csli.dialog.app.calo.topic.classification.topicextraction
Class CsliTopicSegmenterService

java.lang.Object
  csli.dialog.app.calo.topic.classification.topicextraction.CsliOnlineTopicSegmenterService
      csli.dialog.app.calo.topic.classification.topicextraction.CsliTopicSegmenterService

All Implemented Interfaces:

Service, OnlineTopicSegmenterService, TopicSegmenterService

Direct Known Subclasses:

CsliTopicSegmenterServiceImpl

public abstract class CsliTopicSegmenterService
extends CsliOnlineTopicSegmenterService
implements TopicSegmenterService

Nested Class Summary

Nested classes/interfaces inherited from class csli.dialog.app.calo.topic.classification.topicextraction.CsliOnlineTopicSegmenterService

CsliOnlineTopicSegmenterService.TopicSegmenterParameters

Constructor Summary

CsliTopicSegmenterService()

Method Summary

abstract void	clear() Clears all the cache files that have been generated by the TopicSegmenterService.
abstract void	createXML(String meetingName)
abstract Pair<Topic,TopicArea>	csliExtractTopic(String meeting, Pair<XSDDateTime,XSDDateTime> boundaries)
abstract Pair<Topic,TopicArea>	csliExtractTopic(String meeting, XSDDateTime time)
abstract ArrayList<Pair<Topic,TopicArea>>	csliGetTopics(String meeting) Get the Topics discussed in a meeting (using the default method), before adding to/merging with the pool
abstract ArrayList<Pair<Topic,TopicArea>>	csliGetTopicsByMITSegmentation(String meeting, Integer numShifts) Get the Topics discussed in a meeting (forcing segmentation via the MIT generative model), before adding to/merging with the pool
abstract ArrayList<Pair<Topic,TopicArea>>	csliGetTopicsByWordDistribution(String meeting) Get the Topics discussed in a meeting (forcing Stephane's lexical similarity method), before adding to/merging with the pool
Double	csliGetTopicSimilarity(Topic a, Topic b)
abstract TopicArea	csliLocateTopic(String meeting, Topic t)
abstract List<Pair<Double,TopicArea>>	csliSearchTopic(Topic t)
Topic	extractTopic(String meeting, Pair<XSDDateTime,XSDDateTime> boundaries) Extracts the topic discussed in meeting in a certain segment
Topic	extractTopic(String meeting, XSDDateTime time) Extracts the topic discussed in meeting around a specified time.
String[]	getQueries()
List<Topic>	getRelevantTopics(String query, String meeting, int k, boolean useThreshold) Computes the saved topics for a particular meeting that are the most relevant to a certain query.
abstract Set<XSDDateTime>	getTopicBreaks(String meeting) Computes the topic segmentation of the meeting.
Set<XSDDateTime>	getTopicBreaksByMITSegmentation(String meeting) Like getTopicBreaks(), but force segmentation via the MIT generative model.
abstract Set<XSDDateTime>	getTopicBreaksByMITSegmentation(String meeting, Integer numShifts) Like getTopicBreaks(), but force Stephane's lexical similarity method.
abstract Set<XSDDateTime>	getTopicBreaksByWordDistribution(String meeting) Like getTopicBreaks(), but force Stephane's lexical similarity method.
List<Topic>	getTopics(String meeting) Finds the named (i.e.
List<Topic>	getTopicsByMITSegmentation(String meeting) Like getTopics(), but force MIT's generative model.
List<Topic>	getTopicsByMITSegmentation(String meeting, Integer numShifts) Like getTopics(), but force MIT's generative model.
List<Topic>	getTopicsByWordDistribution(String meeting) Like getTopics(), but force Stephane's lexical similarity method.
boolean	isLearning()
abstract ArrayList<Pair<Topic,TopicArea>>	locateDoc(String meetingName)
Pair<XSDDateTime,XSDDateTime>	locateTopic(String meeting, Topic t) Locates the best matching area of a topic in the meeting
void	prepareForNewMeeting(String meetingName) Performs all necessary precomputations when a new meeting is added to the database.
void	setTopicOptions(boolean learning)
abstract boolean	wasDiscussed(Topic t, String meeting) Evaluates whether a certain topic was discussed or not during a past meeting.

