/**
 * Helpers shared by the North-American auction-house connectors of group g7 (Miller & Miller, Wright,
 * LAMA, Doyle, Fanatics Collect, Clean Sweep, HiBid). Source vocabulary and page-shape parsing live
 * here; canonical models stay clean. Everything is pure and unit-testable on saved HTML.
 */
import { parseGradeFromTitle } from '@rareindex/taxonomy';
import type { CurrencyCode } from '@rareindex/shared';
import { hintFromLabel, isBundleTitle, safeYear, slugFromTitle, type DeptHint } from '../_auction-lib/categories.js';
import { gradeOf, lotAttributes, makeLot, popCultureCategory, sportsCategory } from '../_memorabilia-lib/index.js';
import { makeSale } from '../../firecrawl/_carlib/index.js';

export { gradeOf, isBundleTitle, lotAttributes, makeLot, makeSale, safeYear, sportsCategory };

// ---------------------------------------------------------------------------------------------
// JSON scanning inside HTML / JS text
// ---------------------------------------------------------------------------------------------

/** Index just past the balanced JSON value (object/array) that starts at `start`; -1 when incomplete. */
export function scanJsonEnd(text: string, start: number): number {
  const open = text[start];
  if (open !== '{' && open !== '[') return -1;
  let depth = 0;
  let inStr = false;
  for (let i = start; i < text.length; i++) {
    const ch = text[i]!;
    if (inStr) {
      if (ch === '\\') i++;
      else if (ch === '"') inStr = false;
      continue;
    }
    if (ch === '"') inStr = true;
    else if (ch === '{' || ch === '[') depth++;
    else if (ch === '}' || ch === ']') {
      depth--;
      if (depth === 0) return i + 1;
    }
  }
  return -1;
}

/** Parse the balanced JSON value starting at `start` (null when malformed). */
export function parseJsonAt<T = unknown>(text: string, start: number): T | null {
  const end = scanJsonEnd(text, start);
  if (end < 0) return null;
  try {
    return JSON.parse(text.slice(start, end)) as T;
  } catch {
    return null;
  }
}

/**
 * Every JSON object that contains `marker` and starts with `objectStart` (e.g. Auction Mobility rows
 * start with `{"row_id":"` and are typed by `"type":"auction-lot-summary"`). Nested objects that also
 * match are skipped by jumping past each parsed object.
 */
export function jsonObjectsWithMarker<T = Record<string, unknown>>(text: string, marker: string, objectStart = '{"row_id":"'): T[] {
  const out: T[] = [];
  let from = 0;
  for (;;) {
    const i = text.indexOf(marker, from);
    if (i < 0) break;
    const start = text.lastIndexOf(objectStart, i);
    if (start < 0) {
      from = i + marker.length;
      continue;
    }
    const end = scanJsonEnd(text, start);
    if (end < 0 || end < i) {
      from = i + marker.length;
      continue;
    }
    try {
      out.push(JSON.parse(text.slice(start, end)) as T);
    } catch {
      /* malformed slice: skip */
    }
    from = Math.max(end, i + marker.length);
  }
  return out;
}

/** Minimal HTML entity decoding for attribute payloads (Inertia `data-page`). */
export function decodeEntities(s: string): string {
  return s
    .replace(/&quot;/g, '"')
    .replace(/&#0?39;|&apos;/g, "'")
    .replace(/&lt;/g, '<')
    .replace(/&gt;/g, '>')
    .replace(/&nbsp;/g, ' ')
    .replace(/&#(\d+);/g, (_, n: string) => String.fromCodePoint(Number(n)))
    .replace(/&#x([0-9a-f]+);/gi, (_, n: string) => String.fromCodePoint(Number.parseInt(n, 16)))
    .replace(/&amp;/g, '&');
}

/** Inertia.js page payload: `<div id="app" data-page="{…}">` (Wright / LAMA). */
export function inertiaPage<T = Record<string, unknown>>(html: string): T | null {
  const m = html.match(/<div[^>]+id="app"[^>]+data-page="([^"]+)"/);
  if (!m) return null;
  try {
    return JSON.parse(decodeEntities(m[1]!)) as T;
  } catch {
    return null;
  }
}

/** Next.js App Router flight payload: concatenated `self.__next_f.push([1,"…"])` chunks, JS-string decoded. */
export function nextFlightText(html: string): string {
  const re = /self\.__next_f\.push\(\[1,"((?:[^"\\]|\\.)*)"\]\)/g;
  const parts: string[] = [];
  let m: RegExpExecArray | null;
  while ((m = re.exec(html))) {
    try {
      parts.push(JSON.parse(`"${m[1]!}"`) as string);
    } catch {
      /* skip undecodable chunk */
    }
  }
  return parts.join('');
}

