iglu.examples
Class DirectorySearchEngine

java.lang.Object
  |
  +--iglu.examples.DirectorySearchEngine

public class DirectorySearchEngine

extends java.lang.Object

An example for using the iglu.ir package. This class creates a simple search engine from the files in a directory structure, assuming that the files are all text documents. It is rather inefficient, since it must rebuild the index every time it is used, but this simplicity makes it an easy example.

Author:

Ryan Scherle

Field Summary

private RAMSearchEngine	engine
private ValueSortedMap	searchResults
private StopList	stopList
private TFIDFVectorCreator	vecMaker

Constructor Summary

DirectorySearchEngine()
Initializes the the vector creator (which collects keywords from the files and assigns weights to them), and search engine (which allows the keywords to be stored and retrieved).

Method Summary

void	doSearch(java.io.File directory, TermVector query) Indexes all files in the directory, and sends the query to the engine.
private void	indexFile(java.io.File file, boolean indexing) Add a single file to the search engine.
static void	main(java.lang.String[] args) Runs the search engine.
private void	printResults(TermVector query, ValueSortedMap results) Prints out the list of files that match the query.
private void	processFiles(java.io.File file, boolean indexing) Goes through each file in the directory, passing the files to indexFile.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

engine

private RAMSearchEngine engine

vecMaker

private TFIDFVectorCreator vecMaker

searchResults

private ValueSortedMap searchResults

stopList

private StopList stopList

Constructor Detail

DirectorySearchEngine

public DirectorySearchEngine()

Initializes the the vector creator (which collects keywords from the files and assigns weights to them), and search engine (which allows the keywords to be stored and retrieved).

Method Detail

doSearch

public void doSearch(java.io.File directory,
                     TermVector query)

Indexes all files in the directory, and sends the query to the engine. Results are printed to stdout.

processFiles

private void processFiles(java.io.File file,
                          boolean indexing)

Goes through each file in the directory, passing the files to indexFile.

indexFile

private void indexFile(java.io.File file,
                       boolean indexing)

Add a single file to the search engine. This method must be called twice for every file. The first time, indexing is true, indicating that the file's information should be counted to provide the statistical information for the TFIDF algorithm. The second time, indexing is false, indicating that the term vector is actually added to the search engine.

printResults

private void printResults(TermVector query,
                          ValueSortedMap results)

Prints out the list of files that match the query.

main

public static void main(java.lang.String[] args)

Runs the search engine.