Methods inherited from class csli.dialog.app.calo.topic.classification.topicextraction.CsliOnlineTopicSegmenterService

createTopic, createTopicModel, csliSearchTopic, csliSearchTopic, extractTopicBoundaries, findRelevantTopics, getGui, getRelevantTopics, isRelevant, isRelevant, searchTopic

Methods inherited from class java.lang.Object

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

Methods inherited from interface csli.dialog.app.calo.main.services.TopicSegmenterService

clear, extractTopicBoundaries, searchTopic

Methods inherited from interface csli.dialog.app.calo.main.services.OnlineTopicSegmenterService

createTopic, createTopicModel, getRelevantTopics, isRelevant, isRelevant

Methods inherited from interface csli.dialog.app.calo.main.Service

getGui, getTitle, isServiceAlive

Constructor Detail

CsliTopicSegmenterService

public CsliTopicSegmenterService()

Method Detail

getQueries

public String[] getQueries()

Overrides:

getQueries in class CsliOnlineTopicSegmenterService

getTopicBreaks

public abstract Set<XSDDateTime> getTopicBreaks(String meeting)

Description copied from interface: OnlineTopicSegmenterService

Computes the topic segmentation of the meeting. For OnlineTopicSegmenterService, this will necessarily be based on topic break classification. TopicSegmenterService may base the segmentation on lexical similarity or the MIT generative model depending on the config key topic.segmentation.method

Specified by:

getTopicBreaks in interface OnlineTopicSegmenterService

Specified by:

getTopicBreaks in interface TopicSegmenterService

Specified by:

getTopicBreaks in class CsliOnlineTopicSegmenterService

Parameters:

meeting - the meeting we want to segment

Returns:

the set of topic breaks

getTopicBreaksByWordDistribution

public abstract Set<XSDDateTime> getTopicBreaksByWordDistribution(String meeting)

Like getTopicBreaks(), but force Stephane's lexical similarity method.

Parameters:

meeting - the meeting we want to segment

Returns:

the set of topic breaks

See Also:

getTopicBreaks()

getTopicBreaksByMITSegmentation

public Set<XSDDateTime> getTopicBreaksByMITSegmentation(String meeting)

Like getTopicBreaks(), but force segmentation via the MIT generative model. The config key topic.segmentation.numShifts will be used to determine number of shifts (0 for average number, <0 for auto)

Parameters:

meeting - the meeting we want to segment

Returns:

the set of topic breaks

See Also:

getTopicBreaks(), getTopicBreaksByMITSegmentation(String,Integer)

getTopicBreaksByMITSegmentation

public abstract Set<XSDDateTime> getTopicBreaksByMITSegmentation(String meeting,
                                                                 Integer numShifts)

Like getTopicBreaks(), but force Stephane's lexical similarity method.

Parameters:

meeting - the meeting we want to segment

numShifts - if > 0, the fixed number of shifts required; if = 0, fix the number of shifts at the average; if < 0, use the average probability threshold

Returns:

the set of topic breaks

See Also:

getTopicBreaks()

getTopics

public List<Topic> getTopics(String meeting)

Description copied from interface: TopicSegmenterService

Finds the named (i.e. existing in the pool) topics discussed in a meeting. The implementation may base the segmentation used on direct shift classification, lexical similarity or the MIT generative model depending on the config key topic.segmentation.method

Specified by:

getTopics in interface TopicSegmenterService

Parameters:

meeting - The meeting to process.

Returns:

a List of Topics

getTopicsByWordDistribution

public List<Topic> getTopicsByWordDistribution(String meeting)

Like getTopics(), but force Stephane's lexical similarity method.

Parameters:

meeting - The meeting to process.

Returns:

a List of Topics

See Also:

getTopics(String)

getTopicsByMITSegmentation

public List<Topic> getTopicsByMITSegmentation(String meeting)

Like getTopics(), but force MIT's generative model. The config key topic.segmentation.numShifts will be used to determine number of shifts (0 for average number, <0 for auto)

Parameters:

meeting - The meeting to process.

Returns:

a List of Topics

See Also:

getTopics(String)

getTopicsByMITSegmentation

public List<Topic> getTopicsByMITSegmentation(String meeting,
                                              Integer numShifts)

Like getTopics(), but force MIT's generative model.

Parameters:

meeting - The meeting to process.

numShifts - if > 0, the fixed number of shifts required; if = 0, fix the number of shifts at the average; if < 0, use the average probability threshold

Returns:

a List of Topics

See Also:

getTopics(String)

getRelevantTopics

public List<Topic> getRelevantTopics(String query,
                                     String meeting,
                                     int k,
                                     boolean useThreshold)

Description copied from interface: TopicSegmenterService

Computes the saved topics for a particular meeting that are the most relevant to a certain query. Optionally exclude those for which relevance falls beneath a given threshold.

Specified by:

getRelevantTopics in interface TopicSegmenterService

Parameters:

query - The query we are searching.

meeting - The meeting to search.

k - Number of topics we want to return.

useThreshold - if false, return all topics; if true, only those whose relevance exceeds topic.extraction.relevanceThreshold.

