refactor: 全面重构腾讯云人脸识别插件

- 拆分 tencent_cloud_client.py（1311行）为 api_operations.py / image_processor.py / retry.py
- 更新 HA 废弃 API: SupportsResponse.OPTIONAL, async_unload_platforms, native_value, 移除 CONNECTION_CLASS
- 新增动态人员传感器管理（自动增/删）
- 新增 detect_face 服务定义（services.yaml）
- 移除 secret_key 持久化存储，增强安全性
- 完善错误映射前缀匹配 + 异常类型安全退避
- 修复 Coordinator 双重人员 ID 跟踪死代码
- 添加 _sanitize_error 脱敏，ImageCache LRU 缓存
- manifest.json 版本 2.1.0 + Pillow 依赖

const.py

@@ -7,6 +7,7 @@ CONF_SECRET_ID = "secret_id"
 CONF_SECRET_KEY = "secret_key"
 CONF_REGION = "region"
 CONF_PERSON_GROUP_ID = "person_group_id"
+CONF_SCAN_INTERVAL = "scan_interval"
 
 DEFAULT_REGION = "ap-shanghai"
 DEFAULT_PERSON_GROUP_ID = "Hass"
@@ -16,6 +17,7 @@ DEFAULT_MAX_USER_NUM = 5
 DEFAULT_QUALITY_CONTROL = 1
 DEFAULT_NEED_ROTATE_CHECK = 1
 DEFAULT_FACE_MATCH_THRESHOLD = 60.0
+DEFAULT_SCAN_INTERVAL = 300  # seconds (5 minutes)
 
 SERVICE_FACE_SEARCH = "face_search"
 SERVICE_DETECT_FACE = "detect_face"
@@ -45,6 +47,10 @@ ATTR_PERSON_TAG = "person_tag"
 
 EVENT_FACE_DETECTED = "face_detected"
 
+# Sensor identifiers
+SENSOR_STATUS = "status"
+SENSOR_PERSON = "person"
+
 ERROR_INVALID_CONFIG = "无效的配置"
 ERROR_API_ERROR = "API调用错误"
 ERROR_IMAGE_NOT_FOUND = "图片未找到"