Skip to content

A programming framework for agentic AI

Have Console print multimodal messages nicely. (#4179)

afourney4073dd4efb53
diff --git a/python/packages/autogen-agentchat/src/autogen_agentchat/task/_console.py b/python/packages/autogen-agentchat/src/autogen_agentchat/task/_console.py
index 71344eb7a..7e4d1cab9 100644
--- a/python/packages/autogen-agentchat/src/autogen_agentchat/task/_console.py
+++ b/python/packages/autogen-agentchat/src/autogen_agentchat/task/_console.py
@@ -1,11 +1,11 @@
import sys
import time
-from typing import AsyncGenerator
+from typing import AsyncGenerator, List
from autogen_core.components.models import RequestUsage
from autogen_agentchat.base import TaskResult
-from autogen_agentchat.messages import AgentMessage
+from autogen_agentchat.messages import AgentMessage, MultiModalMessage
async def Console(stream: AsyncGenerator[AgentMessage | TaskResult, None]) -> None:
@@ -27,9 +27,22 @@ async def Console(stream: AsyncGenerator[AgentMessage | TaskResult, None]) -> No
)
sys.stdout.write(output)
else:
- output = f"{'-' * 10} {message.source} {'-' * 10}\n{message.content}\n"
+ output = f"{'-' * 10} {message.source} {'-' * 10}\n{_message_to_str(message)}\n"
if message.models_usage:
output += f"[Prompt tokens: {message.models_usage.prompt_tokens}, Completion tokens: {message.models_usage.completion_tokens}]\n"
total_usage.completion_tokens += message.models_usage.completion_tokens
total_usage.prompt_tokens += message.models_usage.prompt_tokens
sys.stdout.write(output)
+
+
+def _message_to_str(message: AgentMessage) -> str:
+ if isinstance(message, MultiModalMessage):
+ result: List[str] = []
+ for c in message.content:
+ if isinstance(c, str):
+ result.append(c)
+ else:
+ result.append("<image>")
+ return "\n".join(result)
+ else:
+ return f"{message.content}"