Sherlok

Distributed restful text mining.

Sherlok is a flexible and powerful open source, distributed, real-time text-mining engine. Sherlok works as a RESTful annotation server based on Apache UIMA. For example, Sherlok can:

• highlight persons and locations in text (using DKPro OpenNLP),
• identify proteins and brain regions in biomedical texts (using Bluima),
• perform sentiment analysis using deep learning (using Stanford Sentiment),
• analyse the syntax of tweets (using TweetNLP),
• analyze clinical text and perform knowledge extraction (using Apache cTAKES)

Getting Started

• Download and unzip the latest Sherlok release
• Install a Java runtime
• Run bin/sherlok (Unix), or bin/sherlok.bat (Windows)

Annotate neuron mentions from Python:

pip install --upgrade sherlok

>>> from sherlok import Sherlok
>>> print list(Sherlok().annotate('neuroner', 'layer 4 neuron'))

[(0, 14, 'layer 4 neuron', u'Neuron', {}),
 (8, 14, 'neuron',  u'Neuron', {}),
 (8, 14, 'neuron',  u'NeuronTrigger', {}),
 (0, 7,  'layer 4', u'Layer', {u'ontologyId': u'HBP_LAYER:0000004'})]

Tag persons and locations with Javascript:

require('sherlok');
var text = 'Jack Burton (born April 29, 1954 in El Paso), also known as Jake Burton, is an American snowboarder and founder of Burton Snowboards.';
sherlok.annotate('opennlp.ners.en', text, function(annotation){
      console$(annotation);
});
{ begin=0, end=11,  value="person"}
{ begin=36, end=43, value="location"}
{ begin=60, end=71, value="person"}

More Built-in Text mining pipelines

• Berkeley parser ([demo](http://localhost:9600/annotate/berkeleyparser.en?text=The blue house of my childhood was bought: what a pity!))
• Malt parser ([demo](http://localhost:9600/annotate/maltparser.en?text=The blue house of my childhood was bought: what a pity!))

Further Documentation

• How to install and deploy Sherlok on your server.
• What are the major concept in Sherlok.
• What is a pipeline.
• What is a bundle.
• How to automatically download remote resources.
• How to communicate with Sherlok.
• What does REST mean.
• What actions can Sherlok perform.
• Using the built-in editor
• How to extend Sherlok capability.