Returns:

the list of at most k topics that are the most relevant to the query, most relevant first.

csliGetTopics

public abstract ArrayList<Pair<Topic,TopicArea>> csliGetTopics(String meeting)

Get the Topics discussed in a meeting (using the default method), before adding to/merging with the pool

Parameters:

meeting -

Returns:

a list of Pairs of Topic and TopicArea

csliGetTopicsByWordDistribution

public abstract ArrayList<Pair<Topic,TopicArea>> csliGetTopicsByWordDistribution(String meeting)

Get the Topics discussed in a meeting (forcing Stephane's lexical similarity method), before adding to/merging with the pool

Parameters:

meeting -

Returns:

a list of Pairs of Topic and TopicArea

See Also:

csliGetTopics(String)

csliGetTopicsByMITSegmentation

public abstract ArrayList<Pair<Topic,TopicArea>> csliGetTopicsByMITSegmentation(String meeting,
                                                                                Integer numShifts)

Get the Topics discussed in a meeting (forcing segmentation via the MIT generative model), before adding to/merging with the pool

Parameters:

meeting -

numShifts - if > 0, the fixed number of shifts required; if = 0, fix the number of shifts at the average; if < 0, use the average probability threshold

Returns:

a list of Pairs of Topic and TopicArea

See Also:

csliGetTopics(String)

csliGetTopicSimilarity

public Double csliGetTopicSimilarity(Topic a,
                                     Topic b)

locateTopic

public Pair<XSDDateTime,XSDDateTime> locateTopic(String meeting,
                                                 Topic t)

Description copied from interface: TopicSegmenterService

Locates the best matching area of a topic in the meeting

Specified by:

locateTopic in interface TopicSegmenterService

Parameters:

meeting - the meeting to search in

t - the topic we want to locate

Returns:

a pair of XSDDateTime giving the beginning and the end of the corresponding segment.

csliLocateTopic

public abstract TopicArea csliLocateTopic(String meeting,
                                          Topic t)

extractTopic

public Topic extractTopic(String meeting,
                          XSDDateTime time)

Description copied from interface: TopicSegmenterService

Extracts the topic discussed in meeting around a specified time. It extends the segment around that time to have a bigger coherent zone.

Specified by:

extractTopic in interface TopicSegmenterService

Parameters:

meeting - the meeting we are interested in.

time - the timestamp we focus on

Returns:

a Topic object showing the word repartition on the segment.

extractTopic

public Topic extractTopic(String meeting,
                          Pair<XSDDateTime,XSDDateTime> boundaries)

Description copied from interface: TopicSegmenterService

Extracts the topic discussed in meeting in a certain segment

Specified by:

extractTopic in interface TopicSegmenterService

Parameters:

meeting - the meeting we are interested in.

boundaries - the beginning and end of the segment of which we want to extract the topic.

Returns:

a Topic object showing the word repartition on that segment.

csliExtractTopic

public abstract Pair<Topic,TopicArea> csliExtractTopic(String meeting,
                                                       XSDDateTime time)

csliExtractTopic

public abstract Pair<Topic,TopicArea> csliExtractTopic(String meeting,
                                                       Pair<XSDDateTime,XSDDateTime> boundaries)

wasDiscussed

public abstract boolean wasDiscussed(Topic t,
                                     String meeting)

Description copied from interface: TopicSegmenterService

Evaluates whether a certain topic was discussed or not during a past meeting.

Specified by:

wasDiscussed in interface TopicSegmenterService

Parameters:

t - the topic

meeting - the meeting

Returns:

a boolean answering the question.

csliSearchTopic

public abstract List<Pair<Double,TopicArea>> csliSearchTopic(Topic t)

Specified by:

csliSearchTopic in class CsliOnlineTopicSegmenterService

prepareForNewMeeting

public void prepareForNewMeeting(String meetingName)

Description copied from interface: OnlineTopicSegmenterService

Performs all necessary precomputations when a new meeting is added to the database.

Specified by:

prepareForNewMeeting in interface OnlineTopicSegmenterService

Specified by:

prepareForNewMeeting in interface TopicSegmenterService

Overrides:

prepareForNewMeeting in class CsliOnlineTopicSegmenterService

Parameters:

meetingName - The name of the meeting we add in the corpus.

setTopicOptions

public void setTopicOptions(boolean learning)

Specified by:

setTopicOptions in interface OnlineTopicSegmenterService

Overrides:

setTopicOptions in class CsliOnlineTopicSegmenterService

isLearning

public boolean isLearning()

Specified by:

isLearning in interface OnlineTopicSegmenterService

Overrides:

isLearning in class CsliOnlineTopicSegmenterService

clear

public abstract void clear()

Description copied from interface: OnlineTopicSegmenterService

Clears all the cache files that have been generated by the TopicSegmenterService. Mostly for development purpose.

Specified by:

clear in interface OnlineTopicSegmenterService

Specified by:

clear in interface TopicSegmenterService

Specified by:

clear in class CsliOnlineTopicSegmenterService

locateDoc

public abstract ArrayList<Pair<Topic,TopicArea>> locateDoc(String meetingName)

createXML

public abstract void createXML(String meetingName)