A WebSocket server that wraps the Stanford Part-of-Speech tagger for easier operation with javascript-based web apps
Java Shell CSS
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
js
.gitignore
README.md
StanfordTaggerServer.java
build.xml
start_server.bat
start_server.sh
tag.css
tag.html
websocket-screenshot.png

README.md

Bitdeli Badge

websocket-tagger

==========

Description

This project provides a WebSocket server that wraps the Stanford Part-of-Speech tagger. This makes it easier to get part-of-speech tags from JavaScript for arbitrary text. Please note that this project is just a prototype that illustrates the utility of WebSockets for NLP.

Requirements

Installation:

  1. Clone the Java WebSocket repository and build java_websocket.jar (you may also download the jar directly from the repository). Add this jar to your CLASSPATH.
  2. Download stanford-corenlp-1.3.5.jar and stanford-corenlp-1.3.5-models.jar and add them to your CLASSPATH (Note: make sure to download the latest version which at this point happens to be 1.3.5)
  3. Run ant to build the StanfordTaggerServer class.
  4. Run the included script start_server.sh to start the WebSocket server.
  5. Load tag.html into Firefox.
  6. Tag away.

Notes:

  1. Please note that currently the server needs to be killed manually since this is what I want. Obviously, this can be easily changed with the same thinking that goes into how to properly kill a regular socket server.
  2. It should be simple to extend the server to do anything else CoreNLP can do besides tagging.

Screenshot:

Here's a screenshot of the jQuery powered tag.html page in action: screenshot