Ensure both open and close tags have been read.

Include tests for other sources.

Add most common sources for Japanese language datasets.