mcSearch: Scientific Article Search

1
mcSearch Scientific Article Search

• Maxym Mykhalchuk
• Information Retrieval and Search Engines

2
Prequisites

• NetBeans IDE 4.1
• Has built-in Tomcat, and allows for an easy
  web-development
• Lucene
• Demo code runs (almost) perfectly
• Test data
• Text
• Lots of trash

3
What was done (1) Indexer

• Some modifications to extract a title and an
  abstract
• Title
• everything before Abstract
• Abstract
• Everything between Abstract and Introduction

4
What was done (2) Search

• Search design was ripped off
• Yandex search engine http//ya.ru
• Filter was created
• Allow/disallow trash files search
• Trash those that dont have title/abstract
  detected

5
Thats all

• Creating a search engine with Lucene is easy
• Updating the index is slow
• One has first to delete the document
• Then add it again
• Not nice for the fast-evolving Web