fastzy 0.1.0

Python library for fast fuzzy search over a big file leveraging C++ and mbleven algorithm

Python library for fast fuzzy search over a big file leveraging C++ and mbleven algorithm

Table of Contents

• Table of Contents
• About The Project
  • Built With
  • Performance
  • Prerequisites
  • Installation
• Usage
• License
• Contact

About The Project

Fastzy is a library written in C++ used for searching over a file for a text based on its distance (levenshtein). The library uses mbleven algorithm for a k-bounded levenshtein distance measurement. When the max distance requested is above 3, where mbleven should be slower, the distance algorithm is replaced with Wagner–Fischer.The library at first, loads the whole file into memory, and created a lightweight index, based on the length of the line. It helps to narrow down the amount of lookups to only potential lines.

Built With

• mbleven

Performance

Library	Text Size	Function	Time	#Results	Improvement Factor
python-Levenshtein	500mb	Levenshtein.distance('text')	24.2 s	1249	1.0x
fastzy	500mb	fastzy.lookup('text)	22.2 ms	1249	1090.0x

Prerequisites

In order to compile this package you should have GCC & Python development package installed.

• Fedora

sudo dnf install python3-devel gcc-c++

• Ubuntu 18.04

sudo apt install python3-dev g++-8

Installation

pip3 install fastzy

Usage

import fastzy

# open a file and index it in memory
searcher = fastzy.Searcher(
    input_file_path='input_text_file.txt',
    separator='',
)

# lookup for the input text 'text' with the distance of 1
searcher.lookup(
    pattern='text',
    max_distance=1,
)
['test', 'texts', 'next']

License

Distributed under the MIT License. See LICENSE for more information.

Contact

Gal Ben David - [e-mail]

Project Link: https://github.com/wavenator/fastzy