Rdzleo dbdd304905 代码初始化:
本项目为触摸版项目代码复制而来,基于此版本进行按键功能的适配!
2026-03-23 11:14:56 +08:00

2.1 KiB

ChangeLog

v0.7.3 - 2026-02-25

Bug Fixes:

  • fix(task): change OpenAI task core affinity from core 1 to tskNO_AFFINITY to avoid error on ESP32-C5

v0.7.2 - 2026-02-02

Breaking Changes:

  • break(repo): migrate agent helper from service::helper::AgentOpenai to agent::helper::Openai
  • break(repo): change esp_peer dependency to private
  • break(base): rename lifecycle methods: on_start -> on_startup, on_stop -> on_shutdown
  • break(base): change on_wakeup() return type to bool
  • break(base): move encoder_feed_data_size to encoder.fetch_data_size in AudioConfig
  • break(base): replace general_event_wait_timeout_ms with operation_timeout in AgentAttributes

Enhancements:

  • feat(Kconfig): add BROOKESIA_AGENT_OPENAI_ENABLE_AUTO_REGISTER option for automatic plugin registration
  • feat(Kconfig): select ESP_WS_CLIENT_SEPARATE_TX_LOCK for WebSocket client
  • feat(base): inherit from service::ServiceBase with get_function_schemas(), get_event_schemas(), get_function_handlers() implementations

v0.7.1~1 - 2026-01-14

Enhancements:

  • feat(CMake): use linker option -u to register plugin instead of WHOLE_ARCHIVE

Bug Fixes

  • fix(agent): call try_load_data() in set_info()

v0.7.1 - 2025-12-25

Bug Fixes

  • fix(repo): add missing return statement in on_init()

v0.7.0 - 2025-12-24

Initial Release

  • feat(repo): add OpenAI agent implementation for ESP-Brookesia ecosystem
  • feat(repo): implement OpenAI Realtime API integration with WebRTC data channel for real-time voice and text interaction
  • feat(repo): support peer-to-peer communication using esp_peer with signaling and data channels
  • feat(repo): implement audio codec support with OPUS format at 16kHz sample rate and 24kbps bitrate
  • feat(repo): support multiple data channel message types (audio streams, text streams, function calls, session management)
  • feat(repo): support NVS-based data persistence for agent configuration
  • feat(repo): implement real-time audio input/output with low-latency voice interaction
  • feat(repo): integrate with agent manager framework for unified lifecycle management