agno 2.0.0rc1__py3-none-any.whl → 2.0.1__py3-none-any.whl

agno/run/agent.py

@@ -5,7 +5,7 @@ from typing import Any, Dict, List, Optional, Sequence, Union
 
 from pydantic import BaseModel
 
-from agno.media import AudioArtifact, AudioResponse, File, ImageArtifact, VideoArtifact
+from agno.media import Audio, File, Image, Video
 from agno.models.message import Citations, Message
 from agno.models.metrics import Metrics
 from agno.models.response import ToolExecution
@@ -43,6 +43,8 @@ class RunEvent(str, Enum):
     output_model_response_started = "OutputModelResponseStarted"
     output_model_response_completed = "OutputModelResponseCompleted"
 
+    custom_event = "CustomEvent"
+
 
 @dataclass
 class BaseAgentRunEvent(BaseRunOutputEvent):
@@ -95,8 +97,8 @@ class RunContentEvent(BaseAgentRunEvent):
     content_type: str = "str"
     reasoning_content: Optional[str] = None
     citations: Optional[Citations] = None
-    response_audio: Optional[AudioResponse] = None  # Model audio response
-    image: Optional[ImageArtifact] = None  # Image attached to the response
+    response_audio: Optional[Audio] = None  # Model audio response
+    image: Optional[Image] = None  # Image attached to the response
     references: Optional[List[MessageReferences]] = None
     additional_input: Optional[List[Message]] = None
     reasoning_steps: Optional[List[ReasoningStep]] = None
@@ -117,10 +119,10 @@ class RunCompletedEvent(BaseAgentRunEvent):
     content_type: str = "str"
     reasoning_content: Optional[str] = None
     citations: Optional[Citations] = None
-    images: Optional[List[ImageArtifact]] = None  # Images attached to the response
-    videos: Optional[List[VideoArtifact]] = None  # Videos attached to the response
-    audio: Optional[List[AudioArtifact]] = None  # Audio attached to the response
-    response_audio: Optional[AudioResponse] = None  # Model audio response
+    images: Optional[List[Image]] = None  # Images attached to the response
+    videos: Optional[List[Video]] = None  # Videos attached to the response
+    audio: Optional[List[Audio]] = None  # Audio attached to the response
+    response_audio: Optional[Audio] = None  # Model audio response
     references: Optional[List[MessageReferences]] = None
     additional_input: Optional[List[Message]] = None
     reasoning_steps: Optional[List[ReasoningStep]] = None
@@ -201,9 +203,9 @@ class ToolCallCompletedEvent(BaseAgentRunEvent):
     event: str = RunEvent.tool_call_completed.value
     tool: Optional[ToolExecution] = None
     content: Optional[Any] = None
-    images: Optional[List[ImageArtifact]] = None  # Images produced by the tool call
-    videos: Optional[List[VideoArtifact]] = None  # Videos produced by the tool call
-    audio: Optional[List[AudioArtifact]] = None  # Audio produced by the tool call
+    images: Optional[List[Image]] = None  # Images produced by the tool call
+    videos: Optional[List[Video]] = None  # Videos produced by the tool call
+    audio: Optional[List[Audio]] = None  # Audio produced by the tool call
 
 
 @dataclass
@@ -226,6 +228,11 @@ class OutputModelResponseCompletedEvent(BaseAgentRunEvent):
     event: str = RunEvent.output_model_response_completed.value
 
 
+@dataclass
+class CustomEvent(BaseAgentRunEvent):
+    event: str = RunEvent.custom_event.value
+
+
 RunOutputEvent = Union[
     RunStartedEvent,
     RunContentEvent,
@@ -246,6 +253,7 @@ RunOutputEvent = Union[
     ParserModelResponseCompletedEvent,
     OutputModelResponseStartedEvent,
     OutputModelResponseCompletedEvent,
+    CustomEvent,
 ]
 
 
@@ -270,6 +278,7 @@ RUN_EVENT_TYPE_REGISTRY = {
     RunEvent.parser_model_response_completed.value: ParserModelResponseCompletedEvent,
     RunEvent.output_model_response_started.value: OutputModelResponseStartedEvent,
     RunEvent.output_model_response_completed.value: OutputModelResponseCompletedEvent,
+    RunEvent.custom_event.value: CustomEvent,
 }
 