/** First JSON object following `key` (e.g. `"prefetchedItemData":`) in a flight/JS text. */
export function jsonAfterKey<T = Record<string, unknown>>(text: string, key: string, from = 0): T | null {
  const i = text.indexOf(key, from);
  if (i < 0) return null;
  const start = text.indexOf('{', i + key.length);
  if (start < 0 || start - (i + key.length) > 4) return null;
  return parseJsonAt<T>(text, start);
}

/** HiBid: `<script id="hibid-state" type="application/json">` → Apollo normalised cache. */
export function hibidApolloState(html: string): Record<string, Record<string, unknown>> | null {
  const m = html.match(/<script id="hibid-state" type="application\/json">([\s\S]*?)<\/script>/);
  if (!m) return null;
  try {
    const st = JSON.parse(decodeEntities(m[1]!)) as Record<string, unknown>;
    const cache = (st['apollo.state'] ?? st) as Record<string, Record<string, unknown>>;
    return cache && typeof cache === 'object' ? cache : null;
  } catch {
    return null;
  }
}

/** Resolve an Apollo `{ __ref }` (or inline object) against the cache. */
export function apolloRef<T = Record<string, unknown>>(cache: Record<string, Record<string, unknown>>, v: unknown): T | null {
  if (!v || typeof v !== 'object') return null;
  const ref = (v as { __ref?: string }).__ref;
  if (ref) return (cache[ref] as T | undefined) ?? null;
  return v as T;
}

// ---------------------------------------------------------------------------------------------
// Dates & money
// ---------------------------------------------------------------------------------------------

const MONTHS: Record<string, number> = { jan: 0, feb: 1, mar: 2, apr: 3, may: 4, jun: 5, jul: 6, aug: 7, sep: 8, sept: 8, oct: 9, nov: 10, dec: 11 };

/** "Sep 2, 2026 10:00 EST" | "Apr 16, 2026" → UTC date (US Eastern wall clock converted when a time zone is given). */
export function parseUsDate(s: string | null | undefined): Date | null {
  if (!s) return null;
  const m = s.match(/([A-Za-z]{3,9})\.?\s+(\d{1,2}),?\s+(\d{4})(?:\s+(\d{1,2}):(\d{2})\s*([AP]M)?\s*(E[SD]T|C[SD]T|M[SD]T|P[SD]T|UTC)?)?/);
  if (!m) return null;
  const mo = MONTHS[m[1]!.slice(0, 4).toLowerCase()] ?? MONTHS[m[1]!.slice(0, 3).toLowerCase()];
  if (mo === undefined) return null;
  const y = Number(m[3]);
  const d = Number(m[2]);
  if (!m[4]) return new Date(Date.UTC(y, mo, d));
  let hour = Number(m[4]);
  if (m[6] === 'PM' && hour < 12) hour += 12;
  if (m[6] === 'AM' && hour === 12) hour = 0;
  const offsets: Record<string, number> = { EST: 5, EDT: 4, CST: 6, CDT: 5, MST: 7, MDT: 6, PST: 8, PDT: 7, UTC: 0 };
  const off = offsets[m[7] ?? 'EST'] ?? 5;
  return new Date(Date.UTC(y, mo, d, hour + off, Number(m[5])));
}

/** "april-2025" | "April 2025" → first day of that month (UTC). Precision is the month — callers must say so. */
export function monthYearDate(s: string | null | undefined): Date | null {
  if (!s) return null;
  const m = s.match(/([A-Za-z]{3,9})[\s-]+(\d{4})/);
  if (!m) return null;
  const mo = MONTHS[m[1]!.slice(0, 3).toLowerCase()];
  if (mo === undefined) return null;
  return new Date(Date.UTC(Number(m[2]), mo, 1));
}

/** ISO 8601 (with or without zone; zone-less values are treated as UTC) → Date or null. */
export function isoDate(s: string | null | undefined): Date | null {
  if (!s) return null;
  const iso = /[zZ]$|[+-]\d{2}:?\d{2}$/.test(s) ? s : `${s}Z`;
  const d = new Date(iso);
  return Number.isNaN(d.getTime()) ? null : d;
}

/** "$2,048" | "2048.00" | 2048 → positive number or null. */
export function amount(v: unknown): number | null {
  if (v === null || v === undefined || v === '') return null;
  const n = typeof v === 'number' ? v : Number.parseFloat(String(v).replace(/[^0-9.-]/g, ''));
  return Number.isFinite(n) && n > 0 ? n : null;
}

export function isCurrency(s: string | null | undefined): s is CurrencyCode {
  return s === 'USD' || s === 'CAD' || s === 'EUR' || s === 'GBP' || s === 'CHF' || s === 'HKD' || s === 'AUD' || s === 'JPY';
}

