Package dk.netarkivet.heritrix3.monitor

Class NASEnvironment

java.lang.Object

dk.netarkivet.heritrix3.monitor.NASEnvironment

public class NASEnvironment
extends java.lang.Object

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
static class	NASEnvironment.StringMatcher

Field Summary

Fields

Modifier and Type	Field	Description
static java.lang.String	contextPath
java.lang.String	h3AdminName
java.lang.String	h3AdminPassword
java.util.List<NASEnvironment.StringMatcher>	h3HostPortAllowRegexList
Heritrix3JobMonitorThread	h3JobMonitorThread
HttpLocaleHandler	httpLocaleUtils
I18n	I18N
protected java.lang.String	login_template_name
protected com.antiaction.common.templateengine.login.LoginTemplateHandler<NASUser>	loginHandler
java.lang.String	NAS_GROOVY_SCRIPT
static java.lang.String	servicePath
protected javax.servlet.ServletConfig	servletConfig	servletConfig.
com.antiaction.common.templateengine.TemplateMaster	templateMaster
java.io.File	tempPath

Constructor Summary

Constructors

Constructor	Description
NASEnvironment(javax.servlet.ServletContext servletContext, javax.servlet.ServletConfig theServletConfig)

Method Summary

Modifier and Type	Method	Description
void	cleanup()	Do some cleanup.
java.util.stream.Stream<java.lang.String>	getCrawledUrls(long jobId, Heritrix3JobMonitor h3Job)	Get the (attempted) crawled URLs of the crawllog for the running job with the given job id
java.lang.String	getResourceAsString(java.lang.String resource)
boolean	isH3HostnamePortEnabled(java.lang.String h3HostnamePort)
boolean	jobHarvestsDomain(long jobId, java.lang.String domainName, Heritrix3JobMonitor h3Job)	Find out whether the given job harvests given domain.
void	replaceH3HostnamePortRegexList(java.util.List<java.lang.String> h3HostnamePortRegexList, java.util.List<java.lang.String> invalidPatternsList)
void	start()

Methods inherited from class java.lang.Object

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

Field Detail

NAS_GROOVY_SCRIPT

public java.lang.String NAS_GROOVY_SCRIPT

servletConfig

protected javax.servlet.ServletConfig servletConfig

servletConfig.

templateMaster

public com.antiaction.common.templateengine.TemplateMaster templateMaster

login_template_name

protected java.lang.String login_template_name

loginHandler

protected com.antiaction.common.templateengine.login.LoginTemplateHandler<NASUser> loginHandler

tempPath

public java.io.File tempPath

h3AdminName

public java.lang.String h3AdminName

h3AdminPassword

public java.lang.String h3AdminPassword

h3JobMonitorThread

public Heritrix3JobMonitorThread h3JobMonitorThread

contextPath

public static java.lang.String contextPath

servicePath

public static java.lang.String servicePath

httpLocaleUtils

public HttpLocaleHandler httpLocaleUtils

h3HostPortAllowRegexList

public final java.util.List<NASEnvironment.StringMatcher> h3HostPortAllowRegexList

I18N

public final I18n I18N

Constructor Detail

NASEnvironment

public NASEnvironment(javax.servlet.ServletContext servletContext,
                      javax.servlet.ServletConfig theServletConfig)
               throws javax.servlet.ServletException

Throws:

javax.servlet.ServletException

Method Detail

getResourceAsString

public java.lang.String getResourceAsString(java.lang.String resource)
                                     throws java.io.IOException

Throws:

java.io.IOException

start

public void start()

cleanup

public void cleanup()

Do some cleanup. This waits for the different workflow threads to stop running.

replaceH3HostnamePortRegexList

public void replaceH3HostnamePortRegexList(java.util.List<java.lang.String> h3HostnamePortRegexList,
                                           java.util.List<java.lang.String> invalidPatternsList)

isH3HostnamePortEnabled

public boolean isH3HostnamePortEnabled(java.lang.String h3HostnamePort)

getCrawledUrls

public java.util.stream.Stream<java.lang.String> getCrawledUrls(long jobId,
                                                                Heritrix3JobMonitor h3Job)

Get the (attempted) crawled URLs of the crawllog for the running job with the given job id

Parameters:

jobId - Id of the running job

h3Job - Heritrix3JobMonitor from which to get the job for the given jobId

Returns:

The (attempted) crawled URLs of the crawllog for given job

jobHarvestsDomain

public boolean jobHarvestsDomain(long jobId,
                                 java.lang.String domainName,
                                 Heritrix3JobMonitor h3Job)

Find out whether the given job harvests given domain.

Parameters:

jobId - The job

domainName - The domain

Returns:

whether the given job harvests given domain