Extract readable plain text from an HTML fragment or document. Strips all tags, collapses whitespace, and drops blank lines — handy for feeding scraped HTML to an LLM or computing a text length.