 
@@ -297,9 +306,9 @@ class RunInput:
     """
 
     input_content: Optional[Union[str, List, Dict, Message, BaseModel, List[Message]]] = None
-    images: Optional[Sequence[ImageArtifact]] = None
-    videos: Optional[Sequence[VideoArtifact]] = None
-    audios: Optional[Sequence[AudioArtifact]] = None
+    images: Optional[Sequence[Image]] = None
+    videos: Optional[Sequence[Video]] = None
+    audios: Optional[Sequence[Audio]] = None
     files: Optional[Sequence[File]] = None
 
     def to_dict(self) -> Dict[str, Any]:
@@ -336,15 +345,15 @@ class RunInput:
         """Create RunInput from dictionary"""
         images = None
         if data.get("images"):
-            images = [ImageArtifact.model_validate(img_data) for img_data in data["images"]]
+            images = [Image.model_validate(img_data) for img_data in data["images"]]
 
         videos = None
         if data.get("videos"):
-            videos = [VideoArtifact.model_validate(vid_data) for vid_data in data["videos"]]
+            videos = [Video.model_validate(vid_data) for vid_data in data["videos"]]
 
         audios = None
         if data.get("audios"):
-            audios = [AudioArtifact.model_validate(aud_data) for aud_data in data["audios"]]
+            audios = [Audio.model_validate(aud_data) for aud_data in data["audios"]]
 
         files = None
         if data.get("files"):
@@ -380,10 +389,10 @@ class RunOutput:
 
     tools: Optional[List[ToolExecution]] = None
 
-    images: Optional[List[ImageArtifact]] = None  # Images attached to the response
-    videos: Optional[List[VideoArtifact]] = None  # Videos attached to the response
-    audio: Optional[List[AudioArtifact]] = None  # Audio attached to the response
-    response_audio: Optional[AudioResponse] = None  # Model audio response
+    images: Optional[List[Image]] = None  # Images attached to the response
+    videos: Optional[List[Video]] = None  # Videos attached to the response
+    audio: Optional[List[Audio]] = None  # Audio attached to the response
+    response_audio: Optional[Audio] = None  # Model audio response
 
     # Input media and messages from user
     input: Optional[RunInput] = None
@@ -478,7 +487,7 @@ class RunOutput:
         if self.images is not None:
             _dict["images"] = []
             for img in self.images:
-                if isinstance(img, ImageArtifact):
+                if isinstance(img, Image):
                     _dict["images"].append(img.to_dict())
                 else:
                     _dict["images"].append(img)
@@ -486,7 +495,7 @@ class RunOutput:
         if self.videos is not None:
             _dict["videos"] = []
             for vid in self.videos:
-                if isinstance(vid, VideoArtifact):
+                if isinstance(vid, Video):
                     _dict["videos"].append(vid.to_dict())
                 else:
                     _dict["videos"].append(vid)
@@ -494,13 +503,13 @@ class RunOutput:
         if self.audio is not None:
             _dict["audio"] = []
             for aud in self.audio:
-                if isinstance(aud, AudioArtifact):
+                if isinstance(aud, Audio):
                     _dict["audio"].append(aud.to_dict())
                 else:
                     _dict["audio"].append(aud)
 
         if self.response_audio is not None:
-            if isinstance(self.response_audio, AudioResponse):
+            if isinstance(self.response_audio, Audio):
                 _dict["response_audio"] = self.response_audio.to_dict()
             else:
                 _dict["response_audio"] = self.response_audio
@@ -556,16 +565,16 @@ class RunOutput:
         tools = [ToolExecution.from_dict(tool) for tool in tools] if tools else None
 
         images = data.pop("images", [])
-        images = [ImageArtifact.model_validate(image) for image in images] if images else None
+        images = [Image.model_validate(image) for image in images] if images else None
 
         videos = data.pop("videos", [])
-        videos = [VideoArtifact.model_validate(video) for video in videos] if videos else None
+        videos = [Video.model_validate(video) for video in videos] if videos else None
 
         audio = data.pop("audio", [])
-        audio = [AudioArtifact.model_validate(audio) for audio in audio] if audio else None
+        audio = [Audio.model_validate(audio) for audio in audio] if audio else None
 
         response_audio = data.pop("response_audio", None)
-        response_audio = AudioResponse.model_validate(response_audio) if response_audio else None
+        response_audio = Audio.model_validate(response_audio) if response_audio else None
 
