fix: offload map parsing to worker thread to prevent event loop blocking

[allenporter]

Description

Offload CPU-intensive map parsing and rendering to worker threads via asyncio.to_thread to prevent blocking the asyncio event loop.

When devices are actively cleaning and returning map updates, parsing the raw map payload and rendering the map into PNG format runs synchronously on the asyncio event loop. In multi-device setups or frequent update cycles, this locks the event loop and leads to high latency spikes in downstream consumers like Home Assistant (see home-assistant/core#182656).

Changes

• In MapContentTrait.refresh(): offload self.converter.convert(response) to a worker thread using await asyncio.to_thread(self.converter.convert, response).
• In HomeTrait:
  • Factor cached map parsing into a helper function _parse_cached_home_map_content(...) and run it via await asyncio.to_thread(...) once, rather than running asyncio.to_thread per item in a loop.
  • Avoid redundant reparsing in _parse_cached_home_map_content when the cached bytes match the previously parsed raw_api_response.

Fixes home-assistant/core#182656

[Copilot]

Copilot review overview

🔵 Needs a closer look

Concurrent refreshes can apply stale map data; serialization or generation checks are needed.

Review effort: Lite
Findings: None

What changed in this PR

This pull request offloads V1 map parsing and rendering to worker threads to reduce asyncio event-loop blocking.

Changes:

• Runs map conversion via asyncio.to_thread.
• Adds threaded cached-map parsing.
• Reuses unchanged cached map content.

File	Description
roborock/​devices/​traits/​v1/​map_content.py	Offloads map conversion to a worker thread.
roborock/​devices/​traits/​v1/​home.py	Offloads cached parsing and avoids redundant reparsing.

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.