{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "https://example.invalid/ai-agent-engineer/document-manifest.schema.json",
  "title": "离线文档解析清单",
  "type": "object",
  "additionalProperties": false,
  "required": [
    "schema_version",
    "parser",
    "config",
    "config_sha256",
    "root",
    "documents",
    "summary",
    "issues"
  ],
  "properties": {
    "schema_version": { "const": "2.0" },
    "parser": {
      "type": "object",
      "additionalProperties": false,
      "required": ["name", "version"],
      "properties": {
        "name": { "const": "stdlib-document-inspector" },
        "version": { "type": "string", "minLength": 1 }
      }
    },
    "config": { "type": "object" },
    "config_sha256": { "$ref": "#/$defs/sha256" },
    "root": { "const": "." },
    "documents": {
      "type": "array",
      "items": { "$ref": "#/$defs/document" }
    },
    "summary": {
      "type": "object",
      "additionalProperties": false,
      "required": [
        "gate",
        "discovered_file_count",
        "processed_file_count",
        "consumed_bytes",
        "status_counts",
        "error_count",
        "warning_count"
      ],
      "properties": {
        "gate": { "enum": ["pass", "review_required", "fail"] },
        "discovered_file_count": { "type": "integer", "minimum": 0 },
        "processed_file_count": { "type": "integer", "minimum": 0 },
        "consumed_bytes": { "type": "integer", "minimum": 0 },
        "status_counts": {
          "type": "object",
          "additionalProperties": false,
          "required": ["parsed", "external_adapter_required", "rejected"],
          "properties": {
            "parsed": { "type": "integer", "minimum": 0 },
            "external_adapter_required": { "type": "integer", "minimum": 0 },
            "rejected": { "type": "integer", "minimum": 0 }
          }
        },
        "error_count": { "type": "integer", "minimum": 0 },
        "warning_count": { "type": "integer", "minimum": 0 }
      }
    },
    "issues": {
      "type": "array",
      "items": { "$ref": "#/$defs/issue" }
    }
  },
  "$defs": {
    "sha256": {
      "type": "string",
      "pattern": "^[0-9a-f]{64}$"
    },
    "issue": {
      "type": "object",
      "additionalProperties": false,
      "required": ["code", "severity", "message"],
      "properties": {
        "code": { "type": "string", "minLength": 1 },
        "severity": { "enum": ["info", "warning", "error"] },
        "message": { "type": "string", "minLength": 1 }
      }
    },
    "element": {
      "type": "object",
      "additionalProperties": false,
      "required": [
        "element_id",
        "kind",
        "text",
        "text_sha256",
        "order",
        "location",
        "section_path",
        "attributes"
      ],
      "properties": {
        "element_id": {
          "type": "string",
          "pattern": "^elm_[0-9a-f]{64}$"
        },
        "kind": { "type": "string", "minLength": 1 },
        "text": { "type": "string" },
        "text_sha256": { "$ref": "#/$defs/sha256" },
        "order": { "type": "integer", "minimum": 1 },
        "location": {
          "type": "object",
          "additionalProperties": false,
          "required": ["coordinate_space", "line_start", "line_end"],
          "properties": {
            "coordinate_space": {
              "const": "normalized-text-lines-1-based-inclusive-v1"
            },
            "line_start": { "type": "integer", "minimum": 1 },
            "line_end": { "type": "integer", "minimum": 1 }
          }
        },
        "section_path": {
          "type": "array",
          "items": { "type": "string" }
        },
        "attributes": { "type": "object" }
      }
    },
    "document": {
      "type": "object",
      "additionalProperties": false,
      "required": [
        "relative_path",
        "source_id",
        "raw_sha256",
        "size_bytes",
        "extension_media_type",
        "detected_media_type",
        "detection_method",
        "encoding",
        "status",
        "parser",
        "parser_version",
        "parse_revision_sha256",
        "elements",
        "issues"
      ],
      "properties": {
        "relative_path": { "type": "string", "minLength": 1 },
        "source_id": { "type": "string", "minLength": 1 },
        "raw_sha256": {
          "oneOf": [{ "$ref": "#/$defs/sha256" }, { "type": "null" }]
        },
        "size_bytes": { "type": "integer", "minimum": 0 },
        "extension_media_type": { "type": ["string", "null"] },
        "detected_media_type": { "type": ["string", "null"] },
        "detection_method": { "type": "string", "minLength": 1 },
        "encoding": { "type": ["string", "null"] },
        "status": { "enum": ["parsed", "external_adapter_required", "rejected"] },
        "parser": { "type": ["string", "null"] },
        "parser_version": { "type": ["string", "null"] },
        "parse_revision_sha256": {
          "oneOf": [{ "$ref": "#/$defs/sha256" }, { "type": "null" }]
        },
        "elements": {
          "type": "array",
          "items": { "$ref": "#/$defs/element" }
        },
        "issues": {
          "type": "array",
          "items": { "$ref": "#/$defs/issue" }
        }
      }
    }
  }
}