         input_data = data.pop("input", None)
         input_obj = None

agno/run/base.py

@@ -4,7 +4,7 @@ from typing import Any, Dict
 
 from pydantic import BaseModel
 
-from agno.media import AudioArtifact, AudioResponse, ImageArtifact, VideoArtifact
+from agno.media import Audio, Image, Video
 from agno.models.message import Citations, Message, MessageReferences
 from agno.models.metrics import Metrics
 from agno.models.response import ToolExecution
@@ -60,21 +60,15 @@ class BaseRunOutputEvent:
         if hasattr(self, "images") and self.images is not None:
             _dict["images"] = []
             for img in self.images:
-                if isinstance(img, ImageArtifact):
+                if isinstance(img, Image):
                     _dict["images"].append(img.to_dict())
                 else:
                     _dict["images"].append(img)
 
-        if hasattr(self, "image") and self.image is not None:
-            if isinstance(self.image, ImageArtifact):
-                _dict["image"] = self.image.to_dict()
-            else:
-                _dict["image"] = self.image
-
         if hasattr(self, "videos") and self.videos is not None:
             _dict["videos"] = []
             for vid in self.videos:
-                if isinstance(vid, VideoArtifact):
+                if isinstance(vid, Video):
                     _dict["videos"].append(vid.to_dict())
                 else:
                     _dict["videos"].append(vid)
@@ -82,13 +76,13 @@ class BaseRunOutputEvent:
         if hasattr(self, "audio") and self.audio is not None:
             _dict["audio"] = []
             for aud in self.audio:
-                if isinstance(aud, AudioArtifact):
+                if isinstance(aud, Audio):
                     _dict["audio"].append(aud.to_dict())
                 else:
                     _dict["audio"].append(aud)
 
         if hasattr(self, "response_audio") and self.response_audio is not None:
-            if isinstance(self.response_audio, AudioResponse):
+            if isinstance(self.response_audio, Audio):
                 _dict["response_audio"] = self.response_audio.to_dict()
             else:
                 _dict["response_audio"] = self.response_audio
@@ -140,23 +134,19 @@ class BaseRunOutputEvent:
 
         images = data.pop("images", None)
         if images:
-            data["images"] = [ImageArtifact.model_validate(image) for image in images]
-
-        image = data.pop("image", None)
-        if image:
-            data["image"] = ImageArtifact.model_validate(image)
+            data["images"] = [Image.model_validate(image) for image in images]
 
         videos = data.pop("videos", None)
         if videos:
-            data["videos"] = [VideoArtifact.model_validate(video) for video in videos]
+            data["videos"] = [Video.model_validate(video) for video in videos]
 
         audio = data.pop("audio", None)
         if audio:
-            data["audio"] = [AudioArtifact.model_validate(audio) for audio in audio]
+            data["audio"] = [Audio.model_validate(audio) for audio in audio]
 
         response_audio = data.pop("response_audio", None)
         if response_audio:
-            data["response_audio"] = AudioResponse.model_validate(response_audio)
+            data["response_audio"] = Audio.model_validate(response_audio)
 
         additional_input = data.pop("additional_input", None)
         if additional_input is not None:

agno/run/team.py

@@ -1,11 +1,11 @@
 from dataclasses import asdict, dataclass, field
 from enum import Enum
 from time import time
-from typing import Any, Dict, List, Optional, Union
+from typing import Any, Dict, List, Optional, Sequence, Union
 
 from pydantic import BaseModel
 
-from agno.media import AudioArtifact, AudioResponse, ImageArtifact, VideoArtifact
+from agno.media import Audio, File, Image, Video
 from agno.models.message import Citations, Message
 from agno.models.metrics import Metrics
 from agno.models.response import ToolExecution
@@ -40,6 +40,8 @@ class TeamRunEvent(str, Enum):
     output_model_response_started = "TeamOutputModelResponseStarted"
     output_model_response_completed = "TeamOutputModelResponseCompleted"
 
