org.wandora.application.tools.extractors

Class RSSExtractor.RSSParser

java.lang.Object

org.wandora.application.tools.extractors.RSSExtractor.RSSParser

All Implemented Interfaces:

org.xml.sax.ContentHandler, org.xml.sax.ErrorHandler

Enclosing class:

RSSExtractor

private static class RSSExtractor.RSSParser
extends java.lang.Object
implements org.xml.sax.ContentHandler, org.xml.sax.ErrorHandler

Field Summary

Fields

Modifier and Type	Field and Description
static java.lang.String	CHANNEL_CATEGORY_SI
static java.lang.String	CHANNEL_COPYRIGHT_SI
static java.lang.String	CHANNEL_DESCRIPTION_SI
static java.lang.String	CHANNEL_DOCS_SI
static java.lang.String	CHANNEL_GENERATOR_SI
static java.lang.String	CHANNEL_IMAGE_DESCRIPTION_SI
static java.lang.String	CHANNEL_IMAGE_LINK_SI
static java.lang.String	CHANNEL_IMAGE_SI
static java.lang.String	CHANNEL_ITEM_AUTHOR_SI
static java.lang.String	CHANNEL_ITEM_CATEGORY_SI
static java.lang.String	CHANNEL_ITEM_DESCRIPTION_SI
static java.lang.String	CHANNEL_ITEM_GUID_SI
static java.lang.String	CHANNEL_ITEM_LINK_SI
static java.lang.String	CHANNEL_ITEM_PUBDATE_SI
static java.lang.String	CHANNEL_ITEM_SI
static java.lang.String	CHANNEL_ITEM_TITLE_SI
static java.lang.String	CHANNEL_LANGUAGE_SI
static java.lang.String	CHANNEL_LASTBUILDDATE_SI
static java.lang.String	CHANNEL_LINK_SI
static java.lang.String	CHANNEL_MANAGINGEDITOR_SI
static java.lang.String	CHANNEL_PUBDATE_SI
static java.lang.String	CHANNEL_RATING_SI
static java.lang.String	CHANNEL_SI
static java.lang.String	CHANNEL_TEXTINPUT_SI
static java.lang.String	CHANNEL_TTL_SI
static java.lang.String	CHANNEL_WEBMASTER_SI
int	channelCount
private java.lang.String	data_channel_category
private java.lang.String	data_channel_copyright
private java.lang.String	data_channel_description
private java.lang.String	data_channel_docs
private java.lang.String	data_channel_generator
private java.lang.String	data_channel_image_description
private java.lang.String	data_channel_image_link
private java.lang.String	data_channel_image_title
private java.lang.String	data_channel_image_url
private java.lang.String	data_channel_item_author
private java.lang.String	data_channel_item_category
private java.lang.String	data_channel_item_description
private java.lang.String	data_channel_item_guid
private java.lang.String	data_channel_item_link
private java.lang.String	data_channel_item_pubdate
private java.lang.String	data_channel_item_title
private java.lang.String	data_channel_language
private java.lang.String	data_channel_lastbuilddate
private java.lang.String	data_channel_link
private java.lang.String	data_channel_managingeditor
private java.lang.String	data_channel_pubdate
