myre - Human Readable Regular Expression

Language: English | 简体中文

myre is a composable regular expression library that transforms regex patterns into first-class objects supporting algebraic operations. It provides a human-friendly way to combine, filter, and transform pattern matching in Python, maintaining full compatibility with the standard re module interface.

Features

• Algebraic Operations: Combine patterns using | (OR), & (AND), + (SEQUENCE), / (SPLIT), ^ (XOR/EXCLUDE), @ (MASK)
• Type-Safe: Protocol-based design with full type hints
• re-Compatible: Drop-in replacement for standard re module patterns
• Composable: Patterns are immutable objects that can be combined like mathematical sets
• Zero Dependencies: Only requires typing-extensions for modern type hints

Installation

pip install myre

Quick Start

import myre

# OR operation: match any pattern (default)
pattern = myre.compile(r"hello", r"world", r"foo")
for match in pattern.finditer("hello world foo"):
    print(match.group())  # Prints: hello, world, foo

# AND operation: all patterns must match
from myre import Mode
pattern = myre.compile(r"hello", r"world", mode=Mode.ALL)
for match in pattern.finditer("hello world"):
    print(match.group())  # Prints: hello world

# SEQUENCE operation: match patterns in order
pattern = myre.compile(r"hello") + myre.compile(r"world")
for match in pattern.finditer("hello world"):
    print(match.group())  # Prints: hello world

# SPLIT operation: split by delimiter, match pattern in each segment
pattern = myre.compile(r"\d{2}") / r"[,-]"
for match in pattern.finditer("12,34-56"):
    print(match.group())  # Prints: 12, 34, 56 (each with correct position)

# XOR operation: match pattern but exclude another
pattern = myre.compile(r"h\wllo") ^ r"ell"
for match in pattern.finditer("Hello Hallo Hillo"):
    print(match.group())  # Prints: Hallo, Hillo (excludes "Hello")

# MASK operation: ignore specific patterns by masking with placeholders
pattern = myre.compile(r"test_{3}value") @ (r"\d+", "_")
for match in pattern.finditer("test123value"):
    print(match.group())  # Prints: test123value

Core Concepts

Pattern Types

MatchAny (Mode.ANY - default)

Matches any of the given patterns, avoiding overlapping matches:

from myre import Mode

pattern = myre.compile(r"abc", r"bcd", r"cde", mode=Mode.ANY)
# or simply: pattern = myre.compile(r"abc", r"bcd", r"cde")
for match in pattern.finditer("abc bcd cde"):
    print(match.group())  # Non-overlapping matches

MatchALL (Mode.ALL)

Requires all patterns to match somewhere in the search range. Returns a combined match:

pattern = myre.compile(r"hello", r"world", mode=Mode.ALL)
for match in pattern.finditer("hello world"):
    print(match.group())  # "hello world"

Note: MatchALL collects the first match from each pattern and combines them into a single match result.

Operators

Operator	Name	Description	Example
|	OR	Match any pattern	p1 | p2 → MatchAny(p1, p2)
&	AND	Match all patterns	p1 & p2 → MatchALL(p1, p2)
+	SEQUENCE	Match patterns in order	p1 + p2 → MatchSeq(p1, p2)
/	SPLIT	Split by delimiter, match in segments	p1 / p2 → SplitMatch(p1, p2)
^	XOR	Exclude matches containing deny pattern	p1 ^ p2 → Match p1, exclude if text contains p2
@	MASK	Mask patterns with placeholders	p1 @ (mask, placeholder) → Replace mask with placeholder, match p1

Operator Usage Notes

^ (XOR): Currently operates at the string level - if the entire search string contains the deny pattern, all matches are rejected. This is different from traditional set XOR.

