org.apache.oro.text.regex.PatternCompiler Java Examples

/** 
 * 字符串下标 regx在src中首次出现的位置 
 * @param src     src
 * @param regx    regx
 * @param begin   有效开始位置 
 * @return return
 */ 
public static int indexOf(String src, String regx, int begin){ 
	int idx = -1; 
	try{ 
		PatternCompiler patternCompiler = new Perl5Compiler(); 
		Pattern pattern = patternCompiler.compile(regx, Perl5Compiler.CASE_INSENSITIVE_MASK); 
		PatternMatcher matcher = new Perl5Matcher(); 
		PatternMatcherInput input = new PatternMatcherInput(src); 
		 
		while(matcher.contains(input, pattern)){ 
			MatchResult matchResult = matcher.getMatch(); 
			int tmp = matchResult.beginOffset(0); 
			if(tmp >= begin){//匹配位置从begin开始 
				idx = tmp; 
				break; 
			} 
		} 
	}catch(Exception e){
		log.error("[提取异常][src:{}][regx:{}]", src, regx); 
		e.printStackTrace();
	} 
	return idx; 
}
 

protected String doRereplace( String _theString, String _theRE, String _theSubstr, boolean _casesensitive, boolean _replaceAll ) throws cfmRunTimeException{
	int replaceCount = _replaceAll ? Util.SUBSTITUTE_ALL : 1; 
	PatternMatcher matcher = new Perl5Matcher();
	Pattern pattern = null;
	PatternCompiler compiler = new Perl5Compiler();
   
	try {
		if ( _casesensitive ){
			pattern = compiler.compile( _theRE, Perl5Compiler.SINGLELINE_MASK );
		}else{
			pattern = compiler.compile( _theRE, Perl5Compiler.CASE_INSENSITIVE_MASK | Perl5Compiler.SINGLELINE_MASK );
		}
		
	} catch(MalformedPatternException e){ // definitely should happen since regexp is hardcoded
		cfCatchData catchD = new cfCatchData();
		catchD.setType( "Function" );
		catchD.setMessage( "Internal Error" );
		catchD.setDetail( "Invalid regular expression ( " + _theRE + " )" );
		throw new cfmRunTimeException( catchD );
	}

	// Perform substitution and print result.
	return Util.substitute(matcher, pattern, new Perl5Substitution( processSubstr( _theSubstr ) ), _theString, replaceCount );
}
 

@Override
public Pattern load(String pattern) throws Exception {
    try {
        PatternCompiler pc = new Perl5Compiler();
        return pc.compile(
                pattern,
                Perl5Compiler.CASE_INSENSITIVE_MASK
                        | Perl5Compiler.READ_ONLY_MASK
                        | Perl5Compiler.SINGLELINE_MASK);
    } catch (MalformedPatternException e) {
        throw new NullPointerException("PatternUtils error!");
    }
}
 

public Pattern apply(String pattern) {
    try {
        PatternCompiler pc = new Perl5Compiler();
        return pc.compile(pattern,
            Perl5Compiler.CASE_INSENSITIVE_MASK
                    | Perl5Compiler.READ_ONLY_MASK
                    | Perl5Compiler.SINGLELINE_MASK);
    } catch (MalformedPatternException e) {
        throw new RuntimeException(e);
    }
}
 

public Pattern apply(String pattern) {
    try {
        PatternCompiler pc = new Perl5Compiler();
        return pc.compile(pattern, Perl5Compiler.CASE_INSENSITIVE_MASK | Perl5Compiler.READ_ONLY_MASK);
    } catch (MalformedPatternException e) {
        throw new RuntimeException("Regex failed!", e);
    }
}
 

@Override
public Pattern apply(String pattern) {
    try {
        PatternCompiler pc = new Perl5Compiler();
        return pc.compile(pattern,
                Perl5Compiler.CASE_INSENSITIVE_MASK
                        | Perl5Compiler.READ_ONLY_MASK
                        | Perl5Compiler.SINGLELINE_MASK);
    } catch (MalformedPatternException e) {
        throw new CanalFilterException(e);
    }
}
 

public static String escapeHtml(String str) {
	try {
		PatternMatcher matcher = new Perl5Matcher();
		PatternCompiler compiler = new Perl5Compiler();

		Pattern pattern = compiler.compile("&(([a-z][a-zA-Z0-9]*)|(#\\d{2,6});)", Perl5Compiler.CASE_INSENSITIVE_MASK | Perl5Compiler.SINGLELINE_MASK);

