Response Object Schemas

Overview

The testing-api follows a standardized response format to ensure consistency across the Supervised AI platform. All API responses are returned as JSON objects containing a status indicator, a descriptive message, and the primary payload.

By adhering to these schemas, client-side integrations can implement uniform error handling and data parsing logic.

Standard Success Response

A success response is returned for any request that is processed correctly by the server (typically HTTP status codes 200, 201, or 204).

Schema Definition

Usage Example

{
  "success": true,
  "message": "Test execution retrieved successfully",
  "data": {
    "id": "test_98765",
    "status": "completed",
    "accuracy": 0.98,
    "latency_ms": 120
  },
  "metadata": {
    "timestamp": "2023-10-27T10:00:00Z",
    "version": "v1.2.0"
  }
}

Standard Error Response

When a request fails due to client error (4xx) or server error (5xx), the API returns a structured error object. This allows developers to programmatically handle specific failure states.

Schema Definition

Usage Example

{
  "success": false,
  "error": {
    "code": "VALIDATION_FAILED",
    "message": "The provided model configuration is invalid.",
    "details": {
      "temperature": "Value must be between 0 and 1.",
      "max_tokens": "Field is required."
    }
  }
}

Paginated Response Schema

For endpoints that return lists of resources (e.g., test logs or dataset entries), the metadata object is populated with pagination details.

Metadata Fields

Usage Example

{
  "success": true,
  "message": "Datasets fetched successfully",
  "data": [
    { "id": "ds_01", "name": "Production_Eval" },
    { "id": "ds_02", "name": "Edge_Cases" }
  ],
  "metadata": {
    "total_count": 45,
    "page": 1,
    "limit": 2,
    "has_next": true
  }
}

Common Error Codes