private java.lang.String	data_channel_rating
private java.lang.String	data_channel_title
private java.lang.String	data_channel_ttl
private java.lang.String	data_channel_webmaster
static java.lang.String	DATE_SI
int	itemCount
static boolean	MAKE_LINK_OCCURRENCE
static boolean	MAKE_LINK_SUBJECT_IDENTIFIER
static boolean	MAKE_LINK_SUBJECT_LOCATOR
static boolean	MAKE_SUBCLASS_OF_WANDORA_CLASS
private RSSExtractor	parent
int	progress
static java.lang.String	RSS_SI
static java.lang.String	SIPREFIX
private int	state
private static int	STATE_CHANNEL
private static int	STATE_CHANNEL_CATEGORY
private static int	STATE_CHANNEL_COPYRIGHT
private static int	STATE_CHANNEL_DESCRIPTION
private static int	STATE_CHANNEL_DOCS
private static int	STATE_CHANNEL_GENERATOR
private static int	STATE_CHANNEL_IMAGE
private static int	STATE_CHANNEL_IMAGE_DESCRIPTION
private static int	STATE_CHANNEL_IMAGE_LINK
private static int	STATE_CHANNEL_IMAGE_TITLE
private static int	STATE_CHANNEL_IMAGE_URL
private static int	STATE_CHANNEL_ITEM
private static int	STATE_CHANNEL_ITEM_AUTHOR
private static int	STATE_CHANNEL_ITEM_CATEGORY
private static int	STATE_CHANNEL_ITEM_DESCRIPTION
private static int	STATE_CHANNEL_ITEM_GUID
private static int	STATE_CHANNEL_ITEM_LINK
private static int	STATE_CHANNEL_ITEM_PUBDATE
private static int	STATE_CHANNEL_ITEM_TITLE
private static int	STATE_CHANNEL_LANGUAGE
private static int	STATE_CHANNEL_LASTBUILDDATE
private static int	STATE_CHANNEL_LINK
private static int	STATE_CHANNEL_MANAGINGEDITOR
private static int	STATE_CHANNEL_PUBDATE
private static int	STATE_CHANNEL_RATING
private static int	STATE_CHANNEL_TEXTINPUT
private static int	STATE_CHANNEL_TITLE
private static int	STATE_CHANNEL_TTL
private static int	STATE_CHANNEL_WEBMASTER
private static int	STATE_RSS
private static int	STATE_START
static java.lang.String	TAG_AUTHOR
static java.lang.String	TAG_CATEGORY
static java.lang.String	TAG_CHANNEL
static java.lang.String	TAG_COPYRIGHT
static java.lang.String	TAG_DESCRIPTION
static java.lang.String	TAG_DOCS
static java.lang.String	TAG_GENERATOR
static java.lang.String	TAG_GUID
static java.lang.String	TAG_IMAGE
static java.lang.String	TAG_ITEM
static java.lang.String	TAG_LANGUAGE
static java.lang.String	TAG_LASTBUILDDATE
static java.lang.String	TAG_LINK
static java.lang.String	TAG_MANAGINGEDITOR
static java.lang.String	TAG_PUBDATE
static java.lang.String	TAG_RATING
static java.lang.String	TAG_RSS
static java.lang.String	TAG_TEXTINPUT
static java.lang.String	TAG_TITLE
static java.lang.String	TAG_TTL
static java.lang.String	TAG_URL
static java.lang.String	TAG_WEBMASTER
private Topic	theChannel
private Topic	theItem
private TopicMap	tm

