crawler.transformer+fessXpathTransformer.xml Maven / Gradle / Ivy

Go to download
<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE components PUBLIC "-//DBFLUTE//DTD LastaDi 1.0//EN"
	"http://dbflute.org/meta/lastadi10.dtd">
<components>
	<include path="crawler/transformer_basic.xml"/>

	<component name="fessXpathTransformer" class="org.codelibs.fess.crawler.transformer.CustomFessXpathTransformer" instance="singleton">
		<property name="name">"fessXpathTransformer"</property>
		<property name="featureMap">defaultFeatureMap</property>
		<property name="propertyMap">defaultPropertyMap</property>
		<property name="childUrlRuleMap">htmlUrlRuleMap</property>
		<!--
		<property name="invalidUrlPattern">@java.util.regex.Pattern@compile("^\\s*javascript:|^\\s*mailto:|^\\s*irc:|^\\s*skype:|^\\s*callto:",@java.util.regex.Pattern@CASE_INSENSITIVE)</property>
		 -->
		<property name="convertUrlMap">
			{"feed:" : "http:"}
		</property>
		<!-- segment -->
		<postConstruct name="addFieldRule">
			<arg>"title"</arg>
			<arg>"//TITLE"</arg>
			<arg>true</arg>
		</postConstruct>
		<postConstruct name="addFieldRule">
			<arg>"important_content"</arg>
			<arg>"//*[self::H1 or self::H2 or self::H3]"</arg>
			<arg>true</arg>
		</postConstruct>
	</component>
</components>