		String tmp = Util.substitute(matcher, pattern, new Perl5Substitution("&$1"), str, Util.SUBSTITUTE_ALL);

		return replaceChars(tmp, new char[] { '<', '>', '\"' }, new String[] { "&lt;", "&gt;", "&quot;" });
	} catch (Exception e) {
		return str;
	}// won't happen

}
 

public static boolean regexMatches(String str, String re) throws MalformedPatternException {
	PatternMatcher matcher = new Perl5Matcher();
	PatternCompiler compiler = new Perl5Compiler();
	PatternMatcherInput input = new PatternMatcherInput(str);

	Pattern pattern = compiler.compile(re, Perl5Compiler.SINGLELINE_MASK);
	return matcher.matches(input, pattern);
}
 

/**
 * compiles fileNameRegExp into a regular expression and pattern matches on
 * each file's name, and returns those that match.
 * 
 * @param files
 * @param fileNameRegExp
 * 
 * @return String[] of file names that match the expresion.
 */
public String[] matchFileNamesWithPattern(File[] files,
		String fileNameRegExp) throws SshException {
	// set up variables for regexp matching
	Pattern mpattern = null;
	PatternCompiler aPCompiler = new Perl5Compiler();
	PatternMatcher aPerl5Matcher = new Perl5Matcher();
	// Attempt to compile the pattern. If the pattern is not valid,
	// throw exception
	try {
		mpattern = aPCompiler.compile(fileNameRegExp);
	} catch (MalformedPatternException e) {
		throw new SshException("Invalid regular expression:"
				+ e.getMessage(), SshException.BAD_API_USAGE);
	}

	Vector<String> matchedNames = new Vector<String>();

	for (int i = 0; i < files.length; i++) {
		if ((!files[i].getName().equals("."))
				&& (!files[i].getName().equals(".."))
				&& (!files[i].isDirectory())) {
			if (aPerl5Matcher.matches(files[i].getName(), mpattern)) {
				// call get for each match, passing true, so that it doesnt
				// repeat the search
				matchedNames.addElement(files[i].getName());
			}
		}
	}

	// return (String[]) matchedNames.toArray(new String[0]);
	String[] matchedNamesStrings = new String[matchedNames.size()];
	matchedNames.copyInto(matchedNamesStrings);
	return matchedNamesStrings;
}
 

/**
 * compiles fileNameRegExp into a regular expression and pattern matches on
 * each file's name, and returns those that match.
 * 
 * @param files
 * @param fileNameRegExp
 * 
 * @return SftpFile[] of files that match the expresion.
 */
public SftpFile[] matchFilesWithPattern(SftpFile[] files,
		String fileNameRegExp) throws SftpStatusException, SshException {
	// set up variables for regexp matching
	Pattern mpattern = null;
	PatternCompiler aPCompiler = new Perl5Compiler();
	PatternMatcher aPerl5Matcher = new Perl5Matcher();
	// Attempt to compile the pattern. If the pattern is not valid,
	// throw exception
	try {
		mpattern = aPCompiler.compile(fileNameRegExp);
	} catch (MalformedPatternException e) {
		throw new SshException("Invalid regular expression:"
				+ e.getMessage(), SshException.BAD_API_USAGE);
	}

	Vector<SftpFile> matchedNames = new Vector<SftpFile>();

	for (int i = 0; i < files.length; i++) {
		if ((!files[i].getFilename().equals("."))
				&& (!files[i].getFilename().equals(".."))
				&& (!files[i].isDirectory())) {
			if (aPerl5Matcher.matches(files[i].getFilename(), mpattern)) {
				// call get for each match, passing true, so that it doesnt
				// repeat the search
				matchedNames.addElement(files[i]);
			}
		}
	}

	// return (SftpFile[]) matchedNames.toArray(new SftpFile[0]);
	SftpFile[] matchedNamesSftpFiles = new SftpFile[matchedNames.size()];
	matchedNames.copyInto(matchedNamesSftpFiles);
	return matchedNamesSftpFiles;
}
 

/**
 * compiles fileNameRegExp into a regular expression and pattern matches on
 * each file's name, and returns those that match.
 * 
 * @param files
 * @param fileNameRegExp
 * 
 * @return String[] of files that match the expresion.
 */
public String[] matchFileNamesWithPattern(File[] files,
		String fileNameRegExp) throws SshException {
	// set up variables for regexp matching
	Pattern mpattern = null;
	PatternCompiler aGCompiler = new GlobCompiler();
	PatternMatcher aPerl5Matcher = new Perl5Matcher();
	// Attempt to compile the pattern. If the pattern is not valid,
	// throw exception
	try {
		mpattern = aGCompiler.compile(fileNameRegExp);
	} catch (MalformedPatternException e) {
		throw new SshException("Invalid regular expression:"
				+ e.getMessage(), SshException.BAD_API_USAGE);
	}

