Same as allow and deny args, it holds a string, a regex or an iterable of. Links whose text don't match one of the regex are filtered out.
DOC restrict_text in LxmlLinkExtractor
* tests for other link extractors are moved to test_linkextractors_deprecated.py
* in Python 3 Link is converted to use native strings for urls
* minor cleanups
* rename LinkExtractorTestCase to BaseSgmlLinkExtractorTestCase
* add BaseLinkExtractorTestCase link extractor tests can inherit from
and decouple it from SgmlLinkExtractor
* add an extra check for deny_extensions
* xfail test_restrict_xpaths_with_html_entities for LxmlLinkExtractor explicitly