org.qbroker.node

Class SelectNode

java.lang.Object

org.qbroker.node.Node

org.qbroker.node.SelectNode

All Implemented Interfaces:

MessageNode

public class SelectNode
extends Node

SelectNode parses the payload of JMS Messages and selects the portions of the content according to predefined rules. For each selected chunk of the content, SelectNode creates a new TextMessage and stores the data into it. It also copies the properties of the original message into the new message and passes it through the outlink of done. According to the rulesets, SelectNode routes the incoming messages into three outlinks: bypass for those original messages, nohit for those messages do not belong to any rulesets, failure for the messages failed at the selection process. However, the outlink of done can be shared with other outlinks. Since the payload of the message may be very large, SelectNode tries to ignore any filters on message body. It will not display message body of the incoming messages either.

SelectNode contains a number of predefined rulesets. These rulesets categorize messages into non-overlapping groups. Therefore, each ruleset defines a unique message group. The ruleset also defines the operations to select content from the messages and other parameters for the group. If a ruleset has the preferredOutLink defined, SelectNode will route the messages to its outlink without any actions. Otherwise, each selected item will be built into a message for delivery. StringProperty determines what properties to be copied over from the original message to the new messages. If it is null, nothing will be copied over. If it is empty, all properties will be copied over. Otherwise, only those properties defined in StringProperty will be copied over. If StoreField is defined for a ruleset, it must be a full path to a local file. SelectNode will retrieve the content from the external file. This is good in case the size of content is too large for a message. All the delivered new messages will be tracked by RULE_PEND of their rulesets.

The default operation is to split the content via the given delimiters. The split items will be selected for the ruleset. Alternatively, SelectNode also supports other 3x2 different ways to select items from the message payload. They are based on Pattern, XPath or JSONPath. For example, if Pattern is defined in a ruleset, SelectNode will parse the content with the pattern. All the matched items will be selected for the ruleset. If XPath is defined, SelectNode will use it to select DOM nodes from the XML document. The text representation of each selected DOM node will be stored to the outgoing message. Both Pattern and XPath support selections with dynamic variables. For either Pattern or XPath, the value can be a Map contains multiple static patterns or XPaths. SelectNode will use each of them to select items. Their keys will be set to the TagField on the new messages.

SelectNode also allows developers to plug-in their own methods to select. In this case, the full ClassName of the selection implementation and its SelectorArguments must be well defined in the rulesets. The requirement on the plug-ins is minimum. The class must have a public method of select(String text) that takes a String as the only argument. The returned object must be an array of String for selected items on success. Each of the array element will be stored into the newly created message. In case of failure, null should be returned. It should also have a constructor taking a Map, or a List, or just a String as the only argument for the configurations. SelectNode will invoke its public method to get the list of the selected items from either message body or certain message property of each incoming message.

If CountField is defined for a ruleset, the total count of selected items will be set to the specified field of the incoming messages. In case of the split operation with given delimiters, the total count will be set at the end of the operation. It means those new messages will not have this property set. For other cases, the total count will be set right after the items are selected. Therefore, the total count can be copied over to the newly created messages. The downstream can use this count to terminate the aggregation sessions or loops. In order to set the property at CountField, the incoming message must be writeable. In most of the cases, you can use the receivers to set a dummy property on all incoming messages since it will make them writeable.

If XAMode is defined in a rulset, it will overwrite the default value defined on the node level. If the XAMode of a ruleset is on, SelectNode will make sure all the outgoing messages are delivered before routing the original message to bypass. Otherwise, the original message will be routed to bypass right after those newly created messages dispatched. In this case, SelectNode will not wait on their delivery.

You are free to choose any names for the four fixed outlinks. But SelectNode always assumes the first outlink for done, the second for bypass, the third for failure and the last for nohit. They all can share the same name, too. For the first outlink, if it is partitioned, you must define the same Partition explicitly on every selection ruleset.

Author:

yannanlu@yahoo.com

Field Summary

Fields inherited from class org.qbroker.node.Node

assetList, bufferSize, callbackMethod, capacity, cbw, cells, cfgList, chkptExpiration, debug, DEBUG_COLL, DEBUG_DIFF, DEBUG_FBAK, DEBUG_INIT, DEBUG_PASS, DEBUG_PROP, DEBUG_REPT, DEBUG_UPDT, displayMask, displayPropertyName, linkName, LOAD_HIGH, LOAD_LOW, LOAD_MEDIUM, msgList, name, operation, RESET_ALL, RESET_MAP, RESET_NONE, RESET_SOME, ruleList, status, TYPE_ACTION, TYPE_AGGREGATE, TYPE_BYPASS, TYPE_CACHE, TYPE_COLLECT, TYPE_DELIVER, TYPE_FORMAT, TYPE_JSONPATH, TYPE_JSONT, TYPE_MAPREDUCE, TYPE_NONE, TYPE_PARSER, TYPE_PICKUP, TYPE_PIPE, TYPE_PUBLISH, TYPE_ROUNDROBIN, TYPE_SCREEN, TYPE_SCRIPT, TYPE_SERVICE, TYPE_SORT, TYPE_SPLIT, TYPE_SPREAD, TYPE_WEIGHTED, TYPE_XPATH, TYPE_XSLT, waitTime, xaMode