Constructor Summary

Constructors

Constructor and Description
RSSParser(TopicMap tm, RSSExtractor parent)

Method Summary

Modifier and Type	Method and Description
void	characters(char[] ch, int start, int length)
void	endDocument()
void	endElement(java.lang.String uri, java.lang.String localName, java.lang.String qName)
void	endPrefixMapping(java.lang.String prefix)
void	error(org.xml.sax.SAXParseException exception)
void	fatalError(org.xml.sax.SAXParseException exception)
private Topic	getOrCreateTopic(java.lang.String si)
private Topic	getOrCreateTopic(java.lang.String si, java.lang.String bn)
void	ignorableWhitespace(char[] ch, int start, int length)
void	processingInstruction(java.lang.String target, java.lang.String data)
void	setDocumentLocator(org.xml.sax.Locator locator)
void	skippedEntity(java.lang.String name)
void	startDocument()
void	startElement(java.lang.String uri, java.lang.String localName, java.lang.String qName, org.xml.sax.Attributes atts)
void	startPrefixMapping(java.lang.String prefix, java.lang.String uri)
void	warning(org.xml.sax.SAXParseException exception)

Methods inherited from class java.lang.Object

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

Field Detail

MAKE_LINK_OCCURRENCE

public static boolean MAKE_LINK_OCCURRENCE

MAKE_LINK_SUBJECT_IDENTIFIER

public static boolean MAKE_LINK_SUBJECT_IDENTIFIER

MAKE_LINK_SUBJECT_LOCATOR

public static boolean MAKE_LINK_SUBJECT_LOCATOR

MAKE_SUBCLASS_OF_WANDORA_CLASS

public static boolean MAKE_SUBCLASS_OF_WANDORA_CLASS

progress

public int progress

itemCount

public int itemCount

channelCount

public int channelCount

tm

private TopicMap tm

parent

private RSSExtractor parent

TAG_RSS

public static final java.lang.String TAG_RSS

See Also:

Constant Field Values

TAG_CHANNEL

public static final java.lang.String TAG_CHANNEL

See Also:

Constant Field Values

TAG_TITLE

public static final java.lang.String TAG_TITLE

See Also:

Constant Field Values

TAG_LINK

public static final java.lang.String TAG_LINK

See Also:

Constant Field Values

TAG_DESCRIPTION

public static final java.lang.String TAG_DESCRIPTION

See Also:

Constant Field Values

TAG_LANGUAGE

public static final java.lang.String TAG_LANGUAGE

See Also:

Constant Field Values

TAG_PUBDATE

public static final java.lang.String TAG_PUBDATE

See Also:

Constant Field Values

TAG_LASTBUILDDATE

public static final java.lang.String TAG_LASTBUILDDATE

See Also:

Constant Field Values

TAG_DOCS

public static final java.lang.String TAG_DOCS

See Also:

Constant Field Values

TAG_GENERATOR

public static final java.lang.String TAG_GENERATOR

See Also:

Constant Field Values

TAG_MANAGINGEDITOR

public static final java.lang.String TAG_MANAGINGEDITOR

See Also:

Constant Field Values

TAG_WEBMASTER

public static final java.lang.String TAG_WEBMASTER

See Also:

Constant Field Values

TAG_TTL

public static final java.lang.String TAG_TTL

See Also:

Constant Field Values

TAG_TEXTINPUT

public static final java.lang.String TAG_TEXTINPUT

See Also:

Constant Field Values

TAG_COPYRIGHT

public static final java.lang.String TAG_COPYRIGHT

See Also:

Constant Field Values

TAG_RATING

public static final java.lang.String TAG_RATING

See Also:

Constant Field Values

TAG_ITEM

public static final java.lang.String TAG_ITEM

See Also:

Constant Field Values

TAG_GUID

public static final java.lang.String TAG_GUID

See Also:

Constant Field Values

TAG_CATEGORY

public static final java.lang.String TAG_CATEGORY

See Also:

Constant Field Values

TAG_AUTHOR

public static final java.lang.String TAG_AUTHOR

See Also:

Constant Field Values

TAG_IMAGE

public static final java.lang.String TAG_IMAGE

See Also:

Constant Field Values

TAG_URL

public static final java.lang.String TAG_URL

See Also:

Constant Field Values

STATE_START

private static final int STATE_START

See Also:

Constant Field Values

STATE_RSS

private static final int STATE_RSS

See Also:

Constant Field Values

STATE_CHANNEL

private static final int STATE_CHANNEL

See Also:

Constant Field Values

STATE_CHANNEL_TITLE

private static final int STATE_CHANNEL_TITLE

See Also:

Constant Field Values

STATE_CHANNEL_LINK

private static final int STATE_CHANNEL_LINK

See Also:

Constant Field Values

STATE_CHANNEL_DESCRIPTION

private static final int STATE_CHANNEL_DESCRIPTION

See Also:

Constant Field Values

STATE_CHANNEL_LANGUAGE

private static final int STATE_CHANNEL_LANGUAGE

See Also:

Constant Field Values

STATE_CHANNEL_PUBDATE

private static final int STATE_CHANNEL_PUBDATE

See Also:

