{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "https://dkharlanau.github.io/datasets/schema.json",
  "title": "Dzmitryi Kharlanau Dataset Byte Schema",
  "type": "object",
  "description": "Loose schema for JSON bytes used as reusable building blocks (articles, prompts, playbooks). The schema is intentionally permissive to avoid breaking domain-specific structures.",
  "required": ["meta"],
  "properties": {
    "id": {
      "type": "string",
      "description": "Primary identifier for a byte (preferred)."
    },
    "byte_id": {
      "type": "string",
      "description": "Alternate identifier field used by some datasets."
    },
    "title": {
      "type": "string",
      "description": "Human-readable title. May be inferred if missing.",
      "minLength": 1
    },
    "tags": {
      "type": "array",
      "items": { "type": "string" }
    },
    "meta": {
      "type": "object",
      "required": ["schema", "schema_version", "dataset", "source_project", "source_path", "creator", "attribution"],
      "properties": {
        "schema": { "type": "string", "const": "dkharlanau.dataset.byte" },
        "schema_version": { "type": "string" },
        "dataset": { "type": "string" },
        "source_project": { "type": "string" },
        "source_path": { "type": "string" },
        "canonical_url": { "type": "string" },
        "created_at_utc": { "type": "string" },
        "updated_at_utc": { "type": "string" },
        "title_inferred": { "type": "boolean" },
        "entity_type": { "type": "string" },
        "entity_subtype": { "type": "string" },
        "summary": { "type": "string" },
        "creator": {
          "type": "object",
          "required": ["name", "role"],
          "properties": {
            "name": { "type": "string" },
            "role": { "type": "string" },
            "website": { "type": "string" },
            "linkedin": { "type": "string" }
          },
          "additionalProperties": true
        },
        "links": {
          "type": "object",
          "properties": {
            "website": { "type": "string" },
            "linkedin": { "type": "string" }
          },
          "additionalProperties": true
        },
        "contact": {
          "type": "object",
          "properties": {
            "preferred": { "type": "string" },
            "linkedin": { "type": "string" }
          },
          "additionalProperties": true
        },
        "provenance": {
          "type": "object",
          "additionalProperties": true
        },
        "attribution": {
          "type": "object",
          "required": ["attribution_required", "preferred_citation"],
          "properties": {
            "attribution_required": { "type": "boolean" },
            "preferred_citation": { "type": "string" }
          },
          "additionalProperties": true
        },
        "doi": {
          "type": "object",
          "properties": {
            "concept": { "type": "string" },
            "version": { "type": "string" },
            "repository": { "type": "string" }
          },
          "additionalProperties": true
        },
        "license": {
          "type": "object",
          "properties": {
            "name": { "type": "string" },
            "spdx": { "type": "string" },
            "url": { "type": "string" }
          },
          "additionalProperties": true
        }
      },
      "additionalProperties": true
    }
  },
  "additionalProperties": true
}