sunlabs.brazil.handler

Class HtmlRewriter

public class HtmlRewriter extends Object

This class helps with parsing and rewriting an HTML document. The source document is not changed; a new HTML document is built.

The user can sequentially examine and rewrite each token in the source HTML document. As each token in the document is seen, the user has two choices:

If the user modifies (or replaces, deletes, etc.) the current token, then the resultant HTML document will contain that modification. On the other hand, if the user doesn't do anything with the current token, it will appear, unchanged, in the resultant HTML document.

Parsing is implemented lazily, meaning, for example, that unless the user actually asks for attributes of an HTML tag, this parser does not have to spend the time breaking up the attributes.

This class is used by HTML filters to maintain the state of the document and allow the filters to perform arbitrary rewriting.

Version: @(#)HtmlRewriter.java 2.6

Author: Colin Stevens (colin.stevens@sun.com)

Field Summary
LexHTMLlex
The parser for the source HTML document.
StringBuffersb
Storage holding the resultant HTML document.
Constructor Summary
HtmlRewriter(LexHTML lex)
Creates a new HtmlRewriter from the given HTML parser.
HtmlRewriter(String str)
Creates a new HtmlRewriter that will operate on the given string.
Method Summary
booleanaccumulate(boolean accumulate)
Turns on or off the automatic accumulation of each token.
voidappend(String str)
Instead of modifying an existing token, this method allows the user to completely replace the current token with arbitrary new content.
voidappendToken()
Appends the current token to the resultant HTML document.
Stringget(String key)
Returns the value that the specified case-insensitive key maps to in the attributes for the current tag.
StringgetArgs()
Gets the arguments of the current token as a string.
StringgetBody()
Gets the body of the current token as a string.
StringMapgetMap()
Return a copy of the StringMap of attributes.
StringgetTag()
Gets the current tag's name.
StringgetToken()
Gets the raw string making up the entire current token, including the angle brackets or comment delimiters, if applicable.
intgetType()
Gets the type of the current token.
booleanisSingleton()
See if the current tag a singleton.
Enumerationkeys()
Returns an enumeration of the keys in the current tag's attributes.
voidkillToken()
Tells this HtmlRewriter not to append the current token to the resultant HTML document.
booleannextTag()
A convenence method built on top of nextToken.
booleannextToken()
Advances to the next token in the source HTML document.
voidpushback()
Puts the current token back.
voidput(String key, String value)
Maps the given case-insensitive key to the specified value in the current tag's attributes.
static Stringquote(String str)
Helper class to quote a attribute's value when the value is being written to the resultant HTML document.
voidremove(String key)
Removes the given case-insensitive key and its corresponding value from the current tag's attributes.
voidreset()
Forgets all the tokens that have been appended to the resultant HTML document so far, including the current token.
voidsetSingleton(boolean singleton)
Make the current tag a singleton.
voidsetTag(String tag)
Changes the current tag's name.
voidsetType(int type)
Sets the type of the current token.
inttagCount()
Return count of tags seen so far
inttokenCount()
Return count of tokens seen so far
StringtoString()
Returns the "new" rewritten HTML document.

Field Detail

lex

public LexHTML lex
The parser for the source HTML document.

sb

public StringBuffer sb
Storage holding the resultant HTML document.

Constructor Detail

HtmlRewriter

public HtmlRewriter(LexHTML lex)
Creates a new HtmlRewriter from the given HTML parser.

Parameters: lex The HTML parser.

HtmlRewriter

public HtmlRewriter(String str)
Creates a new HtmlRewriter that will operate on the given string.

Parameters: str The HTML document.

Method Detail

accumulate

public boolean accumulate(boolean accumulate)
Turns on or off the automatic accumulation of each token.

After each token is processed, the current token is appended to to the resultant HTML document unless the user has already appended something else. By setting accumulate to false, this behavior is turned off. The user must then explicitly call appendToken to cause the current token to be appended.

Turning off accumulation takes effect immediately, while turning on accumulation takes effect on the next token. In other words, whether the user turns this setting off or on, the current token will not be added to the resultant HTML document unless the user explicitly calls appendToken.

