refactor(transformer): add Document-accepting storeChildUrls overload

[marevol]

Summary

Refactor HtmlTransformer.storeChildUrls to expose a Document-accepting overload, so a caller that has already parsed the response body into a DOM (e.g. an XPath-based content extractor) can reuse that Document for child-URL extraction instead of parsing the same page a second time.

Changes

• Add protected void storeChildUrls(ResponseData, ResultData, Document) containing the existing child-URL extraction logic (base-href resolution, child-URL rules, getChildUrlSet() merge, self-URL removal).
• The existing storeChildUrls(ResponseData, ResultData) keeps identical behavior: it parses the body exactly as before and delegates to the new overload.

Behavior for all current callers (HtmlTransformer, FileTransformer, XpathTransformer) is unchanged:

• The 2-arg method still parses (no BOM handling / setEncoding change) and its try/catch is preserved.
• The new overload declares MalformedURLException, which the 2-arg method's existing catch (Exception) wraps into CrawlerSystemException("Could not store data.", e) exactly as before.
• The overridable helpers (getChildUrlRules, getUrlFromTagAttribute, getBaseHref, convertChildUrlList, addChildUrlFromTagAttribute) are still invoked via dynamic dispatch, so subclass customizations continue to apply.

Testing

• New test asserting the 3-arg overload produces the same child URLs as the 2-arg method for the same page, plus a sanity check that the 2-arg method still works end-to-end.
• Full fess-crawler module test suite passes (1980 tests, 0 failures); mvn formatter:format / license:format / javadoc:jar all clean.

Motivation

This enables a follow-up change in Fess (FessXpathTransformer) to parse each crawled page's DOM once instead of twice per page.