Description
Important Instructions
1. List of Input Files: log.txt, encode.txt, decode.txt. Example files are provided in Moodle and at the end of this PDF.
2. Files to be submitted: ROLLNO GroupNO A4.c/.cpp
3. Your code should output a file ‘output.txt‘ when run. Again, example files are provided in Moodle.
4. Files must be read using file handling APIs of C/C++. Reading through ‘input redirection‘/‘scanf‘ isn’t allowed.
6. ONLY PART I AND PART II ARE COMPULSORY FOR TODAY’S ASSIGNMENT. PART III AND IV ARE BONUS.
7. IF YOU SUBMIT THE BONUS PARTS, THOSE WILL BE GRADED AND CONSIDERED ONLY IN THE CASE OF SOMEONE FALLING ON A CLOSE GRADE BOUNDARY IN THIS COURSE AT THE END OF THE SEMESTER.
8. Everything must be submitted within lab hours.
Huffman Coding
Intro
For communicating through digital mediums, we must encode messages into bit streams first. In such a system, the sender needs an encoder for encoding text messages into bit streams, and the recipient, on receiving this bit stream, uses a decoder for reconstructing the original message. Today we will see such an algorithm called Huffman Coding, which can encode text messages into bit streams of minimal length. The algorithm is based on a binarytree frequency-sorting method that allows to encode any message sequence into a bit stream and reassemble any encoded message into its original version without losing any data. When you compare between fixed-length binary encoding and Huffman coding, the later has the least possible expected message length (under certain constraints).
Objectives of this assignment
This assignment has four parts. In part I, you will have to estimate frequencies of every character from the log file. In part II you will have to compute the Huffman codes for each character. In part III you will have to build the encoder based on Huffman codes. In the final part you will have to build a decoder to retrieve the data back.
Part I – Observing Frequency Distribution
As a first step towards building an efficient prefix code for communication, you would need to gather symbol probabilities first. Thankfully, you have access to a log of previously communicated messages as a text file. The format of the file is as follows,
FILE: log.txt
16 the quick brown fox jumps over a lazy dog the five boxing wizards jump quickly ten10
The first line contains the number of words (messages) recorded in log.txt. From this file you need to read in all the words and build a symbol (character) frequency list. As output of this part you should print the frequency list as the very first line of your output.txt. The set of symbols is limited to lowercase alphanumeric characters. And you must print them in proper order, a to z and then 0 to 9.
Example output:
a=3,b=2,c=2,d=2,e=5,f=2,g=2,h=2,i=5,j=2,k=2,l=2,m=2,n=3,o=5,p=2,q=2,r=3,s=2, t=3,u=4,v=2,w=2,x=2,y=2,z=2,0=1,1=1,2=0,3=0,4=0,5=0,6=0,7=0,8=0,9=0
Note: For any symbol not occurring in the log, set their frequency to 0 and include them in all your computations ahead.
Part II – Finding the Huffman Coding
Step 1
Now that we have our frequency list ready, we will implement the Huffman Coding Algorithm. We will start with a sorted linked list of symbol frequencies. See Figure 1 for an example. When facing equal count, for conflict resolution use the priority order,
a < b < … < z < 0 < 1 < … < 9 (1)
Data structure
The basic data structure should follow the template below. We will explain the use of the pointers ‘left’ and ‘right’ soon.
class Node{ char *symbol; int frequency;
Node *next;
Node *left;
Node *right;
}
Initially, all left and right pointers set to null.
Step 2 (SuperNodes)
Remove the two lowest frequency nodes from the linked list and merge them together to create a SuperNode Ni (i: an incremental id starting from 1). This supernode Ni should have the two merged nodes as children, as shown in Figure 2, with the one with lower frequency as left child.
B Ni.left = lowest frequency node
B Ni.right = the other node
B Ni.frequency = sum of the frequencies of its children in this case, 11 + 23 = 34
Any conflict, if they appear, should be resolved using the priority order
N1 < N2 < … < Ni < a < b < … < z < 0 < 1 < … < 9 (2)
PRIORITY ORDERS MENTIONED IN EQUATION 1 AND EQUATION 2 ARE VERY IMPORTANT FOR THE OUT-
Figure 1: Linked List of symbols
Step 3
Add the supernode to the list of nodes (maintaining sorted order). Check Figure 3
Step 4
Repeat steps 2 and 3 until there is only a single node in the list and the final structure thus obtained is a tree (Figure 4). We will call this the Huffman tree. Note that the single node left in the linked list is the root of this tree.
