When extracting body of a document in HTML source format (i.e. without pandoc conversion before display), use BeautifilSoup, which is more tolerant, instead of xml.dom.minidom, which crashes too often.