{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "urn:orbiplex:schema:artifact-extraction-profile:v1",
  "title": "Artifact Extraction Profile v1",
  "description": "Operator-admitted, content-neutral and offline extraction profile. The profile selects one implementation and bounds parsing and candidate output; it carries no locator, source activation, credentials, artifact authority, admission, or publication authority.",
  "type": "object",
  "additionalProperties": false,
  "x-dia-workflow": "project",
  "x-dia-status": "draft",
  "x-dia-basis": [
    "doc/project/40-proposals/088-pull-based-artifact-acquisition.md"
  ],
  "required": [
    "schema",
    "schema/v",
    "profile/ref",
    "parser/id",
    "implementation/digest",
    "carrier/content-types",
    "selector/kinds",
    "candidate/framing",
    "candidate/cardinality",
    "limits"
  ],
  "properties": {
    "schema": { "const": "artifact-extraction-profile.v1" },
    "schema/v": { "const": 1 },
    "profile/ref": { "$ref": "#/$defs/ref" },
    "parser/id": { "type": "string", "minLength": 1, "maxLength": 256, "pattern": "^[a-z][a-z0-9._:-]*$" },
    "implementation/digest": { "$ref": "#/$defs/digest" },
    "carrier/content-types": {
      "type": "array",
      "minItems": 1,
      "maxItems": 32,
      "uniqueItems": true,
      "items": { "type": "string", "minLength": 3, "maxLength": 128, "pattern": "^[a-z0-9!#$&^_.+-]+/[a-z0-9!#$&^_.+-]+$" }
    },
    "selector/kinds": {
      "type": "array",
      "minItems": 1,
      "maxItems": 16,
      "uniqueItems": true,
      "items": { "enum": ["whole-resource", "whole-text", "line-range", "html-element-id", "mime-part"] }
    },
    "candidate/framing": { "enum": ["portable-package", "orbiplex-armored-package:v1"] },
    "candidate/cardinality": { "enum": ["exact-one", "bounded-many"] },
    "limits": {
      "type": "object",
      "additionalProperties": false,
      "required": [
        "carrier/bytes-max",
        "parse/depth-max",
        "nodes/max",
        "candidates/max",
        "candidate/bytes-max",
        "total-extracted/bytes-max",
        "temporary-storage/bytes-max",
        "duration/ms-max"
      ],
      "properties": {
        "carrier/bytes-max": { "type": "integer", "minimum": 1, "maximum": 67108864 },
        "parse/depth-max": { "type": "integer", "minimum": 1, "maximum": 256 },
        "nodes/max": { "type": "integer", "minimum": 1, "maximum": 1048576 },
        "candidates/max": { "type": "integer", "minimum": 1, "maximum": 64 },
        "candidate/bytes-max": { "type": "integer", "minimum": 1, "maximum": 8388608 },
        "total-extracted/bytes-max": { "type": "integer", "minimum": 1, "maximum": 67108864 },
        "temporary-storage/bytes-max": { "type": "integer", "minimum": 0, "maximum": 134217728 },
        "duration/ms-max": { "type": "integer", "minimum": 1, "maximum": 120000 }
      }
    }
  },
  "$defs": {
    "digest": { "type": "string", "pattern": "^sha256:[A-Za-z0-9_-]{43}$" },
    "ref": { "type": "string", "minLength": 3, "maxLength": 512, "pattern": "^[a-z][a-z0-9.-]*:[^\\s\\u0000-\\u001F\\u007F]+$" }
  }
}