org.apache.manifoldcf.crawler.jobs
Class HopCount.DocumentNode

java.lang.Object
  org.apache.manifoldcf.crawler.jobs.HopCount.DocumentNode

Enclosing class:

HopCount

protected static class HopCount.DocumentNode
extends java.lang.Object

This class keeps track of the data associated with a node in the hash map. This basically includes the following: - the document identifier - the 'question' that was asked, which has the form (link type, maximum distance) - possibly the 'answer' to the question, which is either ">(maximum distance)", or a number. - references to the nodes which care about this answer, if they are still queued. - summary of the information we've gathered from children so far (if answer not known yet) - references to the children of this node that can affect the answer, including link details (if answer not known yet)

Field Summary

protected HopCount.Answer	bestPossibleAnswer The best (lowest) possible answer value for this node.
protected java.util.Map	childReferences Child node references.
protected int	databaseAnswerValue This is the original answer (if any), which is the current value in the database
protected java.lang.Long	databaseRow The original database row, if any
protected boolean	isComplete Answer complete flag.
protected java.util.Map	parentsWhoCare Parent nodes who care (i.e.
protected HopCount.Question	question The question.
protected HopCount.Answer	startingAnswer The answer, as calculated up to the level of all the completed children, which will not include incomplete child references of this node.
protected HopCount.Answer	trialAnswer The current best answer.
protected boolean	writeNeeded This flag is meaningful only if the complete flag is set.

Constructor Summary

HopCount.DocumentNode(HopCount.Question question)
Create a document node.

Method Summary

void	addChild(HopCount.NodeReference childRef) Add a child reference.
void	addParent(HopCount.DocumentNode parent) Add a parent who should be notified if this node's answer changes.
void	clearChildReferences() Clear child references.
void	clearWriteNeeded() Clear the 'write needed' flag, to prevent another write.
HopCount.Answer	getBestPossibleAnswer() Get best possible answer
java.util.Iterator	getCurrentChildren() Iterate through current children.
java.util.Iterator	getCurrentParents() Iterate through all current parents.
java.lang.Long	getDatabaseRow() Get the database row
int	getDatabaseValue() Get the database answer value
int	getFinalAnswer() Get an answer that's final.
HopCount.Question	getQuestion() Get the question.
HopCount.Answer	getStartingAnswer() Get the starting (base) answer.
HopCount.Answer	getTrialAnswer() Get the current best answer.
boolean	hasChildren() Check if there are children.
boolean	isAnswerComplete() Check if the answer is complete.
boolean	isAnswerNeeded() Check if answer is still needed.
boolean	isComplete() Check if the node is complete, given the question it represents.
boolean	isWriteNeeded() Check if a write of the answer is needed to the database
void	makeComplete() Mark the node as being "complete", with a write needed.
void	makeCompleteNoWrite() Mark the answer as being "complete", and not needing a write.
void	removeChild(HopCount.NodeReference childRef) Remove a child reference.
void	removeParent(HopCount.DocumentNode parent) Remove a parent.
void	reset() Reset back to an "unknown" state.
void	setBestPossibleAnswer(HopCount.Answer answer) Set best possible answer
void	setSource(java.lang.Long rowID, int answerValue) Set the database row and answer value
void	setStartingAnswer(HopCount.Answer answer) Set the starting (base) answer.
void	setTrialAnswer(HopCount.Answer answer) Set the answer for this node.

Methods inherited from class java.lang.Object

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

Field Detail

question

protected HopCount.Question question

The question.

databaseAnswerValue

protected int databaseAnswerValue

This is the original answer (if any), which is the current value in the database

databaseRow

protected java.lang.Long databaseRow

The original database row, if any

startingAnswer

protected HopCount.Answer startingAnswer

The answer, as calculated up to the level of all the completed children, which will not include incomplete child references of this node. This is a starting point for every reassessment of this node's current answer. It is adjust only when additional children are noted as being complete.

trialAnswer

protected HopCount.Answer trialAnswer

The current best answer. This takes into account the current status of all the child nodes. If the node is not complete, then the answer must be viewed as being less than or equal to this value.