	Vector<String> matchedNames = new Vector<String>();
	for (int i = 0; i < files.length; i++) {
		if ((!files[i].getName().equals("."))
				&& (!files[i].getName().equals(".."))) {
			if (aPerl5Matcher.matches(files[i].getName(), mpattern)) {
				// call get for each match, passing true, so that it doesnt
				// repeat the search
				matchedNames.addElement(files[i].getAbsolutePath());
			}
		}
	}

	// return (String[]) matchedNames.toArray(new String[0]);
	String[] matchedNamesStrings = new String[matchedNames.size()];
	matchedNames.copyInto(matchedNamesStrings);
	return matchedNamesStrings;
}
 

/**
 * compiles fileNameRegExp into a regular expression and pattern matches on
 * each file's name, and returns those that match.
 * 
 * @param files
 * @param fileNameRegExp
 * 
 * @return SftpFile[] of files that match the expresion.
 */
public SftpFile[] matchFilesWithPattern(SftpFile[] files,
		String fileNameRegExp) throws SftpStatusException, SshException {
	// set up variables for regexp matching
	Pattern mpattern = null;
	PatternCompiler aGCompiler = new GlobCompiler();
	PatternMatcher aPerl5Matcher = new Perl5Matcher();
	// Attempt to compile the pattern. If the pattern is not valid,
	// throw exception
	try {
		mpattern = aGCompiler.compile(fileNameRegExp);
	} catch (MalformedPatternException e) {
		throw new SshException("Invalid regular expression:"
				+ e.getMessage(), SshException.BAD_API_USAGE);
	}

	Vector<SftpFile> matchedNames = new Vector<SftpFile>();

	for (int i = 0; i < files.length; i++) {
		if ((!files[i].getFilename().equals("."))
				&& (!files[i].getFilename().equals(".."))
				&& (!files[i].isDirectory())) {
			if (aPerl5Matcher.matches(files[i].getFilename(), mpattern)) {
				// call get for each match, passing true, so that it doesnt
				// repeat the search
				matchedNames.addElement(files[i]);
			}
		}
	}

	// return (SftpFile[]) matchedNames.toArray(new SftpFile[0]);
	SftpFile[] matchedNamesSftpFiles = new SftpFile[matchedNames.size()];
	matchedNames.copyInto(matchedNamesSftpFiles);
	return matchedNamesSftpFiles;
}
 

public Pattern apply(String pattern) {
    try {
        PatternCompiler pc = new Perl5Compiler();
        return pc.compile(pattern,
            Perl5Compiler.CASE_INSENSITIVE_MASK
                    | Perl5Compiler.READ_ONLY_MASK
                    | Perl5Compiler.SINGLELINE_MASK);
    } catch (MalformedPatternException e) {
        throw new RuntimeException(e);
    }
}
 

/**
 * Extracts <code>Outlink</code> from given plain text and adds anchor
 * to the extracted <code>Outlink</code>s
 * 
 * @param plainText the plain text from wich URLs should be extracted.
 * @param anchor    the anchor of the url
 * 
 * @return Array of <code>Outlink</code>s within found in plainText
 */
public static Outlink[] getOutlinks(final String plainText, String anchor, Configuration conf) {
  long start = System.currentTimeMillis();
  final List<Outlink> outlinks = new ArrayList<Outlink>();

  try {
    final PatternCompiler cp = new Perl5Compiler();
    final Pattern pattern = cp.compile(URL_PATTERN,
        Perl5Compiler.CASE_INSENSITIVE_MASK | Perl5Compiler.READ_ONLY_MASK
            | Perl5Compiler.MULTILINE_MASK);
    final PatternMatcher matcher = new Perl5Matcher();

    final PatternMatcherInput input = new PatternMatcherInput(plainText);

    MatchResult result;
    String url;

    //loop the matches
    while (matcher.contains(input, pattern)) {
      // if this is taking too long, stop matching
      //   (SHOULD really check cpu time used so that heavily loaded systems
      //   do not unnecessarily hit this limit.)
      if (System.currentTimeMillis() - start >= 60000L) {
        if (LOG.isWarnEnabled()) {
          LOG.warn("Time limit exceeded for getOutLinks");
        }
        break;
      }
      result = matcher.getMatch();
      url = result.group(0);
      try {
        outlinks.add(new Outlink(url, anchor));
      } catch (MalformedURLException mue) {
        LOG.warn("Invalid url: '" + url + "', skipping.");
      }
    }
  } catch (Exception ex) {
    // if the matcher fails (perhaps a malformed URL) we just log it and move on
    if (LOG.isErrorEnabled()) { LOG.error("getOutlinks", ex); }
  }

  final Outlink[] retval;

