memind/mindspace-geo-meta.mjs

import { extractCoverSignals } from './mindspace-thumbnails.mjs';

function stripHtml(value) {
  return String(value ?? '')
    .replace(/<[^>]+>/g, ' ')
    .replace(/\s+/g, ' ')
    .trim();
}

function htmlWithoutEmbeddedCode(html) {
  return String(html ?? '')
    .replace(/<script\b[\s\S]*?<\/script>/gi, ' ')
    .replace(/<style\b[\s\S]*?<\/style>/gi, ' ');
}

function sanitizeFaqItem(item) {
  const q = stripHtml(item?.q ?? item?.question).slice(0, 120);
  const a = stripHtml(item?.a ?? item?.answer).slice(0, 500);
  if (!q || !a || q.length < 2 || a.length < 4) return null;
  if (/document\.|function\s*\(|createElement|innerHTML|querySelector/i.test(`${q} ${a}`)) return null;
  return { q, a };
}

function decodeMetaAttribute(value) {
  return String(value ?? '').replaceAll('&quot;', '"').replaceAll('&amp;', '&').replaceAll('&lt;', '<');
}

function extractGeoMetaJsonRaw(html) {
  const source = String(html ?? '');
  const marker = source.match(/<meta[^>]*name=["']mindspace-geo["'][^>]*content\s*=\s*/i);
  if (!marker) return null;
  let rest = source.slice(marker.index + marker[0].length);
  if (rest[0] === '"') {
    let raw = '';
    for (let i = 1; i < rest.length; i += 1) {
      if (rest[i] === '&' && rest.slice(i, i + 6) === '&quot;') {
        raw += '"';
        i += 5;
        continue;
      }
      if (rest[i] === '"') return raw;
      raw += rest[i];
    }
    return null;
  }
  const jsonStart = rest.indexOf('{');
  if (jsonStart < 0) return null;
  let depth = 0;
  let inString = false;
  let escape = false;
  for (let i = jsonStart; i < rest.length; i += 1) {
    const ch = rest[i];
    if (inString) {
      if (escape) {
        escape = false;
        continue;
      }
      if (ch === '\\') {
        escape = true;
        continue;
      }
      if (ch === '"') inString = false;
      continue;
    }
    if (ch === '"') {
      inString = true;
      continue;
    }
    if (ch === '{') depth += 1;
    else if (ch === '}') {
      depth -= 1;
      if (depth === 0) return rest.slice(jsonStart, i + 1);
    }
  }
  return null;
}

export function parseMindspaceGeoMeta(html) {
  const raw = extractGeoMetaJsonRaw(html);
  if (!raw) return {};
  try {
    return JSON.parse(decodeMetaAttribute(raw));
  } catch {
    return {};
  }
}

export function extractFaqFromHeadings(html, { maxItems = 6 } = {}) {
  const source = htmlWithoutEmbeddedCode(html);
  const faq = [];
  const seen = new Set();
  const pattern =
    /<(h[23])[^>]*>([\s\S]*?)<\/\1>\s*(?:<p[^>]*>([\s\S]*?)<\/p>|<div[^>]*>([\s\S]*?)<\/div>)/gi;
  for (const match of source.matchAll(pattern)) {
    const item = sanitizeFaqItem({
      q: match[2],
      a: match[3] ?? match[4],
    });
    if (!item) continue;
    const key = item.q.toLowerCase();
    if (seen.has(key)) continue;
    seen.add(key);
    faq.push(item);
    if (faq.length >= maxItems) break;
  }
  return faq;
}

export function resolveGeoKeywords(html, authorMeta = null) {
  const meta = authorMeta ?? parseMindspaceGeoMeta(html);
  const fromAuthor = Array.isArray(meta.keywords)
    ? meta.keywords.map((item) => String(item ?? '').trim()).filter(Boolean)
    : [];
  if (fromAuthor.length > 0) return fromAuthor.slice(0, 12);

  const keywords = [];
  const seen = new Set();
  const push = (value) => {
    const token = String(value ?? '').trim();
    if (!token || token.length < 2 || seen.has(token)) return;
    seen.add(token);
    keywords.push(token);
  };

  const cover = extractCoverSignals(String(html ?? ''));
  push(cover.tag);
  push(cover.subtitle);

  const title = String(html ?? '').match(/<title[^>]*>([^<]+)<\/title>/i)?.[1]?.trim() ?? '';
  if (title) push(title);

  for (const match of String(html ?? '').matchAll(/<h2[^>]*>([\s\S]*?)<\/h2>/gi)) {
    push(stripHtml(match[1]));
    if (keywords.length >= 12) break;
  }
  return keywords.slice(0, 12);
}

export function resolveGeoFaq(html, authorMeta = null) {
  const meta = authorMeta ?? parseMindspaceGeoMeta(html);
  const fromAuthor = Array.isArray(meta.faq)
    ? meta.faq.map((item) => sanitizeFaqItem(item)).filter(Boolean)
    : [];
  if (fromAuthor.length > 0) return fromAuthor.slice(0, 8);
  return extractFaqFromHeadings(html);
}