Skip to content

Agent Mesh

examples: let chat-a2a consume file attachments

Tomas Tormo5d2cf3542f54
diff --git a/development/examples/chat-a2a/agent.py b/development/examples/chat-a2a/agent.py
index 95216b72..4d15c5c7 100644
--- a/development/examples/chat-a2a/agent.py
+++ b/development/examples/chat-a2a/agent.py
@@ -56,8 +56,18 @@ class ChatExecutor(AgentExecutor):
if data_parts:
# Text-first agent: structured payloads reach Gemini as a labeled block.
prompt += "\n[structured data]: " + json.dumps(data_parts)
+ gemini_parts = [prompt]
+ for part in context.message.parts:
+ if part.WhichOneof("content") != "raw":
+ continue
+ media = part.media_type or "application/octet-stream"
+ if media.startswith("text/"):
+ # Text attachments read best as prompt text, not opaque blobs.
+ gemini_parts[0] += f"\n[attached file {part.filename}]:\n" + part.raw.decode("utf-8", "replace")
+ else:
+ gemini_parts.append(types.Part.from_bytes(data=part.raw, mime_type=media))
started = time.monotonic()
- reply = await chat.send_message(prompt)
+ reply = await chat.send_message(gemini_parts)
print(
f"[chat] context={context.context_id} gemini took "
f"{time.monotonic() - started:.1f}s usage={reply.usage_metadata}",
@@ -82,7 +92,7 @@ agent_card = AgentCard(
description="Gemini-backed conversational agent; remembers the conversation per contextId",
version="0.1.0",
capabilities=AgentCapabilities(streaming=False),
- default_input_modes=["text/plain", "application/json"],
+ default_input_modes=["text/plain", "application/json", "application/pdf", "image/png", "image/jpeg"],
default_output_modes=["text/plain"],
skills=[
AgentSkill(