Constant Field Values

STATE_CHANNEL_LASTBUILDDATE

private static final int STATE_CHANNEL_LASTBUILDDATE

See Also:

Constant Field Values

STATE_CHANNEL_DOCS

private static final int STATE_CHANNEL_DOCS

See Also:

Constant Field Values

STATE_CHANNEL_GENERATOR

private static final int STATE_CHANNEL_GENERATOR

See Also:

Constant Field Values

STATE_CHANNEL_MANAGINGEDITOR

private static final int STATE_CHANNEL_MANAGINGEDITOR

See Also:

Constant Field Values

STATE_CHANNEL_WEBMASTER

private static final int STATE_CHANNEL_WEBMASTER

See Also:

Constant Field Values

STATE_CHANNEL_TTL

private static final int STATE_CHANNEL_TTL

See Also:

Constant Field Values

STATE_CHANNEL_TEXTINPUT

private static final int STATE_CHANNEL_TEXTINPUT

See Also:

Constant Field Values

STATE_CHANNEL_COPYRIGHT

private static final int STATE_CHANNEL_COPYRIGHT

See Also:

Constant Field Values

STATE_CHANNEL_CATEGORY

private static final int STATE_CHANNEL_CATEGORY

See Also:

Constant Field Values

STATE_CHANNEL_RATING

private static final int STATE_CHANNEL_RATING

See Also:

Constant Field Values

STATE_CHANNEL_IMAGE

private static final int STATE_CHANNEL_IMAGE

See Also:

Constant Field Values

STATE_CHANNEL_IMAGE_URL

private static final int STATE_CHANNEL_IMAGE_URL

See Also:

Constant Field Values

STATE_CHANNEL_IMAGE_TITLE

private static final int STATE_CHANNEL_IMAGE_TITLE

See Also:

Constant Field Values

STATE_CHANNEL_IMAGE_LINK

private static final int STATE_CHANNEL_IMAGE_LINK

See Also:

Constant Field Values

STATE_CHANNEL_IMAGE_DESCRIPTION

private static final int STATE_CHANNEL_IMAGE_DESCRIPTION

See Also:

Constant Field Values

STATE_CHANNEL_ITEM

private static final int STATE_CHANNEL_ITEM

See Also:

Constant Field Values

STATE_CHANNEL_ITEM_TITLE

private static final int STATE_CHANNEL_ITEM_TITLE

See Also:

Constant Field Values

STATE_CHANNEL_ITEM_LINK

private static final int STATE_CHANNEL_ITEM_LINK

See Also:

Constant Field Values

STATE_CHANNEL_ITEM_DESCRIPTION

private static final int STATE_CHANNEL_ITEM_DESCRIPTION

See Also:

Constant Field Values

STATE_CHANNEL_ITEM_PUBDATE

private static final int STATE_CHANNEL_ITEM_PUBDATE

See Also:

Constant Field Values

STATE_CHANNEL_ITEM_GUID

private static final int STATE_CHANNEL_ITEM_GUID

See Also:

Constant Field Values

STATE_CHANNEL_ITEM_CATEGORY

private static final int STATE_CHANNEL_ITEM_CATEGORY

See Also:

Constant Field Values

STATE_CHANNEL_ITEM_AUTHOR

private static final int STATE_CHANNEL_ITEM_AUTHOR

See Also:

Constant Field Values

state

private int state

RSS_SI

public static java.lang.String RSS_SI

SIPREFIX

public static java.lang.String SIPREFIX

CHANNEL_SI

public static java.lang.String CHANNEL_SI

CHANNEL_LINK_SI

public static java.lang.String CHANNEL_LINK_SI

CHANNEL_DESCRIPTION_SI

public static java.lang.String CHANNEL_DESCRIPTION_SI

CHANNEL_LANGUAGE_SI

public static java.lang.String CHANNEL_LANGUAGE_SI

CHANNEL_PUBDATE_SI

public static java.lang.String CHANNEL_PUBDATE_SI

CHANNEL_LASTBUILDDATE_SI

public static java.lang.String CHANNEL_LASTBUILDDATE_SI

CHANNEL_DOCS_SI

public static java.lang.String CHANNEL_DOCS_SI

CHANNEL_GENERATOR_SI

public static java.lang.String CHANNEL_GENERATOR_SI

CHANNEL_MANAGINGEDITOR_SI

public static java.lang.String CHANNEL_MANAGINGEDITOR_SI

CHANNEL_WEBMASTER_SI

public static java.lang.String CHANNEL_WEBMASTER_SI

CHANNEL_TTL_SI

public static java.lang.String CHANNEL_TTL_SI

CHANNEL_TEXTINPUT_SI

public static java.lang.String CHANNEL_TEXTINPUT_SI

CHANNEL_COPYRIGHT_SI

public static java.lang.String CHANNEL_COPYRIGHT_SI

CHANNEL_CATEGORY_SI

public static java.lang.String CHANNEL_CATEGORY_SI

CHANNEL_RATING_SI

public static java.lang.String CHANNEL_RATING_SI

CHANNEL_ITEM_SI

public static java.lang.String CHANNEL_ITEM_SI

CHANNEL_ITEM_TITLE_SI

public static java.lang.String CHANNEL_ITEM_TITLE_SI

CHANNEL_ITEM_LINK_SI

public static java.lang.String CHANNEL_ITEM_LINK_SI

CHANNEL_ITEM_DESCRIPTION_SI

public static java.lang.String CHANNEL_ITEM_DESCRIPTION_SI

CHANNEL_ITEM_PUBDATE_SI

public static java.lang.String CHANNEL_ITEM_PUBDATE_SI

CHANNEL_ITEM_GUID_SI

public static java.lang.String CHANNEL_ITEM_GUID_SI

CHANNEL_ITEM_CATEGORY_SI

public static java.lang.String CHANNEL_ITEM_CATEGORY_SI

CHANNEL_ITEM_AUTHOR_SI

public static java.lang.String CHANNEL_ITEM_AUTHOR_SI

CHANNEL_IMAGE_SI

public static java.lang.String CHANNEL_IMAGE_SI

CHANNEL_IMAGE_LINK_SI

public static java.lang.String CHANNEL_IMAGE_LINK_SI

CHANNEL_IMAGE_DESCRIPTION_SI

public static java.lang.String CHANNEL_IMAGE_DESCRIPTION_SI

DATE_SI

public static java.lang.String DATE_SI

data_channel_title

private java.lang.String data_channel_title

data_channel_link

private java.lang.String data_channel_link

data_channel_description

private java.lang.String data_channel_description

data_channel_language

private java.lang.String data_channel_language

data_channel_pubdate

private java.lang.String data_channel_pubdate

data_channel_lastbuilddate

private java.lang.String data_channel_lastbuilddate

data_channel_docs

private java.lang.String data_channel_docs

data_channel_generator

private java.lang.String data_channel_generator

data_channel_managingeditor

private java.lang.String data_channel_managingeditor

data_channel_webmaster

private java.lang.String data_channel_webmaster

data_channel_ttl

private java.lang.String data_channel_ttl

data_channel_copyright

private java.lang.String data_channel_copyright

data_channel_category

private java.lang.String data_channel_category

data_channel_rating

private java.lang.String data_channel_rating

data_channel_item_title

private java.lang.String data_channel_item_title

data_channel_item_link

private java.lang.String data_channel_item_link

data_channel_item_description

private java.lang.String data_channel_item_description

data_channel_item_pubdate

private java.lang.String data_channel_item_pubdate

data_channel_item_guid

private java.lang.String data_channel_item_guid

data_channel_item_category

private java.lang.String data_channel_item_category

data_channel_item_author

private java.lang.String data_channel_item_author

data_channel_image_url

private java.lang.String data_channel_image_url

data_channel_image_title

private java.lang.String data_channel_image_title

data_channel_image_link

private java.lang.String data_channel_image_link

data_channel_image_description

private java.lang.String data_channel_image_description

theChannel

private Topic theChannel

theItem

private Topic theItem

Constructor Detail

RSSParser

public RSSParser(TopicMap tm,
                 RSSExtractor parent)

