How to match “escape” non-printable character in a regex?

Question

I've found a howto, http://answers.oreilly.com/topic/214-how-to-match-nonprintable-characters-with-a-regular-expression/ , but non of the codes, \e, \x1b, \x1B, work for me in Java.

EDIT

I am trying to replace the ANSI escape sequences (specifically, color sequences) of a Linux terminal command's output. In Python the replace pattern would look like "\x1b[34;01m", which means blue bold text. This same pattern does not work in Java. I tried to replace "[34;01m" separately, and it worked, so the problem is \x1b. And I am doing the "[" escaping using Pattern.quote().

EDIT

Map<String,String> escapeMap = new HashMap<String,String>();
escapeMap.put("\\x1b[01;34m", "</span><span style=\"color:blue;font-weight:bold\">");
FileInputStream stream = new FileInputStream(new File("/home/ch00k/gun.output"));
FileChannel fc = stream.getChannel();
MappedByteBuffer bb = fc.map(FileChannel.MapMode.READ_ONLY, 0, fc.size());
String message = Charset.defaultCharset().decode(bb).toString();
stream.close();
String patternString = Pattern.quote(StringUtils.join(escapeMap.keySet(), "|"));
System.out.println(patternString);
Pattern pattern = Pattern.compile(patternString);
Matcher matcher = pattern.matcher(message);
StringBuffer sb = new StringBuffer();
while(matcher.find()) {
    matcher.appendReplacement(sb, escapeMap.get(matcher.group()));
    }
matcher.appendTail(sb);
String formattedMessage = sb.toString();
System.out.println(formattedMessage);

EDIT Here is the code I've ended up with:

import java.io.*;
import java.nio.*;
import java.nio.channels.*;
import java.nio.charset.*;
import java.util.*;
import java.util.regex.*;
import org.apache.commons.lang3.*;

class CreateMessage {
    public static void message() throws IOException {
        FileInputStream stream = new FileInputStream(new File("./gun.output"));
        FileChannel fc = stream.getChannel();
        MappedByteBuffer bb = fc.map(FileChannel.MapMode.READ_ONLY, 0, fc.size());
        String message = Charset.defaultCharset().decode(bb).toString();
        stream.close();
        Map<String,String> tokens = new HashMap<String,String>();
        tokens.put("root", "nobody");
        tokens.put(Pattern.quote("[01;34m"), "qwe");
        String patternString = "(" + StringUtils.join(tokens.keySet(), "|") + ")";
        Pattern pattern = Pattern.compile(patternString);
        Matcher matcher = pattern.matcher(message);
        StringBuffer sb = new StringBuffer();
        while(matcher.find()) {
            System.out.println(tokens.get(matcher.group()));
            matcher.appendReplacement(sb, tokens.get(matcher.group()));
        }
        matcher.appendTail(sb);

        System.out.println(sb.toString());

    }
}

The file gun.output contains the output of ls -la --color=always / Now, the problem is that I'm getting a NullPointerException if I'm trying to match Pattern.quote("[01;34m"). Everything matches fine except of the strings, that contain [, even though I quote them. The exception is the following:

Exception in thread "main" java.lang.NullPointerException
    at java.util.regex.Matcher.appendReplacement(Matcher.java:699)
    at org.minuteware.jgun.CreateMessage.message(CreateMessage.java:32)
    at org.minuteware.jgun.Main.main(Main.java:23)

EDIT

So, according to http://java.sun.com/developer/technicalArticles/releases/1.4regex/, the escape character should be matched with "\u001B", which indeed works in my case. The problem is, if I use tokens.put("\u001B" + Pattern.quote("[01;34m"), "qwe");, I still get the above mentioned NPE.

Answer 1

The proper value for "escape" character in a regexp is \u001B

Answer 2

quote() is to make a pattern that will match the input string verbatim. Your string has pattern language in it. Look at the output from quote() - you'll see that it's trying to literally find the four characters \x1b.

Answer 3

The ansi escape sequences are of the following form [\033[34;01m]

where \033 is ANSI character 033 (oct) or 1b in Hex or 27 in decimal. You need to use the following regexp:

Pattern p = Pattern.compile("\033\\[34;01m");

You can use an octal (\033) or hex (\x1b) representation when you're using a non-printable character in a java string.

Answer 4

FWIW, I've been working on stripping ANSI color codes from colorized log4j files and this little pattern seems to do the trick for all of the cases I've come across:

Pattern.compile("(\\u001B\\[\\d+;\\d+m)+")