Reimplement sanitization of published worksheets

[kcrisman]

See #100, which was later reverted here.

Nonetheless, this is useful, and probably crucial in some ways. Let's see if we can make this work out.

[kcrisman]

I suppose we could go as simple as this part of it:

+except ImportError:
+    def clean_html(text):
+        # looks ugly, but gets the job done
+        return text.replace('<', '&lt;')

But ugly, as he says.