Add ADK compaction flushing persistence

[gvdongen]

This PR makes sure that the ADK integration supports event compaction:

• RestateEventsSummarizer: to wrap LLM summarization calls in durable steps
• Flushing compaction events in append_event to Restate

Example usage

agent = Agent(
    model="gemini-2.5-flash",
    name="assistant",
    instruction="You are a helpful assistant. Be concise and helpful.",
)
app = App(
    name=APP_NAME, 
    root_agent=agent, 
    plugins=[RestatePlugin()],
    events_compaction_config=EventsCompactionConfig(
        summarizer=RestateEventSummarizer.from_llm(llm=agent.canonical_model),
        compaction_interval=3,  
        overlap_size=1         
    ),
)
runner = Runner(app=app, session_service=RestateSessionService())

chat = restate.VirtualObject("Chat")


@chat.handler()
async def message(ctx: restate.ObjectContext, req: ChatMessage) -> str | None:
    events = runner.run_async(
        user_id=ctx.key(),
        session_id=req.session_id,
        new_message=Content(role="user", parts=[Part.from_text(text=req.message)]),
    )
    return await parse_agent_response(events)

Fixes #178

[github-actions]

Test Results

  7 files  ±0    7 suites  ±0   4m 9s ⏱️ +20s
 49 tests ±0   49 ✅ ±0  0 💤 ±0  0 ❌ ±0 
210 runs  ±0  210 ✅ ±0  0 💤 ±0  0 ❌ ±0 

Results for commit 70cc997. ± Comparison against base commit 9437df9.

♻️ This comment has been updated with latest results.