schemas/training_example.schema.json

{
  "$schema": "http://json-schema.org/draft-07/schema#",
  "title": "MythosCoder Training Example",
  "description": "Schema for a single training example in the Mythos-Coder dataset. Each line in JSONL must be one valid JSON object matching this schema.",
  "type": "object",
  "required": [
    "id",
    "task_type",
    "difficulty",
    "language",
    "framework",
    "user_prompt",
    "repo_context",
    "failure_log",
    "investigation_steps",
    "plan",
    "solution",
    "verification",
    "lesson",
    "quality_score"
  ],
  "properties": {
    "id": {
      "type": "string",
      "description": "Unique identifier for this training example (e.g., 'react-import-bug-001')"
    },
    "task_type": {
      "type": "string",
      "enum": [
        "bug_fix",
        "feature_build",
        "website_vibe",
        "refactor",
        "migration",
        "terminal_debug",
        "test_writing",
        "ui_repair"
      ],
      "description": "Category of coding task"
    },
    "difficulty": {
      "type": "string",
      "enum": ["beginner", "intermediate", "advanced", "expert"],
      "description": "Difficulty level of the task"
    },
    "language": {
      "type": "string",
      "description": "Primary programming language (e.g., 'typescript', 'python', 'rust')"
    },
    "framework": {
      "type": "string",
      "description": "Framework or stack (e.g., 'react-vite', 'nextjs', 'flask', 'vanilla')"
    },
    "user_prompt": {
      "type": "string",
      "description": "The user's original request or prompt"
    },
    "repo_context": {
      "type": "string",
      "description": "Description of the codebase structure and relevant context"
    },
    "failure_log": {
      "type": "string",
      "description": "What went wrong - error messages, symptoms, or problem description"
    },
    "investigation_steps": {
      "type": "array",
      "items": {
        "type": "string"
      },
      "description": "Array of diagnostic steps taken to understand the problem"
    },
    "plan": {
      "type": "string",
      "description": "The approach or strategy to solve the problem"
    },
    "solution": {
      "type": "string",
      "description": "The actual fix or implementation"
    },
    "verification": {
      "type": "string",
      "description": "How the solution was tested and confirmed to work"
    },
    "lesson": {
      "type": "string",
      "description": "Key takeaway or engineering principle learned"
    },
    "quality_score": {
      "type": "integer",
      "minimum": 1,
      "maximum": 10,
      "description": "Quality rating from 1-10 based on completeness and educational value"
    }
  },
  "additionalProperties": false
}