Fields inherited from interface org.qbroker.node.MessageNode

ASSET_THR, ASSET_URI, ASSET_XQ, META_MAP, META_MSG, META_OUT, META_RULE, META_XQ, MSG_BID, MSG_CID, MSG_OID, MSG_RID, MSG_TID, MSG_TIME, NODE_CLOSED, NODE_DISABLED, NODE_PAUSE, NODE_READY, NODE_RETRYING, NODE_RUNNING, NODE_STANDBY, NODE_STOPPED, OUT_CAPACITY, OUT_COUNT, OUT_DEQ, OUT_EXTRA, OUT_LENGTH, OUT_MODE, OUT_NRULE, OUT_OFFSET, OUT_ORULE, OUT_QDEPTH, OUT_QIPPS, OUT_QSTATUS, OUT_QTIME, OUT_SIZE, OUT_STATUS, OUT_TIME, RULE_COUNT, RULE_DMASK, RULE_EXTRA, RULE_GID, RULE_MODE, RULE_OID, RULE_OPTION, RULE_PEND, RULE_PID, RULE_SIZE, RULE_STATUS, RULE_TIME, RULE_TTL

Constructor Summary

Constructors

Constructor and Description
SelectNode(java.util.Map props)

Method Summary

Modifier and Type	Method and Description
void	close() It closes all resources and cleans up all list, cache, metadata, etc.
protected void	finalize()
protected java.util.Map<java.lang.String,java.lang.Object>	initRuleset(long tm, java.util.Map ph, long[] ruleInfo) It initializes a new ruleset with the ruleInfo and returns the rule upon success.
void	propagate(XQueue in, XQueue[] out) picks up a message from input queue and evaluates its content to decide which output queue to propagate
int	removeRule(java.lang.String key, XQueue in) It removes the rule from the ruleList and returns the rule id upon success.
int	replaceRule(java.lang.String key, java.util.Map ph, XQueue in) It replaces the existing rule of the key and returns its id upon success.

Methods inherited from class org.qbroker.node.Node

addRule, callback, checkpoint, collectAll, containsRule, containsXQ, feedback, forcethru, getCapacity, getCapacity, getDebugMode, getDisplayMask, getDisplayProperties, getLinkName, getLinkName, getMetaData, getName, getNumberOfOutLinks, getNumberOfRules, getOperation, getOutLinkBoundary, getRuleName, getStatus, getXAMode, getXAMode, internalXQSupported, list, listPendings, passthru, refreshRule, renameRule, resetMetaData, restore, rotateRule, setDebugMode, setDisplayMask, setDisplayProperties, setStatus, show, takeback, takebackEnabled, updateParameters

Methods inherited from class java.lang.Object

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

Constructor Detail

SelectNode

public SelectNode(java.util.Map props)

Method Detail

initRuleset

protected java.util.Map<java.lang.String,java.lang.Object> initRuleset(long tm,
                                                                       java.util.Map ph,
                                                                       long[] ruleInfo)

It initializes a new ruleset with the ruleInfo and returns the rule upon success. Otherwise, it throws an exception or returns null.

Specified by:

initRuleset in class Node

removeRule

public int removeRule(java.lang.String key,
                      XQueue in)

It removes the rule from the ruleList and returns the rule id upon success. It is not MT-Safe.

Specified by:

removeRule in interface MessageNode

Overrides:

removeRule in class Node

replaceRule

public int replaceRule(java.lang.String key,
                       java.util.Map ph,
                       XQueue in)

It replaces the existing rule of the key and returns its id upon success. It is not MT-Safe.

Specified by:

replaceRule in interface MessageNode

Overrides:

replaceRule in class Node

propagate

public void propagate(XQueue in,
                      XQueue[] out)
               throws javax.jms.JMSException

picks up a message from input queue and evaluates its content to decide which output queue to propagate

Specified by:

propagate in interface MessageNode

Specified by:

propagate in class Node

Throws:

javax.jms.JMSException

close

public void close()

Description copied from class: Node

It closes all resources and cleans up all list, cache, metadata, etc. Once it is closed, it can not be used any more.

Specified by:

close in interface MessageNode

Overrides:

close in class Node

finalize

protected void finalize()

Overrides:

finalize in class Node