Method Detail

getOrCreateTopic

private Topic getOrCreateTopic(java.lang.String si)
                        throws TopicMapException

Throws:

TopicMapException

getOrCreateTopic

private Topic getOrCreateTopic(java.lang.String si,
                               java.lang.String bn)
                        throws TopicMapException

Throws:

TopicMapException

startDocument

public void startDocument()
                   throws org.xml.sax.SAXException

Specified by:

startDocument in interface org.xml.sax.ContentHandler

Throws:

org.xml.sax.SAXException

endDocument

public void endDocument()
                 throws org.xml.sax.SAXException

Specified by:

endDocument in interface org.xml.sax.ContentHandler

Throws:

org.xml.sax.SAXException

startElement

public void startElement(java.lang.String uri,
                         java.lang.String localName,
                         java.lang.String qName,
                         org.xml.sax.Attributes atts)
                  throws org.xml.sax.SAXException

Specified by:

startElement in interface org.xml.sax.ContentHandler

Throws:

org.xml.sax.SAXException

endElement

public void endElement(java.lang.String uri,
                       java.lang.String localName,
                       java.lang.String qName)
                throws org.xml.sax.SAXException

Specified by:

endElement in interface org.xml.sax.ContentHandler

Throws:

org.xml.sax.SAXException

characters

public void characters(char[] ch,
                       int start,
                       int length)
                throws org.xml.sax.SAXException

Specified by:

characters in interface org.xml.sax.ContentHandler

Throws:

org.xml.sax.SAXException

warning

public void warning(org.xml.sax.SAXParseException exception)
             throws org.xml.sax.SAXException

Specified by:

warning in interface org.xml.sax.ErrorHandler

Throws:

org.xml.sax.SAXException

error

public void error(org.xml.sax.SAXParseException exception)
           throws org.xml.sax.SAXException

Specified by:

error in interface org.xml.sax.ErrorHandler

Throws:

org.xml.sax.SAXException

fatalError

public void fatalError(org.xml.sax.SAXParseException exception)
                throws org.xml.sax.SAXException

Specified by:

fatalError in interface org.xml.sax.ErrorHandler

Throws:

org.xml.sax.SAXException

ignorableWhitespace

public void ignorableWhitespace(char[] ch,
                                int start,
                                int length)
                         throws org.xml.sax.SAXException

Specified by:

ignorableWhitespace in interface org.xml.sax.ContentHandler

Throws:

org.xml.sax.SAXException

processingInstruction

public void processingInstruction(java.lang.String target,
                                  java.lang.String data)
                           throws org.xml.sax.SAXException

Specified by:

processingInstruction in interface org.xml.sax.ContentHandler

Throws:

org.xml.sax.SAXException

startPrefixMapping

public void startPrefixMapping(java.lang.String prefix,
                               java.lang.String uri)
                        throws org.xml.sax.SAXException

Specified by:

startPrefixMapping in interface org.xml.sax.ContentHandler

Throws:

org.xml.sax.SAXException

endPrefixMapping

public void endPrefixMapping(java.lang.String prefix)
                      throws org.xml.sax.SAXException

Specified by:

endPrefixMapping in interface org.xml.sax.ContentHandler

Throws:

org.xml.sax.SAXException

setDocumentLocator

public void setDocumentLocator(org.xml.sax.Locator locator)

Specified by:

setDocumentLocator in interface org.xml.sax.ContentHandler

skippedEntity

public void skippedEntity(java.lang.String name)
                   throws org.xml.sax.SAXException

Specified by:

skippedEntity in interface org.xml.sax.ContentHandler

Throws:

org.xml.sax.SAXException