de.unidu.is.retrieval.lemur
Class Lemur

java.lang.Object
  de.unidu.is.retrieval.lemur.Lemur

public class Lemur

extends java.lang.Object

A wrapper class for the Lemur (distributed) retrieval system.

Since:

2004-02-18

Version:

$Revision: 1.10 $, $Date: 2005/03/14 17:33:14 $

Author:

Henrik Nottelma

Field Summary

static int	inquery The ID for the INQUERY ranking method.
static int	kl The ID for the KL-Divergance ranking method.
static int	okapi The ID for the OKAPI ranking method.
static int	tfidf The ID for the tf.idf ranking method.

Constructor Summary

Lemur(java.io.File dir, boolean useSample)
Creates a new instance and initis the directories.

Method Summary

java.io.File	createTextFile(java.lang.String coll, java.util.Map termMap) Creates a text file (with the terms in it) in TREC format for the specified collection from the database.
void	createTopicFile(java.io.File topicFile, Query query, boolean append) Creates a topic file (with the terms in it) in TREC format for the specified node.
void	createTopicFile(java.lang.String topicsName, Query query, boolean append) Creates a topic file (with the terms in it) in TREC format for the specified node.
void	evaluate(java.lang.String coll, java.io.File topicsFile, java.io.File resultFile, int mode, int numDocs, boolean doLog) Evaluates the given topics on the specified collection.
void	evaluate(java.lang.String coll, java.lang.String topicsName, int mode, int numDocs, boolean doLog) Evaluates the given topics on the specified collection.
void	evaluateCORI(java.lang.String[] colls, java.io.File topicsFile, java.io.File ranksFile, java.io.File resultFile, int numDocs, int numDLs, boolean doLog) Evaluates the given topics on the specified collections, employing resource selection with CORI.
void	evaluateCORI(java.lang.String[] colls, java.lang.String topicsName, int numDocs, int numDLs, boolean doLog) Evaluates the given topics on the specified collections, employing resource selection with CORI.
void	generateSmooth(java.lang.String coll, boolean doLog) Generates a smoothing file for the collection.
java.util.List[]	getCORIResult(java.lang.String[] colls, Query query, int numDocs, int numDLs, boolean doLog) Returns the result for the specified query after resource selection with CORI.
java.io.File	getDir() Returns the base directory.
java.io.File	getIndexCORIDir() Returns the CORI index directory.
java.io.File	getIndexDir() Returns the index directory.
java.io.File	getRanksCORIFile(java.lang.String topicsName) Returns the file name for the ranks of the CORI evaluation.
java.util.List	getResult(java.lang.String coll, Query query, int mode, int numDocs, boolean doLog) Returns the result for the specified query.
java.io.File	getResultCORIFile(java.lang.String topicsName) Returns the file name for the results of the CORI evaluation.
java.io.File	getResultDir() Returns the result directory.
java.io.File	getResultFile(java.lang.String topicsName, java.lang.String coll) Returns the file name for the evaluation results.
java.io.File	getSampleDir() Returns the sample directory.
java.io.File	getTextDir() Returns the text directory.
java.io.File	getTopicDir() Returns the topics directory.
void	index(java.lang.String coll, boolean doLog) Indexes the collection.
void	indexCORI(java.lang.String[] colls, boolean doLog) Computes a CORI index and creates for each collection a parameter file which will be used for retrieval if the collection will be selected.
void	setDir(java.io.File file) Sets the base directory.
void	setDir(java.io.File file, boolean useSample) Sets the base directory.
void	setIndexCORIDir(java.io.File file) Sets the CORI index directory
void	setIndexDir(java.io.File file) Sets the index directory
void	setResultDir(java.io.File file) Sets the result directory
void	setSampleDir(java.io.File file) Sets the sample directory
void	setTextDir(java.io.File file) Sets the text directory
void	setTopicDir(java.io.File file) Sets the topic directory

Methods inherited from class java.lang.Object

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

Field Detail

tfidf

public static final int tfidf

The ID for the tf.idf ranking method.

See Also:

Constant Field Values

okapi

public static final int okapi

The ID for the OKAPI ranking method.

See Also:

Constant Field Values

kl

public static final int kl

The ID for the KL-Divergance ranking method.

See Also:

Constant Field Values

inquery

public static final int inquery

The ID for the INQUERY ranking method.

See Also:

Constant Field Values

Constructor Detail

Lemur

public Lemur(java.io.File dir,
             boolean useSample)

Creates a new instance and initis the directories.

Parameters:

dir - base directory

useSample - if true, the sample will be used for the CORI index

Method Detail

createTextFile

public java.io.File createTextFile(java.lang.String coll,
                                   java.util.Map termMap)

Creates a text file (with the terms in it) in TREC format for the specified collection from the database.

Parameters:

coll - collection name

termMap - document-termlist map

index

public void index(java.lang.String coll,
                  boolean doLog)

Indexes the collection.

Parameters:

coll - collection name

doLog - if true, the output of the Lemur program is logged

generateSmooth

public void generateSmooth(java.lang.String coll,
                           boolean doLog)

Generates a smoothing file for the collection.

Parameters:

coll - collection name

doLog - if true, the output of the Lemur program is logged

createTopicFile

public void createTopicFile(java.lang.String topicsName,
                            Query query,
                            boolean append)

Creates a topic file (with the terms in it) in TREC format for the specified node.

