Developers’ Guide

This is a rough table of contents for the upcoming Developers Guide. We’ll start filling these things in soon.

Using OpenPipeline

Introduction

Installation

A Quick Walkthrough
Define a connector
Stages, schedule, whatever
Configure a docfilter
Now what?

The Built-in Connectors
Discuss each

The Built-in DocFilters
The output of each, standard format
Reference the tokenizer
Any special considerations

The Built-in Stages
Discuss each, what it does
Tokenizer
UIMA wrapper

The Scheduler
The jobs config file
Different scheduling options

Logging
Where the logs go
Replacing a logger. Log4j.

The Server
Start, stop, admin page

How to run in Websphere, Tomcat, etc.
Configure openpipeline.home

Running OpenPipeline in batch mode
No need for an app server, just define your job and run this handy batch file
to execute the job on demand

Extending OpenPipeline

General Architecture
Show the picture, discuss

Plugins, generally
How to build a plugin, how to register it

The core classes: Items and Annotations
What they are
Logic behind the design
How to use them, sample code

How to write your own connector
What each method does
Creating a multi-threaded connector
Dealing with deletes, etc.
Creating a pipeline
See “why were stages designed this way” in Stages section

How to write your own docfilter
Discuss the sample, lexer, etc.

How to write your own stage
Extend, don’t implement
Calling processItem
Why were stages implemented this way
Stages must be thread-safe.
How to write a multi-threaded stage that has a shared resource
Diskfile, jdbc connection, socket, etc.
Traversing, dealing with annotations

Annotations
Extending the annotation class

This page is wiki editable login or register .