Duppix is a comprehensive regex library that brings Oniguruma-compatible advanced features to Dart, including possessive quantifiers, atomic groups, named backreferences, recursive patterns, and much more.
Dart's built-in RegExp is powerful but lacks many advanced features that other regex engines provide. Duppix fills this gap by implementing a hybrid approach:
- 🚀 Fast fallback: Simple patterns use Dart's optimized
RegExp - 🎯 Advanced features: Complex patterns use our custom engine
- 🔄 Full compatibility: Drop-in replacement for
RegExp - 📚 Oniguruma compatible: Supports the same syntax as Ruby, PHP PCRE, and more
| Feature | Dart RegExp | Duppix | Example |
|---|---|---|---|
| Basic patterns | ✅ | ✅ | \d+, [a-z]* |
| Named groups | ✅ | ✅ | (?<name>\w+) |
| Backreferences | ✅ | \1, \k<name> |
|
| Possessive quantifiers | ❌ | ✅ | \d++, .*+ |
| Atomic groups | ❌ | ✅ | (?>...) |
| Recursive patterns | ❌ | ✅ | (?R), (?0) |
| Subroutine calls | ❌ | ✅ | (?1), (?&name) |
| Conditional patterns | ❌ | ✅ | (?(1)yes|no) |
| Variable lookbehind | ❌ | ✅ | (?<=\w{2,4}) |
| Script runs | ❌ | ✅ | (?script_run:...) |
Add Duppix to your pubspec.yaml:
dependencies:
duppix: ^1.0.0import 'package:duppix/duppix.dart';
void main() {
// Works just like RegExp for simple patterns
final basic = DuppixRegex(r'\d+');
print(basic.firstMatch('Hello 123')?.group); // "123"
// But supports advanced features too!
final advanced = DuppixRegex(r'(?<word>\w+)\s+\k<word>');
final match = advanced.firstMatch('hello hello world');
print(match?.namedGroup('word')); // "hello"
}// Match repeated words
final regex = DuppixRegex(r'(?<word>\w+)\s+\k<word>');
final match = regex.firstMatch('hello hello world');
print(match?.namedGroup('word')); // "hello"
// Case-insensitive backreferences
final regex2 = DuppixRegex(r'(?<tag>\w+).*?</\k<tag>>',
options: DUPPIX_OPTION_IGNORECASE);// Atomic matching - no backtracking
final greedy = DuppixRegex(r'.*abc'); // Can backtrack
final possessive = DuppixRegex(r'.*+abc'); // Cannot backtrack
// Useful for performance optimization
final efficient = DuppixRegex(r'\d++[a-z]'); // Faster than \d+[a-z]// Prevent backtracking within groups
final atomic = DuppixRegex(r'(?>.*?)end');
final match = atomic.firstMatch('start middle end');// Match balanced parentheses
final balanced = DuppixRegex(r'\((?:[^()]|(?R))*\)');
final match = balanced.firstMatch('(a(b(c)d)e)');
print(match?.group); // "(a(b(c)d)e)"
// Match nested structures
final nested = DuppixRegex(r'<(\w+)>(?:[^<>]|(?R))*</\1>');// Define reusable patterns
final regex = DuppixRegex(r'(?<digit>\d)(?<letter>[a-z])(?&digit)(?&letter)');
final match = regex.firstMatch('1a1a');
// Numbered subroutine calls
final numbered = DuppixRegex(r'(\d{2})-(?1)-(?1)'); // Match XX-XX-XX format// Match based on conditions
final conditional = DuppixRegex(r'(?(<tag>)yes|no)');
// Matches "yes" if named group "tag" was captured, "no" otherwise// Script runs - ensure single Unicode script
final scriptRun = DuppixRegex(r'(?script_run:\w+)');
// Character class operations
final intersection = DuppixRegex(r'[a-z&&[^aeiou]]'); // Consonants only// Configure regex behavior
final options = DuppixOptions(
ignoreCase: true,
multiline: true,
singleline: false,
unicode: true,
findLongest: false,
debug: false,
