Function main

scripts/docsearch-index.mjs:221–326

()

}

async function main() {
  console.log(
    sourceDir
      ? `Reading local sitemap: ${path.join(sourceDir, "sitemap.xml")}`
      : `Reading sitemap: ${sitemapUrl}`,
  );
  const sitemap = await readSitemap();
  const urls = extractUrlsFromSitemap(sitemap);

  console.log(`Indexing ${urls.length} URL(s) into ${indexName}`);

  const pageRecords = await mapConcurrent(
    urls,
    async (url, index) => {
      try {
        const html = await readPageHtml(url);
        const records = extractRecords(url, html);
        console.log(`${index + 1}/${urls.length} ${records.length} ${url}`);
        return records;
      } catch (error) {
        console.warn(
          `${index + 1}/${urls.length} skipped ${url}: ${error.message}`,
        );
        return [];
      }
    },
    concurrency,
  );

  const records = pageRecords.flat();
  console.log(`Extracted ${records.length} record(s)`);

  if (records.length === 0) {
    throw new Error("No records extracted; aborting Algolia update.");
  }

  await algoliaRequest(`/1/indexes/${encodeURIComponent(indexName)}/clear`, {});

  await algoliaRequest(
    `/1/indexes/${encodeURIComponent(indexName)}/settings`,
    {
      attributesForFaceting: ["type", "lang", "language", "version", "tags"],
      attributesToRetrieve: [
        "hierarchy",
        "content",
        "anchor",
        "url",
        "url_without_anchor",
        "type",
      ],
      attributesToHighlight: ["hierarchy", "content"],
      attributesToSnippet: ["content:10"],
      searchableAttributes: [
        "unordered(hierarchy.lvl0)",
        "unordered(hierarchy.lvl1)",
        "unordered(hierarchy.lvl2)",
        "unordered(hierarchy.lvl3)",
        "unordered(hierarchy.lvl4)",