Following is sample code that illustrates how to use this method to extract the contents of the <head> of the source HTML document.

 HtmlRewriter hr = new HtmlRewriter(str);
 // Don't accumulate tokens until we see the <head> below.
 hr.accumulate(false);
 while (hr.nextTag()) {
     if (hr.getTag().equals("head")) {
         // Start remembering the contents of the HTML document,
         // not including the <head> tag itself.

         hr.accumulate(true);
     } else if (hr.getTag().equals("/head")) {
         // Return everything accumulated so far.

         return hr.toString();
     }
 }
 
This method can be called any number of times while processing the source HTML document.

Parameters: accumulate true to automatically accumulate tokens in the resultant HTML document, false to require that the user explicitly accumulate them.

Returns: The previous accumulate setting

See Also: HtmlRewriter

append

public void append(String str)
Instead of modifying an existing token, this method allows the user to completely replace the current token with arbitrary new content.

This method may be called multiple times while processing the current token to add more and more data to the resultant HTML document. Before and/or after calling this method, the appendToken method may also be called explicitly in order to add the current token to the resultant HTML document.

Following is sample code illustrating how to use this method to put bold tags around all the <a> tags.

 HtmlRewriter hr = new HtmlRewriter(str);
 while (hr.nextTag()) {
     if (hr.getTag().equals("a")) {
         hr.append("<b>");
         hr.appendToken();
     } else if (hr.getTag().equals("/a")) {
         hr.appendToken();
         hr.append("</b>");
     }
 }
 
The calls to appendToken are necessary. Otherwise, the HtmlRewriter could not know where and when to append the existing token in addition to the new content provided by the user.

Parameters: str The new content to append. May be null, in which case no new content is appended (the equivalent of appending "").

See Also: appendToken HtmlRewriter

appendToken

public void appendToken()
Appends the current token to the resultant HTML document. If the caller has changed the current token using the setTag, set, or remove methods, those changes will be reflected.

By default, this method is automatically called after each token is processed unless the user has already appended something to the resultant HTML document. Therefore, if the user appends something and also wants to append the current token, or if the user wants to append the current token a number of times, this method must be called.

See Also: HtmlRewriter HtmlRewriter

get

public String get(String key)
Returns the value that the specified case-insensitive key maps to in the attributes for the current tag. For keys that were present in the tag's attributes without a value, the value returned is the empty string. In other words, for the tag <table border rows=2>:

Surrounding single and double quote marks that occur in the literal tag are removed from the values reported. So, for the tag <a href="/foo.html" target=_top onclick='alert("hello")'>:

Parameters: The key to lookup in the current tag's attributes.

Returns: The value to which the specified key is mapped, or null if the key was not in the attributes.

See Also: LexHTML

getArgs

public String getArgs()
Gets the arguments of the current token as a string.

Returns: The body.

See Also: LexHTML

getBody

public String getBody()
Gets the body of the current token as a string.

Returns: The body.

See Also: LexHTML

getMap

public StringMap getMap()
Return a copy of the StringMap of attributes.

getTag

public String getTag()
Gets the current tag's name. The name returned is converted to lower case.

Returns: The lower-cased tag name, or null if the current token does not have a tag name

See Also: LexHTML

getToken

public String getToken()
Gets the raw string making up the entire current token, including the angle brackets or comment delimiters, if applicable.

Returns: The current token.

See Also: LexHTML

getType

public int getType()
Gets the type of the current token.

Returns: The type.

See Also: LexHTML

isSingleton

public boolean isSingleton()
See if the current tag a singleton. A Singleton tag ends in "/", as in <
.

keys

public Enumeration keys()
Returns an enumeration of the keys in the current tag's attributes. The elements of the enumeration are the string keys. The keys can be passed to get to get the values of the attributes.

Returns: An enumeration of the keys.

killToken

public void killToken()
Tells this HtmlRewriter not to append the current token to the resultant HTML document. Even if the user hasn't appended anything else, the current token will be ignored rather than appended.

