rector — Regex Generator

Rector turns example strings into optimized regular expressions you can reuse in search, validation, or routing rules. It ships with a trie-based engine that produces compact, deterministic patterns (also used by the CLI).

Installation

pip install rector

Library usage

from rector import TrieRegexGenerator

# Build a minimal trie-based pattern for larger sets or ranges
numbers = (
    TrieRegexGenerator()
    .add_examples([str(n) for n in range(10, 1000)])
    .generate()
)
assert numbers.pattern == "^[1-9][0-9]{1,2}$"

# Generate from small fixed-length samples
words = TrieRegexGenerator().add_examples(["cat", "car", "cap"]).generate()
assert words.pattern == r"^ca[prt]$"

CLI

Pipe examples to rector and it prints the generated pattern to stdout:

printf "cat\ncar\ncap\n" | rector
# -> ^ca[prt]$

The CLI anchors patterns by default and exits with status 1 if no stdin input is provided. Use --dense-threshold to adjust when near-contiguous character sets are collapsed into ranges (accepts fractions.Fraction strings like 3/4 or 0.6; default when flag is present without a value: 8/10).

Behavior notes

• Patterns anchor to start/end unless you pass anchor_start=False or anchor_end=False.
• Examples must be strings; TrieRegexGenerator raises TypeError otherwise.
• TrieRegexGenerator builds a trie, minimizes it, and compacts ranges/quantifiers (using the density threshold for near-contiguous classes); it also accepts empty lines as valid examples.