Huffman Coding (Greedy Algorithms) in Java

Introduction

This repository was created to share my project in "Data Structures and Algorithms in Java" class.
What I did in the project are:

• Implemented Huffman Coding in Java
• Implemented function to automatically generate .dot file for Graphviz software to visualize the Huffman Tree
• Tested the code with Unix's words file (/usr/share/dict/words)
• Get the result of data compression rate

Key techniques used

• Greedy Algorithms
• HashMap
• PriorityQueue (Min Heap)
• Nested Class
• Recursion

Requirements

Graphviz software is required to visualize the Huffman Tree.
Please download it on here

How to test

1. Prepare your own text file

Create a text file named "original.txt" under /files with any contents. In this demo, I used Unix's words file which contains 235,886 dictionary words(2,493,109 characters in total).

$ cd <PATH TO THE PROJECT DIRECTORY>/files
$ cat /usr/share/dict/words > original.txt

2. Run the test code

"Test.java" automatically loads the original.txt, then test encoding/decoding the given text, and will show the result of text compression.
The actual result will be something like below.

----- Test.java START -----
* Loading the file...DONE
* Builiding Huffman Tree and Code Tables... DONE

============= Word Frequency =============
\n occurs 235886 times
- occurs 2 times
A occurs 2559 times
B occurs 1395 times
C occurs 2496 times
D occurs 950 times
E occurs 918 times
F occurs 478 times
G occurs 1018 times
H occurs 1138 times
I occurs 496 times
J occurs 484 times
K occurs 546 times
L occurs 1073 times
M occurs 1907 times
N occurs 682 times
O occurs 630 times
P occurs 2291 times
Q occurs 77 times
R occurs 716 times
S occurs 2403 times
T occurs 1577 times
U occurs 208 times
V occurs 361 times
W occurs 337 times
X occurs 92 times
Y occurs 139 times
Z occurs 230 times
a occurs 196995 times
b occurs 39038 times
c occurs 100944 times
d occurs 67241 times
e occurs 234413 times
f occurs 23687 times
g occurs 46076 times
h occurs 63218 times
i occurs 200536 times
j occurs 2683 times
k occurs 15612 times
l occurs 129390 times
m occurs 68773 times
n occurs 158061 times
o occurs 170062 times
p occurs 75872 times
q occurs 3657 times
r occurs 160269 times
s occurs 137139 times
t occurs 151254 times
u occurs 87145 times
v occurs 19816 times
w occurs 13527 times
x occurs 6840 times
y occurs 51542 times
z occurs 8230 times

========== Huffman Code for each character =============
\n: 001
-: 0110110111010100
A: 0110110110
B: 01101101111
C: 0110110101
D: 01101100000
E: 01101000001
F: 011011000010
G: 01101100010
H: 01101100101
I: 011011000110
J: 011011000011
K: 011011000111
L: 01101100100
M: 0110100001
N: 111100100100
O: 011011011100
P: 0110110011
Q: 0110110111010101
R: 111100100101
S: 0110110100
T: 11110010011
U: 0110100000010
V: 011010000000
W: 0110110111011
X: 011011011101011
Y: 01101101110100
Z: 0110100000011
a: 1101
b: 101010
c: 11001
d: 01001
e: 000
f: 1111000
g: 101011
h: 01000
i: 1110
j: 1111001000
k: 0110101
l: 11111
m: 01100
n: 1000
o: 1011
p: 10100
q: 011010001
r: 1001
s: 0101
t: 0111
u: 11000
v: 0110111
w: 11110011
x: 111100101
y: 111101
z: 01101001

* Encoding the text... DONE
* Decoding the encoded text... DONE

========== RESULT ==========
Original string cost = 17451763 bits
Encoded  string cost = 10872876 bits
% reduction = 37.697549525512116

----- Test DONE ----- 

The program will automatically create "test.dot" for visualizing the tree.

3. Visualize the tree with Graphviz

Run below command and it will create test.pdf which visualized the Huffman Tree built for the test

$ dot -Tpdf test.dot -o test.pdf

It will create a pdf file like below.

Each circle represents a tree node with below properties.

• Unique ID of a node
• Weight of a node
• Character (leaf nodes only)