disclosure-bureau/web/lib/cases.ts

/**
 * cases.ts — server-only loader for the case-file library.
 *
 * Reads /data/ufo/case/reports/*.md, parses frontmatter, derives the
 * magazine title from the body H1, picks the locale-preferred opening
 * paragraph, and infers two filter dimensions from the slug:
 *
 *   - agency: which Department / Bureau owns the records (DoW, FBI, NASA,
 *     CIA, DoE, ODNI, DoD, Other). Inferred from the slug prefix.
 *   - year / decade: the incident year mentioned in the slug, when there is
 *     one (most slugs include it).
 *
 * The home CaseLibrary and the /bureau browser both consume this. Keeping
 * it server-only means filesystem access stays out of the client bundle.
 */
import "server-only";
import { readdir, readFile, stat } from "node:fs/promises";
import path from "node:path";
import { pickLocaleBody } from "@/lib/bilingual";

export interface CaseFile {
  slug: string;
  topic: string;
  topic_pt_br: string | null;
  /** Opening paragraph of the body, trimmed and language-picked. */
  opening: string;
  /** Last modification timestamp, used as default sort. */
  mtimeMs: number;
  /** Incident year mentioned in the slug, or null. */
  year: number | null;
  /** Decade label like "1940s", "2020s", derived from `year`. */
  decade: string | null;
  /** Originating bureau / agency. */
  agency: Agency;
}

export type Agency =
  | "DoW"    // Department of War (recent UAP releases)
  | "FBI"
  | "NASA"
  | "CIA"
  | "DoE"
  | "ODNI"
  | "DoD"    // generic Department of Defense legacy
  | "Other"; // named historic cases (Kenneth Arnold, Mantell, green fireballs, …)

const CASE_ROOT = process.env.CASE_ROOT || "/data/ufo/case";

function parseFrontmatter(md: string): { fm: Record<string, string>; body: string } {
  const m = md.match(/^---\n([\s\S]+?)\n---\n([\s\S]*)$/);
  if (!m) return { fm: {}, body: md };
  const fm: Record<string, string> = {};
  for (const line of m[1].split("\n")) {
    const kv = line.match(/^([a-z_]+):\s*(.+)$/);
    if (!kv) continue;
    let v = kv[2].trim();
    if (v.startsWith('"') && v.endsWith('"')) v = v.slice(1, -1);
    fm[kv[1]] = v;
  }
  return { fm, body: m[2] };
}

function pickOpening(body: string, locale: "pt-br" | "en"): string {
  const lines = body.split("\n");
  const blocks: string[] = [];
  let cur: string[] = [];
  for (const raw of lines) {
    const line = raw.trim();
    if (line.length === 0) {
      if (cur.length > 0) { blocks.push(cur.join(" ")); cur = []; }
      continue;
    }
    if (line.startsWith("#") || line.startsWith(">") || line.startsWith("|")
        || line.startsWith("---") || line.startsWith("```")) {
      if (cur.length > 0) { blocks.push(cur.join(" ")); cur = []; }
      continue;
    }
    cur.push(line);
  }
  if (cur.length > 0) blocks.push(cur.join(" "));

  if (locale === "pt-br") {
    const ptIdx = body.indexOf("(PT-BR)");
    if (ptIdx >= 0) {
      const after = body.slice(ptIdx);
      const m = after.match(/\n\n([^\n#>|`-][^\n]+(?:\n[^\n#>|`-][^\n]+)*)/);
      if (m) return m[1].replace(/\s+/g, " ").trim().slice(0, 400);
    }
  }
  return (blocks[0] ?? "").replace(/\s+/g, " ").trim().slice(0, 400);
}

/** Map a slug to its originating agency by prefix / known fragments. */
function inferAgency(slug: string): Agency {
  if (slug.startsWith("dow-uap-"))   return "DoW";
  if (slug.startsWith("fbi-"))       return "FBI";
  if (slug.startsWith("nasa-"))      return "NASA";
  if (slug.startsWith("cia-"))       return "CIA";
  if (slug.startsWith("doe-"))       return "DoE";
  if (slug.startsWith("odni-"))      return "ODNI";
  // doc-65-hs1-… is the Hoover/FBI bulk dossier; ship it under FBI for the
  // reader even though the slug doesn't say so.
  if (slug.startsWith("doc-65-"))    return "FBI";
  if (slug.startsWith("doc-"))       return "DoD";
  return "Other";
}

/** Pull a 19xx / 20xx that looks like a standalone year token from the slug.
 *  Requires a non-digit boundary on each side, capping the search at the
 *  current year + 1 so embedded ID numbers like "doc-2070-…" don't get
 *  mistaken for a 2070 incident. */
function inferYear(slug: string): number | null {
  const max = new Date().getFullYear() + 1;
  const m = slug.match(/(?:^|[^0-9])(19\d{2}|20\d{2})(?:[^0-9]|$)/);
  if (!m) return null;
  const y = parseInt(m[1], 10);
  return y <= max ? y : null;
}

function decadeOf(year: number | null): string | null {
  if (year === null) return null;
  return `${Math.floor(year / 10) * 10}s`;
}

export async function loadCases(locale: "pt-br" | "en"): Promise<CaseFile[]> {
  const dir = path.join(CASE_ROOT, "reports");
  let files: string[];
  try {
    files = await readdir(dir);
  } catch {
    return [];
  }
  const items: CaseFile[] = [];
  for (const f of files) {
    if (!f.endsWith(".md")) continue;
    try {
      const full = path.join(dir, f);
      const md = await readFile(full, "utf-8");
      const st = await stat(full);
      const { fm, body } = parseFrontmatter(md);
      const slug = f.replace(/\.md$/, "");
      const h1s = [...body.matchAll(/^#\s+(.+)$/gm)].map((m) => m[1].trim());
      const enTitle = h1s[0] ?? fm.topic ?? f;
      const ptTitle = h1s[1] ?? h1s[0] ?? fm.topic_pt_br ?? fm.topic ?? f;
      const year = inferYear(slug);
      items.push({
        slug,
        topic: enTitle,
        topic_pt_br: ptTitle,
        opening: pickOpening(pickLocaleBody(body, locale), locale),
        mtimeMs: st.mtimeMs,
        year,
        decade: decadeOf(year),
        agency: inferAgency(slug),
      });
    } catch { /* skip broken file */ }
  }
  return items.sort((a, b) => b.mtimeMs - a.mtimeMs);
}