Skip to main content

Pure Python Huffman encoder and decoder module

Project description

https://img.shields.io/travis/soxofaan/dahuffman/master.svg https://img.shields.io/badge/license-MIT-blue.svg https://img.shields.io/pypi/v/dahuffman

dahuffman is a pure Python module for Huffman encoding and decoding, commonly used for lossless data compression.

The name of the module refers to the full name of the inventor of the Huffman code tree algorithm: David Albert Huffman (August 9, 1925 – October 7, 1999).

Features and design

  • Pure Python implementation, only using standard library.
  • Leverages iterators and generators internally, allows to be used in streaming fashion.
  • Not limited to byte/unicode string input, can handle other “symbols” or tokens, for example chess moves or sequences of categorical data, as long as these symbols can be used as keys in dictionaries (meaning they should be hashable).
  • Properly handle end of encoded bit stream if it does not align with byte boundaries
  • For Python 3.5 and up

Installation

pip install dahuffman

Usage

Basic usage example, where the code table is built based on given symbol frequencies:

>>> from dahuffman import HuffmanCodec
>>> codec = HuffmanCodec.from_frequencies({'e': 100, 'n':20, 'x':1, 'i': 40, 'q':3})
>>> encoded = codec.encode('exeneeeexniqneieini')
>>> encoded
b'\x86|%\x13i@'
>>> len(encoded)
6
>>> codec.decode(encoded)
'exeneeeexniqneieini'
>>> codec.print_code_table()
bits  code       (value)  symbol
   5  00000      (    0)  _EOF
   1  1          (    1)  'e'
   2  01         (    1)  'i'
   3  001        (    1)  'n'
   4  0001       (    1)  'q'
   5  00001      (    1)  'x'

You can also “train” the codec by providing it data directly:

>>> codec = HuffmanCodec.from_data('hello world how are you doing today foo bar lorem ipsum')
>>> codec.encode('do lo er ad od')
b'^O\x1a\xc4S\xab\x80'
>>> len(_)
7

Using it with sequences of symbols (country codes in this example):

>>> countries = ['FR', 'UK', 'BE', 'IT', 'FR', 'IT', 'GR', 'FR', 'NL', 'BE', 'DE']
>>> codec = HuffmanCodec.from_data(countries)
>>> encoded = codec.encode(['FR', 'IT', 'BE', 'FR', 'UK'])
>>> encoded
b'L\xca'
>>> len(encoded)
2
>>> codec.decode(encoded)
['FR', 'IT', 'BE', 'FR', 'UK']

Doing it in a streaming fashion (generators):

>>> import random
>>> def sample(n, symbols):
...     for i in range(n):
...             if (n-i) % 5 == 1:
...                     print(i)
...             yield random.choice(symbols)
...
>>> codec = HuffmanCodec.from_data(countries)
>>> encoded = codec.encode_streaming(sample(16, countries))
>>> encoded
<generator object encode_streaming at 0x108bd82d0>
>>> decoded = codec.decode_streaming(encoded)
>>> decoded
<generator object decode_streaming at 0x108bd8370>
>>> list(decoded)
0
5
10
15
['DE', 'BE', 'FR', 'GR', 'UK', 'BE', 'UK', 'IT', 'UK', 'FR', 'DE', 'IT', 'NL', 'IT', 'FR', 'UK']

Pre-trained codecs

The dahuffman.codecs package contains a bunch of pre-trained code tables. The codecs can be loaded as follows:

>>> from dahuffman import load_shakespeare
>>> codec = load_shakespeare()
>>> codec.print_code_table()
Bits Code                     Value Symbol
   4 0000                         0 'n'
   4 0001                         1 's'
   4 0010                         2 'h'
   5 00110                        6 'u'
   7 0011100                     28 'k'
   9 001110100                  116 'Y'
  14 00111010100000            3744 '0'
...
>>> len(codec.encode('To be, or not to be; that is the question;'))
24

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Files for dahuffman, version 0.4.0
Filename, size File type Python version Upload date Hashes
Filename, size dahuffman-0.4.0-py2.py3-none-any.whl (17.4 kB) File type Wheel Python version py2.py3 Upload date Hashes View
Filename, size dahuffman-0.4.0.tar.gz (16.5 kB) File type Source Python version None Upload date Hashes View

Supported by

AWS AWS Cloud computing Datadog Datadog Monitoring DigiCert DigiCert EV certificate Facebook / Instagram Facebook / Instagram PSF Sponsor Fastly Fastly CDN Google Google Object Storage and Download Analytics Pingdom Pingdom Monitoring Salesforce Salesforce PSF Sponsor Sentry Sentry Error logging StatusPage StatusPage Status page