+    custom_event = "CustomEvent"
+
 
 @dataclass
 class BaseTeamRunEvent(BaseRunOutputEvent):
@@ -96,8 +98,8 @@ class RunContentEvent(BaseTeamRunEvent):
     content_type: str = "str"
     reasoning_content: Optional[str] = None
     citations: Optional[Citations] = None
-    response_audio: Optional[AudioResponse] = None  # Model audio response
-    image: Optional[ImageArtifact] = None  # Image attached to the response
+    response_audio: Optional[Audio] = None  # Model audio response
+    image: Optional[Image] = None  # Image attached to the response
     references: Optional[List[MessageReferences]] = None
     additional_input: Optional[List[Message]] = None
     reasoning_steps: Optional[List[ReasoningStep]] = None
@@ -118,10 +120,10 @@ class RunCompletedEvent(BaseTeamRunEvent):
     content_type: str = "str"
     reasoning_content: Optional[str] = None
     citations: Optional[Citations] = None
-    images: Optional[List[ImageArtifact]] = None  # Images attached to the response
-    videos: Optional[List[VideoArtifact]] = None  # Videos attached to the response
-    audio: Optional[List[AudioArtifact]] = None  # Audio attached to the response
-    response_audio: Optional[AudioResponse] = None  # Model audio response
+    images: Optional[List[Image]] = None  # Images attached to the response
+    videos: Optional[List[Video]] = None  # Videos attached to the response
+    audio: Optional[List[Audio]] = None  # Audio attached to the response
+    response_audio: Optional[Audio] = None  # Model audio response
     references: Optional[List[MessageReferences]] = None
     additional_input: Optional[List[Message]] = None
     reasoning_steps: Optional[List[ReasoningStep]] = None
@@ -188,9 +190,9 @@ class ToolCallCompletedEvent(BaseTeamRunEvent):
     event: str = TeamRunEvent.tool_call_completed.value
     tool: Optional[ToolExecution] = None
     content: Optional[Any] = None
-    images: Optional[List[ImageArtifact]] = None  # Images produced by the tool call
-    videos: Optional[List[VideoArtifact]] = None  # Videos produced by the tool call
-    audio: Optional[List[AudioArtifact]] = None  # Audio produced by the tool call
+    images: Optional[List[Image]] = None  # Images produced by the tool call
+    videos: Optional[List[Video]] = None  # Videos produced by the tool call
+    audio: Optional[List[Audio]] = None  # Audio produced by the tool call
 
 
 @dataclass
@@ -213,6 +215,11 @@ class OutputModelResponseCompletedEvent(BaseTeamRunEvent):
     event: str = TeamRunEvent.output_model_response_completed.value
 
 
+@dataclass
+class CustomEvent(BaseTeamRunEvent):
+    event: str = TeamRunEvent.custom_event.value
+
+
 TeamRunOutputEvent = Union[
     RunStartedEvent,
     RunContentEvent,
@@ -231,6 +238,7 @@ TeamRunOutputEvent = Union[
     ParserModelResponseCompletedEvent,
     OutputModelResponseStartedEvent,
     OutputModelResponseCompletedEvent,
+    CustomEvent,
 ]
 
 # Map event string to dataclass for team events
@@ -252,6 +260,7 @@ TEAM_RUN_EVENT_TYPE_REGISTRY = {
     TeamRunEvent.parser_model_response_completed.value: ParserModelResponseCompletedEvent,
     TeamRunEvent.output_model_response_started.value: OutputModelResponseStartedEvent,
     TeamRunEvent.output_model_response_completed.value: OutputModelResponseCompletedEvent,
+    TeamRunEvent.custom_event.value: CustomEvent,
 }
 
 
@@ -266,6 +275,76 @@ def team_run_output_event_from_dict(data: dict) -> BaseTeamRunEvent:
     return event_class.from_dict(data)  # type: ignore
 
