Project for Chloe Calvarin, Judge Lee, Rowan Zellers

We are the banana team!

Here is a quick list of files and what they do

• chomsky_converter.py takes in an nltk.grammar CFG, and returns a chomsky normal form grammar equivalent.
  convert_grammar(your_cfg)

• chomsky_tree_converter.py takes in an nltk.tree Tree and returns a Tree that has been converted to chomsky normal form. The rules are systematic in such a way that we can convert a whole corpus without worrying about having new intermediate rules that get mixed up for each other.
  convert_tree(tree)

• cky_parser.py has both a deterministic and probabilistic CKY parser, with few frills added. It does, however, deal with unknown words (provided that there are probabilistic rules that handle unknowns)

  To use, we can run: myparser = cky_parser.ckyparser(grammar,start_symbol)
chart,mytrees = myparser.probabilistic_parse(sentence)
chart,mytrees = myparser.deterministic_parse(sentence)

• main.py was an initial playground for figuring out the corpus, it became where we run things from different files together.

• malt_parser.py
  TODO

• probabilities.py takes in trees and returns a grammar, with an added shortcut if you prefer using the fileids of the treebank. Since we convert our trees first in chomsky_tree_converter, you would feed the resulting list into here
  makeGrammar(tree_list)

• test_cky.py
  This file contains test code for each of the two CKY parsers (deterministic, and probabilistic).

• transition_parser.py
  TODO