fix: add optional confidence field to ReviewAlertEventInfo

[himorishige]

Summary

When using the Event Reviewer with the cv-event-detector example (e.g., GroundingDINO), the CV pipeline naturally includes a detection confidence score in the event payload. However, ReviewAlertEventInfo inherits from ViaBaseModel which sets extra="forbid", causing a Pydantic validation error:

Extra inputs are not permitted [type=extra_forbidden, input_value=0.85, input_type=float]

This PR adds confidence as an Optional[float] field to ReviewAlertEventInfo, so CV pipelines can include detection confidence scores without triggering validation errors.

Changes

• Added confidence: Optional[float] field to ReviewAlertEventInfo in vss_api_models.py
  • Default: None (fully backward-compatible)
  • Constraints: ge=0.0, le=1.0

Reproduction

1. Deploy Event Reviewer on DGX Spark (IS_SBSA=1)
2. Send an alert with event.confidence:

curl -X POST http://localhost:9080/api/v1/alerts \
  -H "Content-Type: application/json" \
  -d '{
    "id": "550e8400-e29b-41d4-a716-446655440000",
    "version": "1.0",
    "@timestamp": "2026-03-08T10:00:00.000Z",
    "sensor_id": "warehouse-cam-1",
    "video_path": "/tmp/alert-media-dir/clip.mp4",
    "confidence": 0.92,
    "alert": {
      "type": "safety_violation",
      "description": "Person without hard hat",
      "severity": "HIGH",
      "status": "REVIEW_PENDING"
    },
    "event": {
      "type": "object_detection",
      "description": "Person detected without PPE",
      "confidence": 0.85
    },
    "vss_params": {
      "vlm_params": {
        "prompt": "Is the person wearing a hard hat? Answer Yes or No.",
        "max_tokens": 200,
        "temperature": 0.3
      }
    }
  }'

3. Without this fix, event.confidence triggers: Extra inputs are not permitted

Environment

• VSS 2.4.1 (Event Reviewer deployment)
• DGX Spark (ARM64 SBSA), DGX OS 7.4.0
• CV pipeline: examples/cv-event-detector with GroundingDINO