// ---------------------------------------------------------------------------------------------
// Category mapping for general-purpose houses
// ---------------------------------------------------------------------------------------------

const SPORTS_LABEL = /sports?\s*(cards?|memorabilia)|hockey|baseball|basketball|football|trading cards?|game[- ]used/i;
const POP_LABEL = /pop culture|toys?|comics?|advertising|petroliana|coin[- ]op|movie|music|entertainment|disney|star wars/i;

/**
 * Department/sale label + lot title → taxonomy slug. Sports-card houses route through `sportsCategory`;
 * pop-culture sales through `popCultureCategory`; everything else through the shared auction mapper.
 * Returns null when nothing confident matched (the connector counts an anomaly and skips the lot).
 */
export function houseCategory(label: string | null | undefined, title: string, fallback: string | null = null): string | null {
  const l = label ?? '';
  if (SPORTS_LABEL.test(l)) return sportsCategory(title);
  if (/petroliana|advertising|soda|gas|oil|signs?/i.test(l)) return /\b(sign|clock|thermometer|display|calendar|poster|globe|tin|can|bottle|crate)\b/i.test(title) ? 'advertising' : slugFromTitle(title, 'unknown') ?? 'advertising';
  if (POP_LABEL.test(l)) return popCultureCategory(title);
  const hint: DeptHint = hintFromLabel(l);
  const slug = slugFromTitle(title, hint);
  if (slug) return slug;
  if (hint === 'design') return 'design_furniture';
  if (hint === 'art' || hint === 'contemporary' || hint === 'prints') return 'art';
  if (hint === 'furniture' || hint === 'decorative' || hint === 'asian' || hint === 'tribal' || hint === 'antiquities') return 'antiques';
  if (hint === 'books') return 'books';
  if (hint === 'jewelry') return 'jewelry';
  if (hint === 'watches') return 'other_watches';
  if (hint === 'coins') return 'coins';
  return fallback;
}

/** Card/TCG-aware mapper for sports-card marketplaces (Fanatics Collect). */
export function cardHouseCategory(title: string): string {
  if (/\bpok[eé]mon\b|\bcharizard\b|\bpikachu\b/i.test(title)) return 'pokemon';
  if (/magic:? the gathering|\bmtg\b|black lotus|\bmox\b/i.test(title)) return 'magic_the_gathering';
  if (/yu-?gi-?oh/i.test(title)) return 'yugioh';
  if (/\bone piece\b/i.test(title) && /\b(card|tcg|op0\d|leader|alt art|manga)\b/i.test(title)) return 'one_piece_card_game';
  if (/\blorcana\b/i.test(title)) return 'disney_lorcana';
  if (/\bdragon ?ball\b/i.test(title) && /\bcard|tcg|fusion world\b/i.test(title)) return 'dragon_ball_tcg';
  if (/\bdigimon\b/i.test(title)) return 'digimon_tcg';
  if (/\bflesh and blood\b/i.test(title)) return 'flesh_and_blood';
  if (/\bweiss schwarz\b/i.test(title)) return 'weiss_schwarz';
  if (/\b(marvel|dc|star wars|garbage pail|wacky pack|non-?sport)\b/i.test(title) && /\bcard|topps|panini|upper deck|fleer|skybox\b/i.test(title)) return 'non_sport_cards';
  if (/\b(sealed|booster box|hobby box|wax box|blaster|case)\b/i.test(title) && !/\bcard\b/i.test(title)) return sportsCategory(title);
  if (/\b(cgc|cbcs)\b.*\b#\s?\d+/i.test(title) && !/\bcard\b/i.test(title)) return popCultureCategory(title);
  return sportsCategory(title);
}

/** Grade for sale records: title grading via taxonomy, plus PSA/DNA-style authentication left as null grader. */
export function saleGrade(title: string): { grader: string | null; grade: string | null; qualifier: string | null } {
  const g = gradeOf(title);
  const q = parseGradeFromTitle(title).qualifier;
  return { grader: g.grader, grade: g.grade, qualifier: q };
}

/** Certification number written in a title, e.g. "PSA 10 (cert 12345678)" / "Cert #: 12345678". */
export function certFromTitle(title: string): string | null {
  const m = title.match(/\bcert(?:ification)?\.?\s*(?:no\.?|#|number)?[:\s#]*(\d{7,12})\b/i) ?? title.match(/\b(?:PSA|BGS|SGC|CGC|CBCS)\b[^()]{0,40}\((\d{7,12})\)/i);
  return m ? m[1]! : null;
}