See Also: HtmlRewriter HtmlRewriter

nextTag

public boolean nextTag()
A convenence method built on top of nextToken. Advances to the next HTML tag. All intervening strings and comments between the last tag and the new current tag are copied through unchanged. This method can be used when the caller wants to process only HTML tags, without having to manually check the type of each token to see if it is actually a tag.

Returns: true if there are tokens left to process, false otherwise.

nextToken

public boolean nextToken()
Advances to the next token in the source HTML document.

The other purpose of this function is to "do the right thing", which is to append the token we just processed to the resultant HTML document, unless the user has already appended something else.

A sample program follows. This program changes all <img> tags to <form> tags, deletes all <table> tags, capitalizes and bolds each string token, and passes all other tokens through unchanged, to illustrate how nextToken interacts with some of the other methods in this class.

 HtmlRewriter hr = new HtmlRewriter(str);
 while (hr.nextToken()) {
     switch (hr.getType()) {
     case LexHTML.TAG: 
         if (hr.getTag().equals("img")) {
             // Change the tag name w/o affecting the attributes.
             
             hr.setTag("form");
         } else if (hr.getTag().equals("table")) {
             // Eliminate the entire "table" token.
             
             hr.killToken();
         } 
         break;
             
     case LexHTML.STRING:
         // Append a new sequence in place of the existing token.

         hr.append("<b>" + hr.getToken().toUpperCase() + "</b>");
         break;
     }
     // Any tokens we didn't modify get copied through unchanged.
 }
 

Returns: true if there are tokens left to process, false otherwise.

pushback

public void pushback()
Puts the current token back. The next time nextToken is called, it will be the current token again, rather than advancing to the next token in the source HTML document.

This is useful when a code fragment needs to read an indefinite number of tokens, but that once some distinguished token is found, needs to push that token back so that normal processing can occur on that token.

put

public void put(String key, String value)
Maps the given case-insensitive key to the specified value in the current tag's attributes.

The value can be retrieved by calling get with a key that is case-insensitive equal to the given key.

If the attributes already contained a mapping for the given key, the old value is forgotten and the new specified value is used. The case of the prior key is retained in that case. Otherwise the case of the new key is used and a new mapping is made.

Parameters: key The new key. May not be null. value The new value. May be not be null.

quote

public static String quote(String str)
Helper class to quote a attribute's value when the value is being written to the resultant HTML document. Values set by the put method are automatically quoted as needed. This method is provided in case the user is dynamically constructing a new tag to be appended with append and needs to quote some arbitrary values.

The quoting algorithm is as follows:
If the string contains double-quotes, put single quotes around it.
If the string contains any "special" characters, put double-quotes around it.

This algorithm is, of course, insufficient for complicated strings that include both single and double quotes. In that case, it is the user's responsibility to escape the special characters in the string using the HTML special symbols like &quot; or &#34;

Returns: The quoted string, or the original string if it did not need to be quoted.

remove

public void remove(String key)
Removes the given case-insensitive key and its corresponding value from the current tag's attributes. This method does nothing if the key is not in the attributes.

Parameters: key The key that needs to be removed. Must not be null.

reset

public void reset()
Forgets all the tokens that have been appended to the resultant HTML document so far, including the current token.

setSingleton

public void setSingleton(boolean singleton)
Make the current tag a singleton. A Singleton tag ends in "/", as in <
.

setTag

public void setTag(String tag)
Changes the current tag's name. The tag's attributes are not changed.

Parameters: tag New tag name

setType

public void setType(int type)
Sets the type of the current token.

tagCount

public int tagCount()
Return count of tags seen so far

tokenCount

public int tokenCount()
Return count of tokens seen so far

toString

public String toString()
Returns the "new" rewritten HTML document. This is normally called once all of the tokens have been processed, and the user wants to send on this rewritten document.

At any time, this method can be called to return the current state of the HTML document. The return value is the result of processing the source document up to this point in time; the unprocessed remainder of the source document is not considered.

Due to the implementation, calling this method may be expensive. Specifically, calling this method a second (or further) time for a given HtmlRewriter may involve copying temporary strings around. The pessimal case would be to call this method every time a new token is appended.

Returns: The rewritten HTML document, up to this point in time.