Skip to main content

The Python NLTK JSON-NLP package

Project description

NLTK-JSON-NLP

(C) 2019 by Damir Cavar, Oren Baldinger, Maanvitha Gongalla, Anurag Kumar, Murali Kammili, Boli Fang

Brought to you by the NLP-Lab.org!

Introduction

NLTK wrapper to JSON-NLP. NLTK has a wide variety of capabilities, but for our purposes we are limiting it to WordNet, VerbNet, and FrameNet. Other packages such as spaCy and Flair are more accurately able to annotate things like part of speech tags and dependency parses. See below for instruction on how to unify outputs from multiple packages.

Microservice

The JSON-NLP repository provides a Microservice class, with a pre-built implementation of [Flask]. To run it, execute:

python nltkjsonnlp/server.py

Since server.py extends the [Flask] app, a WSGI file would contain:

from nltkjsonnlp.server import app as application

Pipeline

JSON-NLP provides a simple Pipeline interface that we implement as NltkPipeline:

pipeline = nltkjsonnlp.NltkPipeline()
print(pipeline.process(text='I am a sentence.'))

Unification

To make the best use of this pipeline, it is best to unify it with a more accurate and complete pipeline such as spaCy-NLP-Json:

class UnifiedPipeline(pyjsonnlp.pipeline.Pipeline):
    def __init__(self):
        super(UnifiedPipeline, self).__init__()
        self.spacy = spacynlpjson.SpacyPipeline()
        self.nltk = nltkjsonnlp.NltkPipeline()

    def process(self, text='', coreferences=True, constituents=False, dependencies=True, expressions=True,
                **kwargs) -> OrderedDict:
        # start with a spacy parse
        spacy_json = self.spacy.process(text, spacy_model='en_core_web_md', constituents=False,
                                        coreferences=coreferences, dependencies=dependencies, expressions=False)
        # the get an nltk parse
        nltk_json = self.nltk.process(text)

        # unify the parses
        return pyjsonnlp.unification.unifier.add_annotation_to_a_from_b(a=spacy_json, 
                                                                        b=nltk_json, annotation='tokens')

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

nltkjsonnlp-0.0.5.tar.gz (9.4 kB view hashes)

Uploaded source

Built Distribution

nltkjsonnlp-0.0.5-py3-none-any.whl (14.7 kB view hashes)

Uploaded py3

Supported by

AWS AWS Cloud computing Datadog Datadog Monitoring Facebook / Instagram Facebook / Instagram PSF Sponsor Fastly Fastly CDN Google Google Object Storage and Download Analytics Huawei Huawei PSF Sponsor Microsoft Microsoft PSF Sponsor NVIDIA NVIDIA PSF Sponsor Pingdom Pingdom Monitoring Salesforce Salesforce PSF Sponsor Sentry Sentry Error logging StatusPage StatusPage Status page