easy-tokenizer 0.0.9

tokenizer tool

Easy-Tokenizer

Description

Most tokenizers are eithor too cumbersom (Neural Network based), or too simple. This simple rule based tokenizer is type, small, and sufficient good. Specially, it handles long strings very often parsed wrong by some simple tokenizers, deal url, email, long digits rather well.

Try with the following script:

easy_tokenizer -s input_text

or

easy_tokenizer -f input_file

CI Status

Requirements

Python 3.6+

Installation

pip install easy-tokenizer

Usage

• easy-tokenizer:

  input:

  • string: input string to tokenize

  • filename: input text file to tokenize

  • output: output filename, optional. print out to STDOUT when not set

  output:

  • a sequence of space separated tokens

examples:

# string input
easy-tokenizer -s "this is   a simple test."

easy-tokenizer -f foo.txt
easy-tokenizer -f foo.txt -o bar.txt

output will be “this is a simple test .”

Development

To install package and its dependencies, run the following from project root directory:

python setup.py install

To work the code and develop the package, run the following from project root directory:

python setup.py develop

To run unit tests, execute the following from the project root directory:

python setup.py test

0.0.9 (2020-01-16)

• [BUG] fixed infinite loop for long url

0.0.8 (2019-11-14)

• [New] added a function module for char/string normalization

0.0.7 (2019-11-14)

• [Bugfix] update the url patter to fix the regexp loop for long url string

0.0.5 (2019-10-23)

• [Bugfix] encryption and doc generation

0.0.3 (2019-10-23)

• Test the CI/CD and auto documentation generation

0.0.2 (2019-10-23)

• support script to output result to a file, add documentation

0.0.1 (2019-10-22)

• Add the first version of the tokenizer