trrex 0.0.4

Transform trie to regular expression

Efficient keyword extraction with regex

This package contains a function for efficiently representing a set of keywords as regex. This regex can be used to replace keywords in sentences or extract keywords from sentences

Why use trrex?

• Pure Python, no other dependencies
• trrex is fast, about 300 times faster than a regex union, and about 2.5 times faster than FlashText
• Plays well with others, can be integrated easily with pandas

Install trrex

Use pip,

pip install trrex

Usage

import trrex as tx

pattern = tx.compile(['baby', 'bat', 'bad'])
hits = pattern.findall('The baby was scared by the bad bat.')
# hits = ['baby', 'bat', 'bad']

pandas

import trrex as tx
import pandas as pd

frame = pd.DataFrame({
    "txt": ["The baby", "The bat"]
})
pattern = tx.make(['baby', 'bat', 'bad'], prefix=r"\b(", suffix=r")\b") # need to specify capturing groups
frame["match"] = frame["txt"].str.extract(pattern)
hits = frame["match"].tolist()
print(hits)
# hits = ['baby', 'bad']

Why the name?

Naming is difficult, but as we had to call it something:

• trex: trie to regex
• trex: Tyrannosaurus rex, a large dinosaur species with small arms (rex meaning "king" in Latin)

Acknowledgments

This project is based on the following resources:

• Speed up regex
• Triegex