 
+@dataclass
+class TeamRunInput:
+    """Container for the raw input data passed to Agent.run().
+    This captures the original input exactly as provided by the user,
+    separate from the processed messages that go to the model.
+    Attributes:
+        input_content: The literal input message/content passed to run()
+        images: Images directly passed to run()
+        videos: Videos directly passed to run()
+        audios: Audio files directly passed to run()
+        files: Files directly passed to run()
+    """
+
+    input_content: Optional[Union[str, List, Dict, Message, BaseModel, List[Message]]] = None
+    images: Optional[Sequence[Image]] = None
+    videos: Optional[Sequence[Video]] = None
+    audios: Optional[Sequence[Audio]] = None
+    files: Optional[Sequence[File]] = None
+
+    def to_dict(self) -> Dict[str, Any]:
+        """Convert to dictionary representation"""
+        result: Dict[str, Any] = {}
+
+        if self.input_content is not None:
+            if isinstance(self.input_content, (str)):
+                result["input_content"] = self.input_content
+            elif isinstance(self.input_content, BaseModel):
+                result["input_content"] = self.input_content.model_dump(exclude_none=True)
+            elif isinstance(self.input_content, Message):
+                result["input_content"] = self.input_content.to_dict()
+            elif (
+                isinstance(self.input_content, list)
+                and self.input_content
+                and isinstance(self.input_content[0], Message)
+            ):
+                result["input_content"] = [m.to_dict() for m in self.input_content]
+            else:
+                result["input_content"] = self.input_content
+
+        if self.images:
+            result["images"] = [img.to_dict() for img in self.images]
+        if self.videos:
+            result["videos"] = [vid.to_dict() for vid in self.videos]
+        if self.audios:
+            result["audios"] = [aud.to_dict() for aud in self.audios]
+
+        return result
+
+    @classmethod
+    def from_dict(cls, data: Dict[str, Any]) -> "TeamRunInput":
+        """Create TeamRunInput from dictionary"""
+        images = None
+        if data.get("images"):
+            images = [Image.model_validate(img_data) for img_data in data["images"]]
+
+        videos = None
+        if data.get("videos"):
+            videos = [Video.model_validate(vid_data) for vid_data in data["videos"]]
+
+        audios = None
+        if data.get("audios"):
+            audios = [Audio.model_validate(aud_data) for aud_data in data["audios"]]
+
+        files = None
+        if data.get("files"):
+            files = [File.model_validate(file_data) for file_data in data["files"]]
+
+        return cls(input_content=data.get("input_content"), images=images, videos=videos, audios=audios, files=files)
+
+
 @dataclass
 class TeamRunOutput:
     """Response returned by Team.run() functions"""
@@ -287,11 +366,14 @@ class TeamRunOutput:
 
     tools: Optional[List[ToolExecution]] = None
 
-    images: Optional[List[ImageArtifact]] = None  # Images from member runs
-    videos: Optional[List[VideoArtifact]] = None  # Videos from member runs
-    audio: Optional[List[AudioArtifact]] = None  # Audio from member runs
+    images: Optional[List[Image]] = None  # Images from member runs
+    videos: Optional[List[Video]] = None  # Videos from member runs
+    audio: Optional[List[Audio]] = None  # Audio from member runs
+
+    response_audio: Optional[Audio] = None  # Model audio response
 
-    response_audio: Optional[AudioResponse] = None  # Model audio response
+    # Input media and messages from user
+    input: Optional[TeamRunInput] = None
 
     reasoning_content: Optional[str] = None
 
@@ -401,6 +483,9 @@ class TeamRunOutput:
                 else:
                     _dict["tools"].append(tool)
 
+        if self.input is not None:
+            _dict["input"] = self.input.to_dict()
+
         return _dict
 
     def to_json(self) -> str:
@@ -454,19 +539,24 @@ class TeamRunOutput:
             references = [MessageReferences.model_validate(reference) for reference in references]
 
         images = data.pop("images", [])
-        images = [ImageArtifact.model_validate(image) for image in images] if images else None
+        images = [Image.model_validate(image) for image in images] if images else None
 
         videos = data.pop("videos", [])
-        videos = [VideoArtifact.model_validate(video) for video in videos] if videos else None
+        videos = [Video.model_validate(video) for video in videos] if videos else None
 
         audio = data.pop("audio", [])
-        audio = [AudioArtifact.model_validate(audio) for audio in audio] if audio else None
+        audio = [Audio.model_validate(audio) for audio in audio] if audio else None
 
         tools = data.pop("tools", [])
         tools = [ToolExecution.from_dict(tool) for tool in tools] if tools else None
 
         response_audio = data.pop("response_audio", None)
-        response_audio = AudioResponse.model_validate(response_audio) if response_audio else None
+        response_audio = Audio.model_validate(response_audio) if response_audio else None
+
+        input_data = data.pop("input", None)
+        input_obj = None
+        if input_data:
+            input_obj = TeamRunInput.from_dict(input_data)
 
         metrics = data.pop("metrics", None)
         if metrics:
@@ -487,6 +577,7 @@ class TeamRunOutput:
             videos=videos,
             audio=audio,
             response_audio=response_audio,
+            input=input_obj,
             citations=citations,
             tools=tools,
             events=events,