corenlp-server

A NLP web serverice supported by Stanford OpenNLP.

Build

Build the service using following command:

cd build
sh download_components.sh
sh build.sh

Run

You can run an example of this service via:

cd build
sh run.sh

The run.sh runs following command:

#!/bin/bash

java -cp jeromq-0.3.5.jar:corenlp-server.jar:stanford-corenlp-3.5.2.jar:stanford-chinese-corenlp-2015-04-20-models.jar so.knowledge.corenlp.server.Server -props StanfordCoreNLP-chinese.properties  -port 5900

Above command start a service for Chinese (according to model jar: stanford-chinese-corenlp-2015-04-20-models.jar and the property file StanfordCoreNLP-chinese.properties) at port 5900.

You can revise the command to load models for other language and listen at another port.

Service Interface

The server provide service using zeromq with REP pattern. It accept following commands:

stop

Client want to stop the server. The server reply 'stopping' and will stop service.

ping

Client ping the server to test if alive. The server gives a 'pong'.

process {text}

Process the text and return the result in json format.

Thanks

This code is based on https://github.com/kowey/corenlp-server.

rudaoshi/corenlp-server