feat: interaction.{output_text,output_image,output_audio,output_video}

google/genai/_interactions/types/interaction.py

@@ -33,7 +33,9 @@
 from .video_content import VideoContent
 from .._legacy_lyria import is_legacy_lyria_response_body
 from .webhook_config import WebhookConfig
+from .user_input_step import UserInputStep
 from .document_content import DocumentContent
+from .model_output_step import ModelOutputStep
 from .dynamic_agent_config import DynamicAgentConfig
 from .text_response_format import TextResponseFormat
 from .audio_response_format import AudioResponseFormat
@@ -212,3 +214,70 @@ def _coerce_outputs_to_steps(cls, values: Any) -> Any:
             def _coerce_outputs_to_steps(cls, data: Any) -> Any:
                 coerced, _ = cls._maybe_coerce_outputs(data)
                 return coerced
+
+    @property
+    def _last_model_output_steps(self) -> List[ModelOutputStep]:
+        if not self.steps:
+            return []
+        trailing: List[ModelOutputStep] = []
+        for step in reversed(self.steps):
+            if isinstance(step, ModelOutputStep):
+                trailing.append(step)
+            else:
+                break
+        trailing.reverse()
+        return trailing
+
+    @property
+    def output_text(self) -> str:
+        """Concatenated last consecutive text from the last consecutive model output steps."""
+        parts: List[str] = []
+        done = False
+        for step in reversed(self._last_model_output_steps):
+            if done:
+                break
+            if step.content:
+                for content in reversed(step.content):
+                    if isinstance(content, TextContent):
+                        parts.append(content.text)
+                    else:
+                        done = True
+                        break
+        parts.reverse()
+        return "".join(parts)
+
+    @property
+    def output_image(self) -> Optional[ImageContent]:
+        """The last image generated by the model in response to the current request."""
+        for step in reversed(self.steps):
+            if isinstance(step, UserInputStep):
+                break
+            if isinstance(step, ModelOutputStep) and step.content:
+                for content in reversed(step.content):
+                    if isinstance(content, ImageContent):
+                        return content
+        return None
+
+    @property
+    def output_audio(self) -> Optional[AudioContent]:
+        """The last audio generated by the model in response to the current request."""
+        for step in reversed(self.steps):
+            if isinstance(step, UserInputStep):
+                break
+            if isinstance(step, ModelOutputStep) and step.content:
+                for content in reversed(step.content):
+                    if isinstance(content, AudioContent):
+                        return content
+        return None
+
+    @property
+    def output_video(self) -> Optional[VideoContent]:
+        """The last video generated by the model in response to the current request."""
+        for step in reversed(self.steps):
+            if isinstance(step, UserInputStep):
+                break
+            if isinstance(step, ModelOutputStep) and step.content:
+                for content in reversed(step.content):
+                    if isinstance(content, VideoContent):
+                        return content
+        return None