mwetoolkit Code

We had to rewrite some old patternlib code that was just not doing
the right thing. The semantics of id-order now works like this:
* For each id-order, print all words matching the pattern `p` with this ID.
* Exception: if a child of `p` has ignore="true", all grandchildren are * skipped.
An ignore="true" in a parent of `p` (or `p` itself) causes no skipping.

Basic testing example:
./bin/grep.py -e 'N{id=a} N{id=b}' --id-order="b:a" --to=PlainCorpus test/inputs/ted500.xml | head