Resources

Software

  • TuLiPA: A parsing environment for variants of tree-adjoining grammar.
  • rparse:  Discontinuous data-driven parsing with PLCFRS
  • uparse: Discontinuous shift-reduce parsing.
  • evalb-lcfrsAn extension of evalb bracket scoring for PLCFRS parses.
  • treetoolsA growing collection of algorithms for treebank tree processing.

More software on github.

Resources

  • DiscosuiteA test suite for German discontinuous constituency structures, based on the TIGER treebank.
  • Penn Treebank coordination annotation: An annotation layer for the Penn Treebank which marks coordinating punctuation. Available through the LDC.
  • Twotiger: A conversion script which reduces the block degree of the TIGER treebank to 2.