Step 5
Obtaining the Huffman Codes: Once the Huffman tree is built, you can obtain the Huffman code for each symbol based on the path from the root to the leaf node holding that symbol. If we denote a path (of length k) by v0v1…vk (root:v0, leaf:vk), the corresponding Huffman code for vk.symbol will be 1r(0,1), 1r(1,2), …, 1r(k−1,k). 1r(i,i+1) is the indicator function for “whether vi+1 is a right child of vi”. For example, in Figure 4 Huffman code for ‘t’ will be ‘00’,
Figure 3: Adding supernodes to the list
Figure 2: Merging nodes to create supernodes
Figure 4: Huffman Tree
but it will be ‘010’ for ‘o’. For finding all such paths use TreeTraverse on the Huffman tree and obtain the Huffman codes on-the-fly.
Algorithm 1: TreeTraverse
Outputs: As output of Part II, you should append the following lines to your output.txt file (created in Part I).
B Print the preorder traversal of the Huffman Tree along with the frequency value for each node. e.g. for the
Huffman Tree given in Fig 4, output should be,
N5(152),N3(61),t(27),N1(34),o(11),i(23),N4(91),e(40),N2(51),a(25),s(26)
B The next 36 lines should contain the Huffman code or prefix code for all alphanumeric characters in alphabetic order, in the following format. During initialization of the linked list use count of 0 for symbols unseen in log.txt.
a 0000 b 01000 c 00100 d 10111 e 0110 … … y 11000 z 01010
0 110100100 1 110100111111 …
9 110100101
Part III – Encoding
** BONUS AND EVALUATED ONLY WHEN YOU ARE AT A GRADE BOUNDARY.
Using the prefix codes found above for each alphanumeric character, encoding new messages is pretty easy. From step 5 in Part I, you should have an array H of strings containing the Huffman codes for each alphanumeric characters. For any given new word(message), you need to read it char-by-char (symbols) and encode each to them in left to right order. To test out your implementation, we will need you to encode the sample messages from the encode.txt file. Output the encoded messages to output.txt as usual, with one encoded message per line.
Notation: In Algorithm 2 and 3, H[c] denotes the Huffman code for character or symbol c.
Algorithm 2: Encode a message
Input: H: Huffman Codes, encodedMessage
Output: R: Encoded Message R = “”; for c in encodedMessage do R = concat(R, H[c])
// Decode function template
void encode(char* message, char* encodedMessage){
// YOUR LOGIC HERE
// Output is saved to encodedMessage
}
Part IV – Decoding
Although we have this great and efficient encoder built, we still need to be able to decode received messages also. That’s where the decoder comes into place. Given an encoded message of 0’s and 1’s, the decoder will convert it back to the original message. Since the encoding scheme is a prefix code, we can start reading from left to right and apply a greedy matching scheme for decoding symbols. Same as before, output the decoded messages in the output.txt.
Algorithm 3: Decode an encoded message
Input: S: Symbol Alphabet, H: Huffman Codes, encodedMessage Output: R: Decoded message
// Decoder function template void decode(char* encodedMessage,
// YOUR LOGIC HERE
} char* decodedOutput){
Example
A complete set of sample input and output files is made available through moodle, please check attachments.
FILE: log.txt
16 the quick brown fox jumps over a lazy dog the five boxing wizards jump quickly ten10
FILE: encode.txt
3 zombie without legs
FILE: decode.txt
3
001011110000110001110011
1011110000101110010110111
0101011000010110111100000010110111
FILE: output.txt
a=3,b=2,c=2,d=2,e=5,f=2,g=2,h=2,i=5,j=2,k=2,l=2,m=2,n=3,o=5,p=2,q=2,r=3,s=2,t=3,u=4,v=2,w=2, x=2,y=2,z=2,0=1,1=1,2=0,3=0,4=0,5=0,6=0,7=0,8=0,9=0
N35(69),N33(31),N29(15),N21(7),t(3),N10(4),N9(2),N8(1),N7(0),N6(0),N5(0),N4(0),N3(0),N2(0), N1(0),2(0),3(0),4(0),5(0),6(0),7(0),8(0),9(0),0(1),1(1),b(2),N22(8),N11(4),c(2),d(2),N12(4), f(2),g(2),N30(16),N23(8),N13(4),h(2),j(2),N14(4),k(2),l(2),N24(8),N15(4),m(2),p(2),N16(4), q(2),s(2),N34(38),N31(17),N25(8),N17(4),v(2),w(2),N18(4),x(2),y(2),N26(9),u(4),N19(5),z(2), a(3),N32(21),N27(10),e(5),i(5),N28(11),o(5),N20(6),n(3),r(3) a 10111 b 00011 c 00100 d 00101 e 1100 f 00110 g 00111 h 01000 i 1101 j 01001 k 01010 l 01011 m 01100 n 11110 o 1110 p 01101 q 01110 r 11111 s 01111 t 0000 u 1010 v 10000 w 10001 x 10010 y 10011 z 10110
0 0001001
1 000101
2 00010000000000
3 00010000000001
4 0001000000001
5 000100000001
6 00010000001
7 0001000001
8 000100001
9 00010001
101101110011000001111011100
100011101000001000111010100000 0101111000011101111 dobby avada kedavda
