{
  "audit_date": "2026-11-07",
  "input_scope": "All unique PDF/DOC/DOCX links embedded in 503 public posts dated 2024-01-01 through 2026-11-07 on mali-losinj.hr",
  "urls_attempted": 1537,
  "format_counts": {
    ".doc": 98,
    ".docx": 490,
    ".pdf": 949
  },
  "downloaded_urls": 1523,
  "unique_file_sha256": 1408,
  "exact_duplicate_url_copies": 115,
  "duplicate_hash_groups": 48,
  "status_counts": {
    "text_extracted": 1192,
    "no_or_minimal_text": 329,
    "parse_error": 2,
    "size_limit": 14
  },
  "text_extracted_unique_sha256": 1094,
  "keyword_candidate_urls": 121,
  "keyword_candidates_are_not_verified_mentions": true,
  "manual_scan_documents_reviewed": 3,
  "manual_scan_pages_reviewed": 6,
  "manual_scan_keys": [
    "cc253f4dfe70ee2c",
    "314aa85d35601268",
    "65c8156196b64ed6"
  ],
  "max_concurrent_downloads": 3,
  "max_file_bytes": 26214400,
  "download_timeout_seconds": 25,
  "parse_timeout_seconds": 35,
  "raw_download_bytes": 790308091,
  "verified_evidence_rows": 19,
  "unresolved_evidence_rows": 1,
  "post_mentions": {
    "directly_confirmed": 9,
    "contextually_confirmed": 1,
    "total": 10,
    "denominator": 503,
    "percentage": 1.9880715705765408,
    "hotel_only": 5,
    "by_year": {
      "2024": {
        "all": 141,
        "mentions": 7
      },
      "2025": {
        "all": 200,
        "mentions": 3
      },
      "2026_to_11_07": {
        "all": 162,
        "mentions": 0
      }
    }
  },
  "limitations": [
    "Text extracted means at least 80 characters, not proof every page is searchable; mixed scanned PDFs can hide further mentions.",
    "No full OCR across 329 minimal-text documents; three selected scans visually checked.",
    "14 files exceeded 25 MiB and were not downloaded; two PDF parsers failed.",
    "Exact-byte deduplication does not merge semantically identical reports with different cover pages.",
    "121 keyword candidate documents include personal names, geographical references and unresolved corporate aliases.",
    "Document dates may differ from parent post dates and upload URL directory dates.",
    "No claim of all web mentions, all institutional contacts, audience reach, ownership influence or favoritism."
  ],
  "evidence_count_scope": "19 odabranih izvora čiji su navedeni nalazi ručno potvrđeni, plus jedan nerazriješen identitet isključen iz pripisivanja grupi; nije broj svih relevantnih dokumenata.",
  "all_text_candidate_documents_manually_reviewed": false,
  "public_inventory_file": "institucije-svi-dokumenti.csv",
  "public_inventory_contains_fulltext": false
}