gov.lanl.archive.rewrite
Class TextDocument

java.lang.Object
  gov.lanl.archive.rewrite.TextDocument

public class TextDocument
extends Object

Class which wraps functionality for converting a Resource(InputStream + HTTP headers) into a StringBuilder, performing several common URL resolution methods against that StringBuilder, inserting arbitrary Strings into the page, and then converting the page back to a byte array.

Version:

$Date: 2010-09-28 16:28:38 -0600 (Tue, 28 Sep 2010) $, $Revision: 3262 $

Author:

luda inspired by brad

Field Summary

StringBuilder

sb the internal StringBuilder

Constructor Summary

TextDocument(Memento resource, String pageUrl, String captureDate, ResultURIConverter uriConverter)

Method Summary

void	addBase()
byte[]	getBytes()
String	getCharSet()
String	getJSIncludeString(String jsUrl)
String	getResult()
void	insertAtEndOfBody(String toInsert)
void	insertAtStartOfBody(String toInsert)
void	insertAtStartOfDocument(String toInsert)
void	insertAtStartOfHead(String toInsert)
void	readFully(InputStreamReader isr)
void	resolveAllPageUrls() Update all URLs inside the page, so they resolve correctly to absolute URLs within the Wayback service.
void	resolveASXRefUrls()
void	resolveCSSUrls()
void	resolvePageUrls() Update URLs inside the page, so those URLs which must be correct at page load time resolve correctly to absolute URLs.
void	setCharSet(String charSet)
void	setResultBytes(byte[] resultBytes)
void	stripHTML()
void	writeToOutputStream(OutputStream os) Write the contents of the page to the client.

Methods inherited from class java.lang.Object

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

Field Detail

sb

public StringBuilder sb

the internal StringBuilder

Constructor Detail

TextDocument

public TextDocument(Memento resource,
                    String pageUrl,
                    String captureDate,
                    ResultURIConverter uriConverter)

Parameters:

resource -

result -

uriConverter -

Method Detail

addBase

public void addBase()

resolvePageUrls

public void resolvePageUrls()

Update URLs inside the page, so those URLs which must be correct at page load time resolve correctly to absolute URLs. This means ensuring there is a BASE HREF tag, adding one if missing, and then resolving: FRAME-SRC, META-URL, LINK-HREF, SCRIPT-SRC tag-attribute pairs against either the existing BASE-HREF, or the page's absolute URL if it was missing.

resolveAllPageUrls

public void resolveAllPageUrls()

Update all URLs inside the page, so they resolve correctly to absolute URLs within the Wayback service.

resolveCSSUrls

public void resolveCSSUrls()

resolveASXRefUrls

public void resolveASXRefUrls()

stripHTML

public void stripHTML()

readFully

public void readFully(InputStreamReader isr)
               throws IOException

Parameters:

charSet -

Throws:

IOException

getBytes

public byte[] getBytes()
                throws UnsupportedEncodingException

Returns:

raw bytes contained in internal StringBuilder

Throws:

UnsupportedEncodingException

getResult

public String getResult()
                 throws UnsupportedEncodingException

Throws:

UnsupportedEncodingException

setResultBytes

public void setResultBytes(byte[] resultBytes)

writeToOutputStream

public void writeToOutputStream(OutputStream os)
                         throws IOException

Write the contents of the page to the client.

Parameters:

os -

Throws:

IOException

insertAtStartOfDocument

public void insertAtStartOfDocument(String toInsert)

Parameters:

toInsert -

insertAtStartOfHead

public void insertAtStartOfHead(String toInsert)

Parameters:

toInsert -

insertAtEndOfBody

public void insertAtEndOfBody(String toInsert)

Parameters:

toInsert -

insertAtStartOfBody

public void insertAtStartOfBody(String toInsert)

Parameters:

toInsert -

getJSIncludeString

public String getJSIncludeString(String jsUrl)

Parameters:

jsUrl -

Returns:

getCharSet

public String getCharSet()

Returns:

the charSet

setCharSet

public void setCharSet(String charSet)

Parameters:

charSet - the charSet to set