product-sourcing-capture/extension/content.js

chrome.runtime.onMessage.addListener((message, sender, sendResponse) => {
  if (message?.type !== "EXTRACT_PRODUCTS") return;
  try {
    sendResponse({ ok: true, items: extractVisibleProducts(message.mode || "page") });
  } catch (error) {
    sendResponse({ ok: false, error: error.message || String(error), items: [] });
  }
});

function extractVisibleProducts(mode) {
  const platform = detectPlatform(location.href);
  const sourcePage = location.href;
  const title = cleanText(document.title);
  const candidates = mode === "main" ? [document.body] : findCandidateNodes();
  const items = [];
  const seen = new Set();

  for (const node of candidates) {
    const item = extractFromNode(node, platform, sourcePage, title);
    if (!item.title && !item.image) continue;
    const key = item.url || `${item.title}|${item.image}`;
    if (seen.has(key)) continue;
    seen.add(key);
    items.push(item);
    if (items.length >= 30) break;
  }

  if (!items.length) {
    const fallback = extractFromNode(document.body, platform, sourcePage, title);
    if (fallback.title || fallback.image) items.push(fallback);
  }

  return items;
}

function findCandidateNodes() {
  const nodes = new Set();
  const imageNodes = Array.from(document.images)
    .filter((img) => {
      const rect = img.getBoundingClientRect();
      const src = img.currentSrc || img.src || img.getAttribute("data-src") || "";
      return src && !src.startsWith("data:") && rect.width >= 80 && rect.height >= 80;
    })
    .slice(0, 80);

  for (const img of imageNodes) {
    nodes.add(findCardRoot(img));
  }

  const linkNodes = Array.from(document.querySelectorAll("a[href]"))
    .filter((link) => {
      const rect = link.getBoundingClientRect();
      const text = cleanText(link.innerText || link.textContent || "");
      return rect.width >= 80 && rect.height >= 40 && (link.querySelector("img") || /[¥￥]\s*\d|\d+(?:\.\d+)?\s*元/.test(text));
    })
    .slice(0, 60);

  for (const link of linkNodes) nodes.add(findCardRoot(link));

  return Array.from(nodes).filter(Boolean).slice(0, 90);
}

function findCardRoot(node) {
  let current = node;
  let best = node;
  for (let i = 0; i < 5 && current?.parentElement; i += 1) {
    current = current.parentElement;
    const rect = current.getBoundingClientRect();
    const text = cleanText(current.innerText || current.textContent || "");
    if (rect.width >= 100 && rect.height >= 80 && text.length <= 900) best = current;
    if (/[¥￥]\s*\d|\d+(?:\.\d+)?\s*元/.test(text) && text.length >= 8) {
      best = current;
      break;
    }
  }
  return best;
}

function extractFromNode(node, site, pageUrl, pageTitle) {
  const text = cleanText(node.innerText || node.textContent || "");
  const linkEl = node.matches?.("a[href]") ? node : node.querySelector?.("a[href]");
  const url = absolutize(linkEl?.href || pageUrl);
  const image = bestImage(node);
  const priceText = findPriceText(text);
  const titleText = bestTitle(node, text, pageTitle);
  const supplier = findSupplier(node, text);

  return {
    platform: site,
    url,
    sourcePage: pageUrl,
    title: titleText,
    image,
    priceText,
    supplier,
    shop: supplier,
    metrics: findMetrics(text),
    notes: `插件采集；页面标题：${pageTitle}`,
  };
}

function bestTitle(node, text, pageTitle) {
  const attrs = [
    node.getAttribute?.("title"),
    node.getAttribute?.("aria-label"),
    node.querySelector?.("[title]")?.getAttribute("title"),
    node.querySelector?.("h1,h2,h3")?.innerText,
    node.querySelector?.("[class*='title'],[class*='name'],[class*='desc']")?.innerText,
  ];
  const fromAttrs = attrs.map(cleanText).find((value) => value && value.length >= 4);
  if (fromAttrs) return limit(fromAttrs, 120);

  const lines = text
    .split(/[\n\r。]/)
    .map(cleanText)
    .filter((line) => line.length >= 4 && !/^[¥￥]?\d/.test(line));
  const bestLine = lines.find((line) => /耳|链|饰|手|项|发|戒|藏|民族|绿松石|珍珠|银|夹|挂/.test(line)) || lines[0];
  return limit(bestLine || pageTitle, 120);
}

function bestImage(node) {
  const images = Array.from(node.querySelectorAll?.("img") || (node.tagName === "IMG" ? [node] : []))
    .map((img) => {
      const rect = img.getBoundingClientRect();
      return {
        src: img.currentSrc || img.src || img.getAttribute("data-src") || img.getAttribute("data-lazy") || "",
        score: rect.width * rect.height + (img.naturalWidth || 0) * (img.naturalHeight || 0),
      };
    })
    .filter((image) => image.src && !image.src.startsWith("data:"))
    .sort((a, b) => b.score - a.score);
  return absolutize(images[0]?.src || "");
}

function findPriceText(text) {
  const match = text.match(/(?:¥|￥)\s*\d+(?:\.\d+)?(?:\s*[-~至]\s*\d+(?:\.\d+)?)?|(?:批发价|价格|券后|到手价)?\s*\d+(?:\.\d+)?\s*元/);
  return cleanText(match?.[0] || "");
}

function findSupplier(node, text) {
  const el = node.querySelector?.("[class*='shop'],[class*='seller'],[class*='supplier'],[class*='company'],[class*='store']");
  const fromEl = cleanText(el?.innerText || el?.textContent || "");
  if (fromEl) return limit(fromEl, 80);
  const match = text.match(/[\u4e00-\u9fa5A-Za-z0-9（）()]{2,30}(?:工厂|厂家|旗舰店|饰品厂|商行|店|公司)/);
  return cleanText(match?.[0] || "");
}

function findMetrics(text) {
  const metrics = {};
  const patterns = {
    sales: /(?:已售|销量|人付款|成交|售出)\s*[:：]?\s*[\d.万wW+]+/,
    comments: /(?:评论|评价)\s*[:：]?\s*[\d.万wW+]+/,
    likes: /(?:点赞|赞|喜欢)\s*[:：]?\s*[\d.万wW+]+/,
    saves: /(?:收藏|加购)\s*[:：]?\s*[\d.万wW+]+/,
  };
  for (const [key, pattern] of Object.entries(patterns)) {
    const match = text.match(pattern);
    if (match) metrics[key] = match[0];
  }
  metrics.visibleText = limit(text, 500);
  return metrics;
}

function detectPlatform(url) {
  const host = new URL(url).hostname;
  if (host.includes("1688.com")) return "1688";
  if (host.includes("xiaohongshu.com") || host.includes("xhslink.com")) return "xiaohongshu";
  if (host.includes("tmall.com")) return "tmall";
  if (host.includes("taobao.com")) return "taobao";
  return "other";
}

function absolutize(url) {
  if (!url) return "";
  try {
    return new URL(url, location.href).href;
  } catch {
    return "";
  }
}

function cleanText(value) {
  return String(value || "").replace(/\s+/g, " ").trim();
}

function limit(value, length) {
  const text = cleanText(value);
  return text.length > length ? `${text.slice(0, length)}…` : text;
}