Add LDC datasets

[mjpost]

If $LDC were defined, LDC datasets installed locally could be extracted in similar fashion without violating any licenses. However this would require writing an XML parser to handle multiple references since Python 3 doesn't include one.

[ozancaglayan]

There's this standard library: https://docs.python.org/3/library/xml.etree.elementtree.html

[mjpost]

Yeah, I don't know why I thought we'd have to write one. For the Anthology we use lxml.

In my experience the problem with many "community-wrapped" XML files (e.g., WMT) is that they do not parse.