Skip to content

Latest commit

 

History

History
94 lines (77 loc) · 3.8 KB

README.md

File metadata and controls

94 lines (77 loc) · 3.8 KB
Raw

RegExpBuilder - a Regular Expression builder for Java.

Regular expressions are powerful but not easily readable, especially when coming back to one after a while. It's not without reason that the 'saying' goes:

Once a programmer had a problem, he tried to solve that with regular expressions, then he had two problems.

The reason regexps are not easily readable is because of their highly technical and compact (aka old fashioned) notation. Less is not always more. Adding a little verbosity sometimes can make things more readable. Like using names for patterns, instead of remembering what the heck \S meant (and no, I did not mean \s). It is not without reason single letter variable names are considered not done in programming.

Now, there are already some implementations of Regex builders, like the one by Sergey. Which probably is a very good one, but I'm not sure I like the API; its resulting code does not make reading the regex any easier to me. So this is my attempt at writing a regular expression builder:

import org.tbee.regexpbuilder.RegExp;
import static org.tbee.regexpbuilder.RE.*;

RegExp regExp = RegExp.of()
                      .startOfLine()
                      .optional("abc")
                      .group("g1", oneOrMore(digit().digit()))
                      .group("g2", notOneOf("xyz"));
Matcher matcher = regExp.toMatcher("abc1234de");
matcher.find();
String g1 = matcher.group(regExp.indexOf("g1")); // results in "1234"

Or a little bit more complex example for log lines:

// RegExp for this: 127.0.0.1 - - [21/Jul/2014:9:55:27 -0800] "GET /home.html HTTP/1.1" 200 2048
RegExp regExp = RegExp.of()
        .group("ip", oneOrMore(nonWhitespace()))
        .text(" ")
        .group("client", oneOrMore(nonWhitespace()))
        .text(" ")
        .group("user", oneOrMore(nonWhitespace()))
        .text(" [")
        .group("datetime", oneOrMore(oneOf(word().or().text(":/"))))
        .text(" ")
        .group("offset", oneOf("+-").followedBy().occurs(4, digit()))
        .text("] \"")
        .group("method", oneOrMore(nonWhitespace()))
        .text(" ")
        .group("url", oneOrMore(nonWhitespace()))
        .text(" ")
        .group("http", oneOrMore(nonWhitespace()))
        .text("\" ")
        .group("status", oneOrMore(digit()))
        .text(" ")
        .group("size", oneOrMore(digit()));

String logLine = "127.0.0.1 - - [21/Jul/2014:9:55:27 -0800] \"GET /home.html HTTP/1.1\" 200 2048";
Matcher matcher = regExp.toMatcher(logLine);
matcher.matches();
matcher.group(regExp.indexOf("datetime"); // returns "21/Jul/2014:9:55:27"

Or pattern reuse:

// for #ABC or #ABCDEF
RegExp colorchar = range("a", "f", "A", "F", "0", "9");
RegExp regExp = RegExp.of()
    .text("#")
    .group(occurs(3, colorchar))
    .optional(group(occurs(3, colorchar)));

The or(), and(), and followedBy() are dummy methods, solely present for readability and may be omitted.

As always; give it a shot, and let me know where I messed up. There are a some places in the API which I'm not totally happy with, so I'd love suggestions. (For example oneOf(word().or().text(":/")) suggests either a word or the text ":/", but it actually is a word or the character ":" or the character "/".)

Support

There is no formal support for RegExpBuilder: this library is an open source hobby project and no claims can be made. Asking for help is always an option. But so is participating, creating pull requests, and other ways of contributing.

Usage

Just include a dependency in your project. For the latest version see Maven central

<dependency>
    <groupId>org.tbee.regexpbuilder</groupId>
    <artifactId>regexpbuilder</artifactId>
    <version>1.1.0</version>
</dependency>