputT]) -> list[EvaluationOutput]:
         """
@@ -68,5 +70,5 @@ async def evaluate_async(self, evaluation_case: EvaluationData[InputT, OutputT])
         evaluation_prompt = compose_test_prompt(
             evaluation_case=evaluation_case, rubric=self.rubric, include_inputs=self.include_inputs
         )
-        result = await evaluator_agent.structured_output_async(EvaluationOutput, evaluation_prompt)
-        return [result]
+        result = await evaluator_agent.invoke_async(evaluation_prompt, structured_output_model=EvaluationOutput)
+        return [cast(EvaluationOutput, result.structured_output)]

src/strands_evals/evaluators/tool_parameter_accuracy_evaluator.py

@@ -1,4 +1,5 @@
 from enum import Enum
+from typing import cast
 
 from pydantic import BaseModel, Field
 from strands import Agent
@@ -56,12 +57,17 @@ def evaluate(self, evaluation_case: EvaluationData[InputT, OutputT]) -> list[Eva
         for tool_input in tool_inputs:
             prompt = self._format_prompt(tool_input)
             evaluator_agent = Agent(model=self.model, system_prompt=self.system_prompt, callback_handler=None)
-            rating = evaluator_agent.structured_output(ToolParameterAccuracyRating, prompt)
+            result = evaluator_agent(prompt, structured_output_model=ToolParameterAccuracyRating)
+            rating = cast(ToolParameterAccuracyRating, result.structured_output)
             normalized_score = self._score_mapping[rating.score]
-            result = EvaluationOutput(
-                score=normalized_score, test_pass=normalized_score == 1.0, reason=rating.reasoning, label=rating.score
+            results.append(
+                EvaluationOutput(
+                    score=normalized_score,
+                    test_pass=normalized_score == 1.0,
+                    reason=rating.reasoning,
+                    label=rating.score,
+                )
             )
-            results.append(result)
 
         return results
 
@@ -72,12 +78,17 @@ async def evaluate_async(self, evaluation_case: EvaluationData[InputT, OutputT])
         for tool_input in tool_inputs:
             prompt = self._format_prompt(tool_input)
             evaluator_agent = Agent(model=self.model, system_prompt=self.system_prompt, callback_handler=None)
-            rating = await evaluator_agent.structured_output_async(ToolParameterAccuracyRating, prompt)
+            result = await evaluator_agent.invoke_async(prompt, structured_output_model=ToolParameterAccuracyRating)
+            rating = cast(ToolParameterAccuracyRating, result.structured_output)
             normalized_score = self._score_mapping[rating.score]
-            result = EvaluationOutput(
-                score=normalized_score, test_pass=normalized_score == 1.0, reason=rating.reasoning, label=rating.score
+            results.append(
+                EvaluationOutput(
+                    score=normalized_score,
+                    test_pass=normalized_score == 1.0,
+                    reason=rating.reasoning,
+                    label=rating.score,
+                )
             )
-            results.append(result)
 
         return results
 

src/strands_evals/evaluators/tool_selection_accuracy_evaluator.py

@@ -1,4 +1,5 @@
 from enum import Enum
+from typing import cast
 
 from pydantic import BaseModel, Field
 from strands import Agent
@@ -56,12 +57,17 @@ def evaluate(self, evaluation_case: EvaluationData[InputT, OutputT]) -> list[Eva
         for tool_input in tool_inputs:
             prompt = self._format_prompt(tool_input)
             evaluator_agent = Agent(model=self.model, system_prompt=self.system_prompt, callback_handler=None)
-            rating = evaluator_agent.structured_output(ToolSelectionRating, prompt)
+            result = evaluator_agent(prompt, structured_output_model=ToolSelectionRating)
+            rating = cast(ToolSelectionRating, result.structured_output)
             normalized_score = self._score_mapping[rating.score]
-            result = EvaluationOutput(
-                score=normalized_score, test_pass=normalized_score == 1.0, reason=rating.reasoning, label=rating.score
+            results.append(
+                EvaluationOutput(
+                    score=normalized_score,
+                    test_pass=normalized_score == 1.0,
+                    reason=rating.reasoning,
+                    label=rating.score,
+                )
             )
-            results.append(result)
 
         return results
 
@@ -72,12 +78,17 @@ async def evaluate_async(self, evaluation_case: EvaluationData[InputT, OutputT])
         for tool_input in tool_inputs:
             prompt = self._format_prompt(tool_input)
             evaluator_agent = Agent(model=self.model, system_prompt=self.system_prompt, callback_handler=None)
-            rating = await evaluator_agent.structured_output_async(ToolSelectionRating, prompt)
+            result = await evaluator_agent.invoke_async(prompt, structured_output_model=ToolSelectionRating)
+            rating = cast(ToolSelectionRating, result.structured_output)
             normalized_score = self._score_mapping[rating.score]
-            result = EvaluationOutput(
-                score=normalized_score, test_pass=normalized_score == 1.0, reason=rating.reasoning, label=rating.score
+            results.append(
+                EvaluationOutput(
+                    score=normalized_score,
+                    test_pass=normalized_score == 1.0,
+                    reason=rating.reasoning,
+                    label=rating.score,
+                )
             )
-            results.append(result)
 
         return results