);
final regex = DuppixRegex(r'pattern', options: options);
// Or use flags (Oniguruma compatible)
final flagged = DuppixRegex(r'pattern',
options: DUPPIX_OPTION_IGNORECASE | DUPPIX_OPTION_MULTILINE);Duppix is designed as a drop-in replacement for RegExp:
// Before (RegExp)
final oldRegex = RegExp(r'\d+');
final oldMatch = oldRegex.firstMatch('123');
// After (Duppix) - same API
final newRegex = DuppixRegex(r'\d+');
final newMatch = newRegex.firstMatch('123');
// All the same methods work
print(newRegex.hasMatch('123'));
print(newRegex.allMatches('1 2 3').length);
print(newRegex.replaceAll('a1b2c', 'X'));Duppix uses a smart hybrid approach:
- Simple patterns → Dart's optimized
RegExp(fastest) - Advanced patterns → Custom engine (feature-rich)
- Automatic detection → No manual configuration needed
// This uses fast RegExp fallback
final simple = DuppixRegex(r'\d+');
// This uses custom engine (detected automatically)
final advanced = DuppixRegex(r'\d++'); // Possessive quantifierRun the comprehensive test suite:
dart testTests cover:
- ✅ All basic RegExp functionality
- ✅ Advanced Oniguruma features
- ✅ Performance edge cases
- ✅ Unicode support
- ✅ Error handling
- ✅ Legacy compatibility
final emailRegex = DuppixRegex(
r'(?<local>[a-zA-Z0-9._%+-]+)@(?<domain>[a-zA-Z0-9.-]+\.[a-zA-Z]{2,})'
);
final match = emailRegex.firstMatch('[email protected]');
print('Local: ${match?.namedGroup('local')}'); // "user"
print('Domain: ${match?.namedGroup('domain')}'); // "example.com"final urlRegex = DuppixRegex(
r'(?<protocol>https?)://(?<domain>(?&subdomain)\.)*(?<tld>\w+)(?<path>/.*)?'
r'(?<subdomain>\w+)'
);final brackets = DuppixRegex(r'\{(?:[^{}]|(?R))*\}');
final json = '{"key": {"nested": "value"}}';
print(brackets.firstMatch(json)?.group); // Full JSON objectfinal htmlTag = DuppixRegex(r'<(?<tag>\w+)>.*?</\k<tag>>');
final html = '<div>Content</div>';
print(htmlTag.firstMatch(html)?.namedGroup('tag')); // "div"- Core regex engine architecture
- Pattern parser with full Oniguruma syntax
- Hybrid fallback system
- Basic quantifiers (*, +, ?, {n,m})
- Character classes and ranges
- Named and numbered capture groups
- Possessive quantifiers (*+, ++, ?+)
- Atomic groups ((?>...))
- Lookahead/lookbehind assertions
- Backreferences (\1, \k)
- Subroutine calls ((?1), (?&name))
- Recursive patterns ((?R))
- Conditional patterns framework
- Comprehensive error handling
- Full RegExp API compatibility
- Unicode property support (\p{Letter}, \p{Script=Latin})
- Anchor improvements (^, $, \b, \B)
- Performance optimizations
- Additional Oniguruma features
- Variable-length lookbehind optimization
- More Unicode features
- JIT compilation for hot patterns
- WASM acceleration
- Additional language support
We welcome contributions! Areas where help is needed:
- Unicode Properties - Implement \p{...} classes
- Performance - Optimize hot paths
- Documentation - More examples and guides
- Testing - Edge cases and real-world patterns
- Features - Additional Oniguruma compatibility
MIT License - see LICENSE for details.
- Oniguruma - The inspiration and syntax reference
- Ruby - For pioneering advanced regex features
- PCRE - For performance insights
- Dart Team - For the excellent base RegExp implementation
Made with ❤️ for the Dart community