It is not important that how to create data but our focus should to how handle large amuont of data so that it can be retrieved in case of necessity.

Created in HTML Slidy: Slide Shows in XHTML

Harsh Vardhan
M Tech I CSE
IIT Bombay
Home Page

XHTML/CSS

Overview

Problem Statement

Solution Design

Implementation Plan

Testing Plan

Context

Problem Statement

This project is about developing a Desktop Search Engine.It takes query from the user parses it and displays the text based documents stored in the drive with some specified ranking criterian.

This Search Engine also initially builds the database of each and every document stored in the form of tokens and searching is performed on this database only.

Minimum Work : A basic search engine would be developed which takes input from user deletes the stop words ie redundant words search through the indexed databse which is prepared earlier and ranks them to single criteria and displays the output in the GUI .

Wish list : Their are two improvements possible
To add a scheduler on the crawler part so that user will not have to specify the path and it would construct a index for all the text files present in it.
To give user freedom to himself self the ranking criteria

Solution Design

In this project LUCENE API's are going to be used provided by Apache
LUCENE provides a kind of framework on which search routine can be made.

Solution Design contd...

Indexer :Indexer module is been used to create the indexed database which is been generated by storing the tokens of every text documents on which searech is to be performed.
Analyser : This module takes input the user query from GUI and analyses it . By Analysing it means that it deletes the unnecessary words called as stop words . And it generates the internal query which would be parsed by the searcher.
Searcher : This module takes input the from the analyser module which a internal query and interprets it. Searcher then executes the query on the indexed database and takes output ranks them in specified criteria and returns back the link to the user of document containing input words.
GUI : This module directly interacts with the user takes the input and displays the output.

Implementation

Language Used	JAVA
API Used for back end	LUCENE API
API Used for front end	JAVA Swing classes
DATA Used	Reasonable amount of text based documents
Amount of code	upto 500 LOC
Current Status	30% of back end coding done

Testing

To test this project we have to divide the test cases into 2 phases

To test modules involving creation of database : We will specify some path and then we will see whether all the text based documents are properly indexed and stored into the database or not .
To test modules regarding the searching : The user will provide some words which are used in many documents and we will see the result whether all those documents are displayed or not and whether they are ranked in proper order or not.

Context

Desktop Search Engine is now a days most required module to be stored in the users hard disk. Since the bulk of data is increasing tremendously as liks that of internet now the necessity has come to search the words in the document stored in our system only.

Since this search module is a general part it can also be incorporated on webpages also . For example some web sites are very bulky and contain thousands of pages so if developer wants to provide user a solution to effectively search to his required data he can prepare the index database for his web sit itself.

This desktop search Engine can be extended to a web based search engine . Instaead of taking documents stored on the web pages it should taks documents from the web itself and store them in form of tokens.

Desktop Search Engine