de.spieleck.app.jacson.filter
Class WebSessionFilter

java.lang.Object
  de.spieleck.app.jacson.PluginBase
      de.spieleck.app.jacson.filter.FilterBase
          de.spieleck.app.jacson.filter.ConstFilter
              de.spieleck.app.jacson.filter.WebSessionFilter

All Implemented Interfaces:

ConfigVerify.Acceptor, FilterNames, JacsonChunkDrain, JacsonFilter, JacsonRegistrable, JacsonReporting, JacsonStately

public class WebSessionFilter

extends ConstFilter

A hack to derive something like a session from a web log file by using the well known ip=user session till timeout logic. This isn't working great, but it is working.
Note: The implementation depends on the fact that timestamps are monotonical increasing in the logfile. If this isn't the case we are in real trouble.
Note: The implementation currently cannot merge sessions between multiple input streams (separated by null chunks)
The plugin understands the following parameters:

• timeout - the timeout to use to terminate sessions in Milliseconds. Optional, defaults to 1200000 = 20 minutes.
• dateregexp - an regular expression to fetch the date/time date from a chunk. Required.
• clientregexp - another regular expression to fetch the client from the chunk. Required. Common for Weblogfiles is to use either IP or IP+BrowserID, better er URL-Session-IDs if you have them.
• All other options for RegExpUtil. These are shared by above regexps.
• dateformat - Format for the datestring to be parsed in the java.text.SimpleDateFormat syntax. Defaults to yyyy-MM-dd HH:mm:ss

Author:

fsn

See Also:

RegExpUtil

Field Summary

static java.lang.String	CLIENTREGEXP_NODE
protected java.text.DateFormat	dateFormat Date parse format
static java.lang.String	DATEFORMAT_NODE
protected org.apache.oro.text.regex.Pattern	datePattern Date regular expression
static java.lang.String	DATEREGEXP_NODE
static java.lang.String	DEFAULT_DATEFORMAT
protected org.apache.oro.text.regex.Pattern	idPattern ID regular expression
protected long	lastTime Object caching for speedup (flyweight pattern)
protected java.lang.Long	lastTimeLong Object caching for speedup (flyweight pattern)
protected RegExpUtil	rutil Utility for regular expressions
protected int	timeout The timeout assumed for a session
static java.lang.String	TIMEOUT_NODE
static java.lang.String	WSF

Fields inherited from class de.spieleck.app.jacson.filter.ConstFilter

inverse

Fields inherited from class de.spieleck.app.jacson.filter.FilterBase

drain

Fields inherited from interface de.spieleck.app.jacson.filter.FilterNames

COUNT_NODE, INVERSE_NODE

Constructor Summary

WebSessionFilter()

Method Summary

boolean	accept(ConfigNode node)
protected void	flushSessions(long limit) Clean up the stored sessions up to a certain point in history.
void	init(ConfigNode node, JacsonRegistry registry)
protected java.lang.String	matching(java.lang.String chunk, org.apache.oro.text.regex.Pattern p) Convenience method to extract a string from a chunk via a match.
void	putChunk(java.lang.String chunk) Feed a single chunk into this filter

Methods inherited from class de.spieleck.app.jacson.filter.ConstFilter

summary

Methods inherited from class de.spieleck.app.jacson.filter.FilterBase

setDrain

Methods inherited from class de.spieleck.app.jacson.PluginBase

getRegReport, getRegState, registerReport, registerState

Methods inherited from class java.lang.Object

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

Methods inherited from interface de.spieleck.app.jacson.JacsonStately

registerState

Methods inherited from interface de.spieleck.app.jacson.JacsonReporting

registerReport

Field Detail

TIMEOUT_NODE

public static final java.lang.String TIMEOUT_NODE

See Also:

Constant Field Values

DATEREGEXP_NODE

public static final java.lang.String DATEREGEXP_NODE

See Also:

Constant Field Values

CLIENTREGEXP_NODE

public static final java.lang.String CLIENTREGEXP_NODE

See Also:

Constant Field Values

DATEFORMAT_NODE

public static final java.lang.String DATEFORMAT_NODE

See Also:

Constant Field Values

DEFAULT_DATEFORMAT

public static final java.lang.String DEFAULT_DATEFORMAT

See Also:

Constant Field Values

WSF

public static final java.lang.String WSF

See Also:

Constant Field Values

rutil

protected RegExpUtil rutil

Utility for regular expressions

timeout

protected int timeout

The timeout assumed for a session

datePattern

protected org.apache.oro.text.regex.Pattern datePattern

Date regular expression

dateFormat

protected java.text.DateFormat dateFormat

Date parse format

idPattern

protected org.apache.oro.text.regex.Pattern idPattern

ID regular expression

lastTime

protected long lastTime

Object caching for speedup (flyweight pattern)

lastTimeLong

protected java.lang.Long lastTimeLong

Object caching for speedup (flyweight pattern)

Constructor Detail

WebSessionFilter

public WebSessionFilter()

Method Detail

init

public void init(ConfigNode node,
                 JacsonRegistry registry)
          throws JacsonConfigException

Specified by:

init in interface JacsonRegistrable

Overrides:

init in class ConstFilter

Throws:

JacsonConfigException

accept

public boolean accept(ConfigNode node)

Specified by:

accept in interface ConfigVerify.Acceptor

Overrides:

accept in class ConstFilter

putChunk

public void putChunk(java.lang.String chunk)
              throws JacsonException

Description copied from interface: JacsonChunkDrain

Feed a single chunk into this filter

Specified by:

putChunk in interface JacsonChunkDrain

Overrides:

putChunk in class ConstFilter

Throws:

JacsonException

flushSessions

protected void flushSessions(long limit)
                      throws JacsonException

Clean up the stored sessions up to a certain point in history.

Throws:

JacsonException

matching

protected java.lang.String matching(java.lang.String chunk,
                                    org.apache.oro.text.regex.Pattern p)

Convenience method to extract a string from a chunk via a match.