Get Started

Installation

The most straight forward way to install PredPatt is via `pip:

$ pip install git+https://github.com/hltcoe/PredPatt.git

Alternatively, one can install from source:

$ git clone https://github.com/hltcoe/PredPatt.git
$ cd PredPatt
$ python setup.py develop

PredPatt supports many input formats: raw text (English only; via Berkeley parser), CoNLLu data formats and Concretely annotated corpora.

Fire up the interactive web demo locally.

$ python bin/server.py

Command-line usage (see --help for other usage)

$ python -m predpatt test/en-ud-dev.conllu

>>> from predpatt import PredPatt
>>> pp = PredPatt.from_sentence('Chris loves silly dogs and clever cats .')
>>> print pp.pprint()

	?a loves ?b	[loves-root]
		?a: Chris	[Chris-nsubj]
		?b: silly dogs and clever cats	[dogs-dobj]

Notes:

The first time you call PredPatt.from_sentence, the Berkeley parser will be downloaded. So it might take a little while and you'll need an internet connection. It will store the parser and it's grammar in ~/.PredPatt.
The parser is generally slow to parse the first sentence and significantly faster on subsequent ones because it lazily loads the parsing model, etc.
There is a caching layer on top of the parser, which caches sentences under ~/.PredPatt. To bypass the caching layer, invoke from_sentence with the cacheable=False option.
There are a substantial number of options for a variety of syntactic phenomena. For more information see the tutorial.