Function createHtmlLoader

packages/cli/src/cli/loaders/html.ts:12–251

()

} from "../utils/element-extraction";

export default function createHtmlLoader(): ILoader<
  string,
  Record<string, string>
> {
  return createLoader({
    async pull(locale, input) {
      const result: Record<string, string> = {};

      // Parse HTML with htmlparser2 (preserves structure, no foster parenting)
      const handler = new DomHandler();
      const parser = new htmlparser2.Parser(handler, {
        lowerCaseTags: false,
        lowerCaseAttributeNames: false,
      });
      parser.write(input);
      parser.end();

      const dom = handler.dom;

      // Get innerHTML equivalent (serialize children)
      function getInnerHTML(element: Element): string {
        return element.children
          .map((child) =>
            DomSerializer.default(child, { encodeEntities: false }),
          )
          .join("");
      }

      // Extract localizable attributes from element
      function extractAttributes(element: Element, path: string): void {
        const tagName = element.name.toLowerCase();
        const attrs = BASE_LOCALIZABLE_ATTRIBUTES[tagName];
        if (!attrs) return;

        for (const attr of attrs) {
          const value = element.attribs?.[attr];
          if (value && value.trim()) {
            result[`${path}#${attr}`] = value.trim();
          }
        }
      }

      // Recursively extract translation units from element tree
      const extractFromElement = createElementExtractor(
        { getInnerHTML, extractAttributes },
        result,
      );

      // Find head and body elements
      const html = domutils.findOne(
        (elem) => elem.type === "tag" && elem.name.toLowerCase() === "html",
        dom,
        true,
      ) as Element | null;

      if (html) {
        const head = domutils.findOne(
          (elem) => elem.type === "tag" && elem.name.toLowerCase() === "head",