Parameters:

topicsName - name for the topic

query - query

append - if true, appends to that file

createTopicFile

public void createTopicFile(java.io.File topicFile,
                            Query query,
                            boolean append)

Creates a topic file (with the terms in it) in TREC format for the specified node.

Parameters:

topicFile - file

query - query

append - if true, appends to that file

getResult

public java.util.List getResult(java.lang.String coll,
                                Query query,
                                int mode,
                                int numDocs,
                                boolean doLog)

Returns the result for the specified query.

Parameters:

coll - collection name

query - query

mode - Lemur mode

numDocs - number of documents

doLog - if true, the output of the Lemur program is logged

Returns:

list of ProbDoc instances

evaluate

public void evaluate(java.lang.String coll,
                     java.lang.String topicsName,
                     int mode,
                     int numDocs,
                     boolean doLog)

Evaluates the given topics on the specified collection.

Parameters:

coll - collection name

topicsName - name of the topics (for the topic and the result file)

mode - Lemur mode

numDocs - number of documents

doLog - if true, the output of the Lemur program is logged

evaluate

public void evaluate(java.lang.String coll,
                     java.io.File topicsFile,
                     java.io.File resultFile,
                     int mode,
                     int numDocs,
                     boolean doLog)

Evaluates the given topics on the specified collection.

Parameters:

coll - collection name

topicsFile - topics file

resultFile - file for the results

mode - Lemur mode

numDocs - number of documents

doLog - if true, the output of the Lemur program is logged

indexCORI

public void indexCORI(java.lang.String[] colls,
                      boolean doLog)

Computes a CORI index and creates for each collection a parameter file which will be used for retrieval if the collection will be selected.

Parameters:

colls - collections to be used for the index

doLog - if true, the output of the Lemur program is logged

getCORIResult

public java.util.List[] getCORIResult(java.lang.String[] colls,
                                      Query query,
                                      int numDocs,
                                      int numDLs,
                                      boolean doLog)

Returns the result for the specified query after resource selection with CORI.

Parameters:

colls - collection names

query - query

numDocs - number of documents

numDLs - number of DLs to be selected

doLog - if true, the output of the Lemur program is logged

Returns:

array of two lists of ProbDoc instances (0: result, 1: ranks)

evaluateCORI

public void evaluateCORI(java.lang.String[] colls,
                         java.lang.String topicsName,
                         int numDocs,
                         int numDLs,
                         boolean doLog)

Evaluates the given topics on the specified collections, employing resource selection with CORI.

Parameters:

colls - collection names

topicsName - name of the topics (for the topic and the result file)

numDocs - number of documents

numDLs - number of DLs to be selected

doLog - if true, the output of the Lemur program is logged

evaluateCORI

public void evaluateCORI(java.lang.String[] colls,
                         java.io.File topicsFile,
                         java.io.File ranksFile,
                         java.io.File resultFile,
                         int numDocs,
                         int numDLs,
                         boolean doLog)

Evaluates the given topics on the specified collections, employing resource selection with CORI.

Parameters:

colls - collection names

topicsFile - topics file

ranksFile - file for the ranks

resultFile - file for the results

numDocs - number of documents

numDLs - number of DLs to be selected

doLog - if true, the output of the Lemur program is logged

getResultFile

public java.io.File getResultFile(java.lang.String topicsName,
                                  java.lang.String coll)

Returns the file name for the evaluation results.

Parameters:

topicsName - name of the topics (for the topic and the result file)

coll - colection name

Returns:

file name for the evaluation results

getRanksCORIFile

public java.io.File getRanksCORIFile(java.lang.String topicsName)

Returns the file name for the ranks of the CORI evaluation.

Parameters:

topicsName - name of the topics (for the topic and the result file)

Returns:

file name for the evaluation ranks

getResultCORIFile

public java.io.File getResultCORIFile(java.lang.String topicsName)

Returns the file name for the results of the CORI evaluation.

Parameters:

topicsName - name of the topics (for the topic and the result file)

Returns:

file name for the evaluation results

getDir

public java.io.File getDir()

Returns the base directory.

Returns:

base directory

getIndexCORIDir

public java.io.File getIndexCORIDir()

Returns the CORI index directory.

Returns:

directory

getIndexDir

public java.io.File getIndexDir()

Returns the index directory.

Returns:

directory

getResultDir

public java.io.File getResultDir()

Returns the result directory.

Returns:

directory

getSampleDir

public java.io.File getSampleDir()

Returns the sample directory.

Returns:

directory

getTextDir

public java.io.File getTextDir()

Returns the text directory.

Returns:

directory

getTopicDir

public java.io.File getTopicDir()

Returns the topics directory.

Returns:

directory

setDir

public void setDir(java.io.File file)

Sets the base directory.

Parameters:

file - base directory

setDir

public void setDir(java.io.File file,
                   boolean useSample)

Sets the base directory.

Parameters:

file - base directory

useSample - if true, uses samples instead of full text

setIndexCORIDir

public void setIndexCORIDir(java.io.File file)

Sets the CORI index directory

Parameters:

file - directory

setIndexDir

public void setIndexDir(java.io.File file)

Sets the index directory

Parameters:

file - directory

setResultDir

public void setResultDir(java.io.File file)

Sets the result directory

Parameters:

file - directory

setSampleDir

public void setSampleDir(java.io.File file)

Sets the sample directory

Parameters:

file - directory

setTextDir

public void setTextDir(java.io.File file)

Sets the text directory

Parameters:

file - directory

setTopicDir

public void setTopicDir(java.io.File file)

Sets the topic directory

Parameters:

file - directory