reggenomedocumentmetadata.schema.json

{
  "Reggenomedocumentmetadata": {
    "@type": "Reggenomedocumentmetadata",
    "schemaVersion": 1,
    "document_id": {
      "@type": "Document_id"
    },
    "title": "String",
    "source_urls": [
      {
        "source_urls(item)": "URL"
      }
    ],
    "publisher": {
      "@type": "Publisher",
      "id": "String",
      "name": "String",
      "jurisdiction": "String"
    },
    "doctype": "String",
    "published": "Datetime",
    "last_updated": "Datetime",
    "crawled": "Datetime",
    "processed": "Datetime",
    "sector": {
      "@type": "Sector",
      "level_1": "String",
      "level_2": "String",
      "level_3": "String",
      "level_4": "String"
    },
    "cyber_score": "Decimal",
    "authoritative": [
      {
        "authoritative(item)": "String"
      }
    ],
    "content_type": "String",
    "restriction": "String",
    "copyright_notice": "String"
  }
}

{
  "Reggenomedocumentmetadata": {
    "@type": "Reggenomedocumentmetadata",
    "schemaVersion": 1,
    "description": "Schema describing every property RegGenome emits for a corpus document in all-documents.json. Derived by walking 8,865 records on 2026-06-09. Source file: Nexus/06 – Engineering/Data Files/all-documents.json. Every record is a flat object at the top level (no nested arrays of records). All fields except `processed` and the deeper sector levels are present in 100% of records.",
    "document_id": {
      "@type": "Document_id",
      "description": "Stable RegGenome identifier. Acts as the corpus-wide primary key and is the basename of the original-format file (e.g., {document_id}.pdf) and of the markdown conversion ({document_id}.md)."
    },
    "title": {
      "description": "Document title as published. May carry leading whitespace or trailing newlines — normalize before bibliographic use. Federal Register documents prefix with 'FR Doc. {year}-{number}:'.",
      "type": "String"
    },
    "source_urls": [
      {
        "description": "Authoritative URLs for the document at its publisher's site. Always exactly one element in the observed corpus. This is the canonical URL — use it as the BibTeX `url` field, NOT the package storage URL.",
        "source_urls(item)": {
          "type": "URL"
        }
      }
    ],
    "publisher": {
      "@type": "Publisher",
      "description": "The issuing authority. The same publisher may have many `id` aliases across the corpus, but `name` is stable.",
      "id": {
        "description": "Slug identifier, kebab-case. Useful for filtering and joining.",
        "type": "String"
      },
      "name": {
        "description": "The name of the item or record.",
        "type": "String"
      },
      "jurisdiction": {
        "description": "ISO 3166-1 alpha-2 country code, plus 'International Standard-Setters' for cross-border bodies (IOSCO, BIS, FATF, etc.). Allowed values: AU, CA, DE, GB, International Standard-Setters, NZ, US.",
        "type": "String"
      }
    },
    "doctype": {
      "description": "RegGenome's controlled vocabulary for document genre. Drives BibTeX entry-type selection (Guidance/Research/Technical Standards → @techreport when publisher is NIST/CISA/etc.; everything else → @misc). Allowed values: Codified rulebooks, Consultation Responses Received, Consultations and Calls for Input, Corporate Actions, Operations and Reporting, Firm-level actions and enforcement, Guidance and FAQs, Individual Guidance, Laws, Legislative Codes, Lists, Forms and Tables, Other, Policies and Decisions, … (16 total).",
      "type": "String"
    },
    "published": {
      "description": "ISO 8601 timestamp of original publication. The year portion drives the BibTeX `year` field.",
      "type": "Datetime"
    },
    "last_updated": {
      "description": "ISO 8601. RegGenome's last metadata refresh of this record (not necessarily the document's last revision).",
      "type": "Datetime"
    },
    "crawled": {
      "description": "ISO 8601. When RegGenome fetched the source URL. Use as BibTeX `urldate`.",
      "type": "Datetime"
    },
    "processed": {
      "description": "ISO 8601. When RegGenome's pipeline last reprocessed the record. Null for 29 records (0.3%).",
      "nullable": true,
      "type": "Datetime"
    },
    "sector": {
      "@type": "Sector",
      "description": "Four-level industry classification. Level 1 is always populated; lower levels narrow progressively and are null when the document is broader than that level.",
      "level_1": {
        "description": "Top-level industry. 18 distinct values across the corpus.",
        "nullable": true,
        "type": "String"
      },
      "level_2": {
        "description": "Mid-tier sub-industry. Null for 33.8% of records.",
        "nullable": true,
        "type": "String"
      },
      "level_3": {
        "description": "Narrow sub-industry. Null for 38.1%.",
        "nullable": true,
        "type": "String"
      },
      "level_4": {
        "description": "Most specific sub-industry. Null for 49.6%.",
        "nullable": true,
        "type": "String"
      }
    },
    "cyber_score": {
      "description": "RegGenome's cyber-relevance signal, 0.0–1.0 (higher = more cyber-relevant). Source heuristic not documented by RegGenome. Useful for downstream filtering when building cyber-only corpora.",
      "type": "Decimal"
    },
    "authoritative": [
      {
        "description": "Tags marking the document as authoritative on a topic. Empty in 99.88% of records; when populated, the only observed value is 'cyber' (11 records). Treat as a hint, not a primary filter.",
        "authoritative(item)": {
          "description": "Allowed values: cyber.",
          "type": "String"
        }
      }
    ],
    "content_type": {
      "description": "MIME type of the original file. Constant across the observed corpus. Allowed values: application/pdf.",
      "type": "String"
    },
    "restriction": {
      "description": "Distribution restriction. Constant 'unrestricted' across the observed corpus. Allowed values: unrestricted.",
      "type": "String"
    },
    "copyright_notice": {
      "description": "Free-text copyright line as published. Often includes the publisher and year.",
      "type": "String"
    }
  }
}