info.ephyra.search.searchers
Class IndriDocumentKM

java.lang.Object
  java.lang.Thread
      info.ephyra.search.searchers.Searcher
          info.ephyra.search.searchers.KnowledgeMiner
              info.ephyra.search.searchers.IndriDocumentKM

All Implemented Interfaces:

java.lang.Runnable

public class IndriDocumentKM
extends KnowledgeMiner

A KnowledgeMiner that deploys the Indri IR system to search a local text corpus. In contrast to IndriKM, whole documents are returned instead of paragraphs.

It runs as a separate thread, so several queries can be performed in parallel.

This class extends the class KnowledgeMiner.

Version:

2007-07-26

Author:

Nico Schlaefer

Nested Class Summary

Nested classes/interfaces inherited from class java.lang.Thread

java.lang.Thread.State, java.lang.Thread.UncaughtExceptionHandler

Field Summary

private static java.lang.String	FORBIDDEN_CHAR Regular expression that matches characters that cause problems in Indri queries and thus should be removed from query strings.
private java.lang.String[]	indriDirs Directories of Indri indices.
private java.lang.String[]	indriUrls URLs of Indri servers.
private static int	MAX_DOCS Maximum number of documents fetched at a time.
private static int	MAX_RESULTS_PERQUERY Maximum number of search results per query.
private static int	MAX_RESULTS_TOTAL Maximum total number of search results.

Fields inherited from class info.ephyra.search.searchers.KnowledgeMiner

firstResult, maxResults

Fields inherited from class info.ephyra.search.searchers.Searcher

query, results

Fields inherited from class java.lang.Thread

MAX_PRIORITY, MIN_PRIORITY, NORM_PRIORITY

Constructor Summary

IndriDocumentKM(java.lang.String[] locations, boolean isServers)
Creates a new Indri knowledge miner and sets the directories of indices or the URLs of servers.

Method Summary

protected Result[]	doSearch() Queries the Indri indices or servers and returns an array containing up to MAX_RESULTS_PERQUERY search results.
KnowledgeMiner	getCopy() Returns a new instance of IndriDocumentKM.
protected int	getMaxResultsPerQuery() Returns the maximum number of search results per query.
protected int	getMaxResultsTotal() Returns the maximum total number of search results.
static java.lang.String	transformQueryString(java.lang.String qs) Returns a representation of the query string that is suitable for Indri.

Methods inherited from class info.ephyra.search.searchers.KnowledgeMiner

getResults, getResults, start, start

Methods inherited from class info.ephyra.search.searchers.Searcher

run

Methods inherited from class java.lang.Thread

activeCount, checkAccess, countStackFrames, currentThread, destroy, dumpStack, enumerate, getAllStackTraces, getContextClassLoader, getDefaultUncaughtExceptionHandler, getId, getName, getPriority, getStackTrace, getState, getThreadGroup, getUncaughtExceptionHandler, holdsLock, interrupt, interrupted, isAlive, isDaemon, isInterrupted, join, join, join, resume, setContextClassLoader, setDaemon, setDefaultUncaughtExceptionHandler, setName, setPriority, setUncaughtExceptionHandler, sleep, sleep, start, stop, stop, suspend, toString, yield

Methods inherited from class java.lang.Object

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

Field Detail

MAX_RESULTS_TOTAL

private static final int MAX_RESULTS_TOTAL

Maximum total number of search results.

See Also:

Constant Field Values

MAX_RESULTS_PERQUERY

private static final int MAX_RESULTS_PERQUERY

Maximum number of search results per query.

See Also:

Constant Field Values

MAX_DOCS

private static final int MAX_DOCS

Maximum number of documents fetched at a time.

See Also:

Constant Field Values

FORBIDDEN_CHAR

private static final java.lang.String FORBIDDEN_CHAR

Regular expression that matches characters that cause problems in Indri queries and thus should be removed from query strings.

Indri allows the following characters:

• '?'..'Ø'
• 'a'..'z'
• 'A'..'Z'
• '0'..'9'
• '_'
• '-'
• '.' (only allowed if in between digits)
• whitespaces
• '"'

However, for some of the special characters Indri fails to retrieve results and therefore they are excluded.

See Also:

Constant Field Values

indriDirs

private java.lang.String[] indriDirs

Directories of Indri indices.

indriUrls

private java.lang.String[] indriUrls

URLs of Indri servers.

Constructor Detail

IndriDocumentKM

public IndriDocumentKM(java.lang.String[] locations,
                       boolean isServers)

Creates a new Indri knowledge miner and sets the directories of indices or the URLs of servers.

Parameters:

locations - directories of indices or URLs of servers

isServers - true iff the first parameter provides URLs of servers

Method Detail

transformQueryString

public static java.lang.String transformQueryString(java.lang.String qs)

Returns a representation of the query string that is suitable for Indri.

Parameters:

qs - query string

Returns:

query string for Indri

getMaxResultsTotal

protected int getMaxResultsTotal()

Returns the maximum total number of search results.

Specified by:

getMaxResultsTotal in class KnowledgeMiner

Returns:

maximum total number of search results

getMaxResultsPerQuery

protected int getMaxResultsPerQuery()

Returns the maximum number of search results per query.

Specified by:

getMaxResultsPerQuery in class KnowledgeMiner

Returns:

maximum total number of search results

doSearch

protected Result[] doSearch()

Queries the Indri indices or servers and returns an array containing up to MAX_RESULTS_PERQUERY search results.

Specified by:

doSearch in class Searcher

Returns:

Indri search results

getCopy

public KnowledgeMiner getCopy()

Returns a new instance of IndriDocumentKM. A new instance is created for each query.

Specified by:

getCopy in class KnowledgeMiner

Returns:

new instance of IndriDocumentKM