#PullEnti SDK wrapper#

This project is just a handy (more or less) wrapper for Pullenti SDK.

PullEnti (Puller of Entities) is a fully-featured NER-library written in C#. It is free for non-commercial usage so just go and try it.

Code is dirty and, probably, buggy a bit but I don't really care. Don't get me wrong - I like to write good code. But the main reason of creation of this wrapper was the need to extract some info from a huge bunch of articles in hurry. Definitely clear, that I didn't want to spend much time polishing it and making it more generic.

##Usage##

###Input###

First you need a JSON-file. Not the whatever-you-found-json-file, but specially prepared one. The file should contain a set of text entities where every entity takes separate line.

{title: "....", content: "......"}
{title: "....", content: "......"}
{title: "....", content: "......"}
...

Remember that it's not an JSON-array: no brackets and no commas between objects.

###Output###

After you run

.\PullEntiCLI.exe path\to\your\file.json

it starts parsing your articles in parallel and writing results into files right into the same directory where your source filename.json is. All output should be splitted to as many files as many cores your CPU have. So if you have your source in D:\txt\ and 4 cores, you'll get:

D:\txt\0.json

D:\txt\1.json

D:\txt\2.json

D:\txt\3.json

The content of output files is slightly modified. Title-key lasts the same, content goes text and new key structure appears. It contains an array of parsed entities.

##Requirements##

This code was tested under C# 3.5+, Mono 3+.