  //create array of the Outlinks
  if (outlinks != null && outlinks.size() > 0) {
    retval = outlinks.toArray(new Outlink[0]);
  } else {
    retval = new Outlink[0];
  }

  return retval;
}
 

public cfData execute(cfSession _session, cfArgStructData argStruct ) throws cfmRunTimeException {
	String regexp = getNamedStringParam(argStruct, "regular", "");
	String strToSearch = getNamedStringParam(argStruct, "string", "");
	boolean bUnique = getNamedBooleanParam(argStruct, "unique",false);
	
	HashSet<String>	uniqueTrack = null;
	if ( bUnique ){
		uniqueTrack	= new HashSet<String>();
	}


	/* Setup the RegEx */
	PatternCompiler compiler = new Perl5Compiler();
	Pattern pattern;
	
	try {
		if (caseSensitiveMatch) {
			pattern = compiler.compile(regexp, Perl5Compiler.SINGLELINE_MASK);
		} else {
			pattern = compiler.compile(regexp, Perl5Compiler.CASE_INSENSITIVE_MASK | Perl5Compiler.SINGLELINE_MASK);
		}
	} catch (MalformedPatternException e) {
		cfCatchData catchD = new cfCatchData();
		catchD.setType("Function");
		catchD.setMessage("REMatch - invalid parameter");
		catchD.setDetail("Invalid regular expression ( " + regexp + " )");
		throw new cfmRunTimeException(catchD);
	}

	/* Perform the search */
	cfArrayData	array	= cfArrayData.createArray(1);
	PatternMatcher matcher = new Perl5Matcher();
	MatchResult result;
	PatternMatcherInput input = new PatternMatcherInput( strToSearch );
	while ( matcher.contains(input, pattern) ) {
		result = matcher.getMatch();
		
		String strResult = result.toString();
		if ( bUnique ){
			if ( !uniqueTrack.contains( strResult ) ){
				array.addElement( new cfStringData( strResult ) );
				uniqueTrack.add( strResult );
			}
		}else		
			array.addElement( new cfStringData( strResult ) );
	}
	
	return array;
}
 

/**
 * Extracts <code>Outlink</code> from given plain text and adds anchor
 * to the extracted <code>Outlink</code>s
 * 
 * @param plainText the plain text from wich URLs should be extracted.
 * @param anchor    the anchor of the url
 * 
 * @return Array of <code>Outlink</code>s within found in plainText
 */
public static Outlink[] getOutlinks(final String plainText, String anchor, Configuration conf) {
  long start = System.currentTimeMillis();
  final List<Outlink> outlinks = new ArrayList<Outlink>();

  try {
    final PatternCompiler cp = new Perl5Compiler();
    final Pattern pattern = cp.compile(URL_PATTERN,
        Perl5Compiler.CASE_INSENSITIVE_MASK | Perl5Compiler.READ_ONLY_MASK
            | Perl5Compiler.MULTILINE_MASK);
    final PatternMatcher matcher = new Perl5Matcher();

    final PatternMatcherInput input = new PatternMatcherInput(plainText);

    MatchResult result;
    String url;

    //loop the matches
    while (matcher.contains(input, pattern)) {
      // if this is taking too long, stop matching
      //   (SHOULD really check cpu time used so that heavily loaded systems
      //   do not unnecessarily hit this limit.)
      if (System.currentTimeMillis() - start >= 60000L) {
        if (LOG.isWarnEnabled()) {
          LOG.warn("Time limit exceeded for getOutLinks");
        }
        break;
      }
      result = matcher.getMatch();
      url = result.group(0);
      try {
        outlinks.add(new Outlink(url, anchor));
      } catch (MalformedURLException mue) {
        LOG.warn("Invalid url: '" + url + "', skipping.");
      }
    }
  } catch (Exception ex) {
    // if the matcher fails (perhaps a malformed URL) we just log it and move on
    if (LOG.isErrorEnabled()) { LOG.error("getOutlinks", ex); }
  }

  final Outlink[] retval;

  //create array of the Outlinks
  if (outlinks != null && outlinks.size() > 0) {
    retval = outlinks.toArray(new Outlink[0]);
  } else {
    retval = new Outlink[0];
  }

  return retval;
}