bestPossibleAnswer

protected HopCount.Answer bestPossibleAnswer

The best (lowest) possible answer value for this node. This value is calculated based on the known child link structure of a node, and can only increase. The value will start low (at 0) and will climb as more knowledge is gained, as the children's best possible answer value increases upon re-evaluation. When the trial answer (above) reaches a value equal to the best possible value, then the node will be immediately marked as "complete", and further processing will be considered unnecessary. As far as dependencies are concerned, the bestPossibleAnswer includes dependencies that have gone into its assessment. These dependencies represent what would need to be changed to invalidate the answer as it stands. (Invalidation means that a smaller best possible answer would be possible, so only add dependencies would need consideration.)

isComplete

protected boolean isComplete

Answer complete flag. Will be set to true only if the value of "trialAnswer" is deemed final.

writeNeeded

protected boolean writeNeeded

This flag is meaningful only if the complete flag is set.

parentsWhoCare

protected java.util.Map parentsWhoCare

Parent nodes who care (i.e. are still queued). This map contains DocumentNode objects.

childReferences

protected java.util.Map childReferences

Child node references. This is a reference to an actual document node object which has a parent reference back to this one. If the child node is modified, there is an obligation to cause the parent node to be re-evaluated. The re-evaluation process examines all child nodes and may adjust the status of the trial answer, and may indeed even remove the reference to the child. This map contains NodeReference objects.

Constructor Detail

HopCount.DocumentNode

public HopCount.DocumentNode(HopCount.Question question)

Create a document node. This will happen only if there is no comparable one already in the hash.

Method Detail

getQuestion

public HopCount.Question getQuestion()

Get the question.

reset

public void reset()

Reset back to an "unknown" state.

clearChildReferences

public void clearChildReferences()

Clear child references.

hasChildren

public boolean hasChildren()

Check if there are children.

getFinalAnswer

public int getFinalAnswer()

Get an answer that's final. Returns "unknown" if the current answer is incomplete.

isAnswerComplete

public boolean isAnswerComplete()

Check if the answer is complete. Returns true if the answer is complete.

isComplete

public boolean isComplete()

Check if the node is complete, given the question it represents.

isWriteNeeded

public boolean isWriteNeeded()

Check if a write of the answer is needed to the database

isAnswerNeeded

public boolean isAnswerNeeded()

Check if answer is still needed.

getBestPossibleAnswer

public HopCount.Answer getBestPossibleAnswer()

Get best possible answer

setBestPossibleAnswer

public void setBestPossibleAnswer(HopCount.Answer answer)

Set best possible answer

getTrialAnswer

public HopCount.Answer getTrialAnswer()

Get the current best answer.

setTrialAnswer

public void setTrialAnswer(HopCount.Answer answer)

Set the answer for this node.

getStartingAnswer

public HopCount.Answer getStartingAnswer()

Get the starting (base) answer.

setStartingAnswer

public void setStartingAnswer(HopCount.Answer answer)

Set the starting (base) answer.

makeComplete

public void makeComplete()

Mark the node as being "complete", with a write needed.

makeCompleteNoWrite

public void makeCompleteNoWrite()

Mark the answer as being "complete", and not needing a write.

addParent

public void addParent(HopCount.DocumentNode parent)

Add a parent who should be notified if this node's answer changes. The parent is responsible for figuring out when this reference should be removed.

clearWriteNeeded

public void clearWriteNeeded()

Clear the 'write needed' flag, to prevent another write.

addChild

public void addChild(HopCount.NodeReference childRef)

Add a child reference.

Parameters:

childRef - is the child node reference to add.

removeChild

public void removeChild(HopCount.NodeReference childRef)

Remove a child reference.

Parameters:

childRef - is the child node reference to remove.

removeParent

public void removeParent(HopCount.DocumentNode parent)

Remove a parent. This method will get called when the parent's answer no longer can be affected by this child's answer (probably because the child's answer has become complete).

getCurrentParents

public java.util.Iterator getCurrentParents()

Iterate through all current parents. This is an iterator over DocumentNode objects.

getCurrentChildren

public java.util.Iterator getCurrentChildren()

Iterate through current children. This is an iterator over NodeReference objects.

setSource

public void setSource(java.lang.Long rowID,
                      int answerValue)

Set the database row and answer value

getDatabaseRow

public java.lang.Long getDatabaseRow()

Get the database row

getDatabaseValue

public int getDatabaseValue()

Get the database answer value