Logical Parsing with Python

LoPPy (LOgical Parsing with PYthon) is a Prolog-inspired tool for writing facts that can then be used to build rules.

The primary goal is to use this tool to build a syntactic parser for grammaticality judgments and part-of-speech tagging to be used in NLP tasks.

• loppy.py - The core classes for defining facts
• examples.py - Some examples of useful methods
• wh_grammar.py - The current syntactic parser with a focus on wh-questions (still a long way to go with this!)
• word_lists/ - Lists of words of various parts of speech for testing

• From the command line, run python wh_grammar.py.
• This should bring up the prompt Enter wh-question >>.
• Enter a wh-question comprising only words from the word_lists/ files, entirely in lower-case and without punctuation, e.g. who finds the cat that eats the bird quickly.
• This should return a tuple with a boolean indicating that this is a grammatical wh-question and a list of the parts-of-speech. Note that if the input is not grammatical, the parts-of-speech list will be empty.
• A prompt will ask if you would like to continue to test different input.

• When, Where, Why, How questions
• Plurals and verb conjugations
• Recursive sentence structures ("The cat the dog chases chases mice.")
• Improved tokenization
• Punctuation handling
• Vocabulary data gathering