@ (MASK): Replaces matched text with placeholder characters (preserving length), then matches the pattern. Returns the original text (not masked text). Choose a placeholder that works with your pattern:

• Use 'x' if pattern uses x+ or x{N} to match placeholders
• Use '_' if pattern uses _{3} etc.
• Default is '.' (dot character)

API Reference

PatternLike Protocol

All pattern objects implement the PatternLike protocol:

class PatternLike(Protocol):
    def search(self, string: str, pos: int = 0, endpos: int = sys.maxsize) -> Optional[MatchLike]: ...
    def findall(self, string: str, pos: int = 0, endpos: int = sys.maxsize) -> List[str]: ...
    def finditer(self, string: str, pos: int = 0, endpos: int = sys.maxsize) -> Iterator[MatchLike]: ...

TODO: Standard library methods not yet implemented:

• match() - Match at the beginning of the string
• fullmatch() - Match the entire string
• split() - Split string by pattern occurrences
• sub() / subn() - Replace pattern matches
• flags - Access pattern flags
• pattern - Access pattern string
• groupindex - Access group name mapping

MatchLike Protocol

All match objects implement the MatchLike protocol:

class MatchLike(Protocol):
    re: PatternLike
    string: str

    def start(self, group: int | str = 0) -> int: ...
    def end(self, group: int | str = 0) -> int: ...
    def span(self, group: int | str = 0) -> Tuple[int, int]: ...
    def group(self, group: int | str = 0) -> str: ...
    def groups(self) -> tuple[str, ...]: ...

TODO: Standard library methods not yet implemented:

• groupdict() - Return dict of named groups
• expand() - Expand template using groups
• lastindex - Last matched group index
• lastgroup - Last matched group name
• pos / endpos - Search boundaries

Pattern Classes

Base

@dataclass
class Base:
    p_trim: PatternLike  # Pre-process filter
    p_deny: PatternLike  # Post-process filter

    def finditer(self, string: str, pos: int = 0, endpos: int = sys.maxsize) -> Iterator[MatchLike]
    def findall(self, string: str, pos: int = 0, endpos: int = sys.maxsize) -> List[str]
    def search(self, string: str, pos: int = 0, endpos: int = sys.maxsize) -> Optional[MatchLike]

MatchAny

@dataclass
class MatchAny(Base):
    patterns: tuple[PatternLike, ...]

    @classmethod
    def compile(cls, *patterns: Any, flag: int = 0) -> Self

MatchALL

@dataclass
class MatchALL(MatchAny):
    patterns: tuple[PatternLike, ...]

    @classmethod
    def compile(cls, *patterns: Any, flag: int = 0) -> Self

Advanced Usage

Working with re.Pattern

import re
import myre

# Mix regex strings and compiled patterns
regex1 = re.compile(r"hello")
pattern = myre.compile(regex1, r"world", flag=re.IGNORECASE)

Nested Operations

# Complex pattern: (hello OR world) BUT NOT (foo OR bar)
base = myre.compile(r"hello", r"world")
exclude = myre.compile(r"foo", r"bar")
pattern = base ^ exclude

Position-Aware Matching

# Mask punctuation before matching
pattern = myre.compile(r"\bhello\b") @ (r"[.,!?;:]", " ")
for match in pattern.finditer("Hello, world! Hello."):
    print(match.group())  # Prints both "Hello," and "Hello."

Architecture

The library is built on three architectural layers:

1. Protocol Layer (protocol.py): Defines structural types for pattern/match compatibility
2. Pattern Layer (pattern.py): Implements algebraic operations using Template Method pattern
3. Match Layer (match.py): Provides position-aware match result wrappers

Key Design Patterns

• Template Method: Base.finditer() orchestrates the matching pipeline
• Protocol-based Polymorphism: Structural subtyping over nominal inheritance
• Immutable Operators: All operations create new objects via deepcopy

Development

# Install dependencies
uv sync --extra dev

# Run tests
pytest tests/

# Type checking
mypy myre/

# Linting and formatting
ruff check .
ruff format .

Contributing

Contributions are welcome! Please ensure:

• All tests pass: pytest tests/
• Type checking passes: mypy myre/
• Code is formatted: ruff format .

License

MIT License - see LICENSE for details.

Acknowledgments

Inspired by the need for composable, human-readable regular expressions in Python.