prajna.structure

Class DocCorpus

java.lang.Object

prajna.structure.DocCorpus

Direct Known Subclasses:

PollingCorpus

public class DocCorpus
extends java.lang.Object

Represents a collection of document data elements. This class provides utility methods for generating different data views of the corpus. It is designed to be sub-classed, with subclasses implementing specific interfaces to particular document sets.

Author:

Edward Swing

Constructor Summary

Constructors

Constructor and Description
DocCorpus()

Method Summary

Methods

Modifier and Type	Method and Description
protected void	addDocData(DocData doc) Add a DocData element to the corpus.
protected void	clearCorpus() This method clears all data from the corpus.
DocCorpus	createSubCorpus(java.util.Collection<java.lang.String> keys) Derive a sub-corpus of documents which match the given set of keys.
DocData	findDocData(java.lang.String name) Find the Document data for the document with the given name.
protected java.util.Set<java.lang.String>	getAllKeys() Gets all document keys in all documents within this DocCorpus
java.util.Set<DocData>	getCorpus() Gets the corpus.
int	getCorpusSize() Gets the corpus size, which is the number of documents in the corpus.
Graph<java.lang.String,AccumEdge<DocData>>	getKeyGraph() Get the key graph.
Graph<DocData,Weighted>	getSimilarityGraph() Get the similarity graph.
boolean	hasMoreData() Return whether the underlying data source has more data to be loaded into this document corpus.
boolean	isGraphDynamic() Get whether the graphs are dynamically created by this corpus.
void	loadCorpus() Load the document corpus from a data repository.
void	removeDoc(DocData doc) Remove a document from the corpus
void	setGraphDynamic(boolean dynamic) Set whether the graphs are dynamically created.

Methods inherited from class java.lang.Object

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

Constructor Detail

DocCorpus

public DocCorpus()

Method Detail

addDocData

protected void addDocData(DocData doc)

Add a DocData element to the corpus.

Parameters:

doc - the new document data element

clearCorpus

protected void clearCorpus()

This method clears all data from the corpus. Use with care

createSubCorpus

public DocCorpus createSubCorpus(java.util.Collection<java.lang.String> keys)

Derive a sub-corpus of documents which match the given set of keys. The resulting DocCorpus will contain a subset of the documents within this corpus. A document will be added to the sub-corpus if it matches any of the keys provided.

Parameters:

keys - the keys to match

Returns:

the derived subCorpus

findDocData

public DocData findDocData(java.lang.String name)

Find the Document data for the document with the given name.

Parameters:

name - the name of the document

Returns:

the DocData element for the document, or null if there is no such document

getAllKeys

protected java.util.Set<java.lang.String> getAllKeys()

Gets all document keys in all documents within this DocCorpus

Returns:

the union of all document keys from all documents.

getCorpus

public java.util.Set<DocData> getCorpus()

Gets the corpus.

Returns:

a defensive copy of the corpus

getCorpusSize

public int getCorpusSize()

Gets the corpus size, which is the number of documents in the corpus.

Returns:

the corpus size

getKeyGraph

public Graph<java.lang.String,AccumEdge<DocData>> getKeyGraph()

Get the key graph. The key graph is a graph with a node for each key in the corpus. The edges are accumulations of documents which include the two keys.

Returns:

the key graph

getSimilarityGraph

public Graph<DocData,Weighted> getSimilarityGraph()

Get the similarity graph. The similarity graph is a graph with a node for each document. The edges between document nodes are NamedWeights. The weights are equal to the number of keys that the two nodes have in common.

Returns:

the Similarity graph

hasMoreData

public boolean hasMoreData()

Return whether the underlying data source has more data to be loaded into this document corpus. The default behavior always returns false.

Returns:

false

isGraphDynamic

public boolean isGraphDynamic()

Get whether the graphs are dynamically created by this corpus.

Returns:

true if the graphs are dynamically created, false if they are created at request time.

loadCorpus

public void loadCorpus()

Load the document corpus from a data repository. The default implementation does nothing. Specific implementations of a document corpus should override this method.

removeDoc

public void removeDoc(DocData doc)

Remove a document from the corpus

Parameters:

doc - the document to remove

setGraphDynamic

public void setGraphDynamic(boolean dynamic)

Set whether the graphs are dynamically created. If the dynamic flag is set, the graphs are updated whenever a new document is added to the corpus. This means the updates may take a bit longer, and the corpus requires more memory, but the responses to retrieve the graph are instantaneous. If the dynamic flag is not set, the corpus loads faster, but creating the graphs might take significant time. A particular corpus can set the flag based upon expected usage patterns. By default, the flag is true.

Parameters:

dynamic - the flag indicating whether to generate the key graph and similarity graph dynamically.