/**
 * PDF text extraction (SPEC §1): auction houses that publish prices-realised lists as PDFs.
 * pdfjs-dist is loaded lazily so connectors that never touch PDFs pay nothing.
 */
export interface PdfPage {
  page: number;
  /** lines reconstructed from text items grouped by y coordinate */
  lines: string[];
}

export async function extractPdfPages(data: Uint8Array, opts: { maxPages?: number } = {}): Promise<PdfPage[]> {
  const pdfjs = (await import('pdfjs-dist/legacy/build/pdf.mjs')) as typeof import('pdfjs-dist/legacy/build/pdf.mjs');
  const doc = await pdfjs.getDocument({ data, useSystemFonts: true, disableFontFace: true, isEvalSupported: false }).promise;
  const pages: PdfPage[] = [];
  const n = Math.min(doc.numPages, opts.maxPages ?? doc.numPages);
  for (let i = 1; i <= n; i++) {
    const page = await doc.getPage(i);
    const content = await page.getTextContent();
    const rows = new Map<number, Array<{ x: number; s: string }>>();
    for (const it of content.items as Array<{ str?: string; transform?: number[] }>) {
      if (!it.str || !it.transform) continue;
      const y = Math.round(it.transform[5]! / 2) * 2; // 2pt buckets
      const x = it.transform[4]!;
      const row = rows.get(y) ?? [];
      row.push({ x, s: it.str });
      rows.set(y, row);
    }
    const lines = [...rows.entries()]
      .sort((a, b) => b[0] - a[0])
      .map(([, cells]) => cells.sort((a, b) => a.x - b.x).map((c) => c.s).join(' ').replace(/\s+/g, ' ').trim())
      .filter(Boolean);
    pages.push({ page: i, lines });
  }
  await doc.destroy();
  return pages;
}

export async function extractPdfText(data: Uint8Array, opts: { maxPages?: number } = {}): Promise<string> {
  const pages = await extractPdfPages(data, opts);
  return pages.map((p) => p.lines.join('\n')).join('\n\f\n');
}

/**
 * Parse "lot number … price" rows from prices-realised text. Returns null-safe rows; the connector
 * still decides currency and date. Handles "123  1,250", "Lot 123: $1,250", "123 ........ 1.250,00".
 */
export function parsePricesRealised(lines: string[]): Array<{ lot: string; priceText: string; line: string }> {
  const out: Array<{ lot: string; priceText: string; line: string }> = [];
  const re = /^(?:lot\s*)?(\d{1,6}[A-Za-z]?)\b[\s.:\-–]*.*?([$€£¥]?\s?\d{1,3}(?:[.,\s]\d{3})*(?:[.,]\d{2})?)\s*$/i;
  for (const line of lines) {
    const m = line.match(re);
    if (!m) continue;
    if (m[1] === m[2]) continue;
    out.push({ lot: m[1]!, priceText: m[2]!.trim(), line });
  }
  return out;
}