How do I write Huffman encoded characters to a file?

0
votes

I am trying to read the contents of file and then write them to a different file with Huffman coding. So I have created a Huffman Tree, with each node containing the character, frequency of the character, and a binary string representing the frequency. What I am having trouble understanding is writing the Huffman coded characters to a file.

I tried writing the binary string to the file but realized that it was just writing string and not actually the coded data. So I then converted the binary strings to bytes and wrote the bytes to the file but that would just give me a blank file the same size as the original. I feel like I am missing something when it comes to actually writing the file.

Edit: After taking a look back at my code I realized that my Tree wasn't completely correct and I'm now able to (I think) combine bit string together to make a byte array that I can write to a file (code updated to reflect that). For my test case I am reading in the text AAA_BB_C but when I look at the file the output is <0x1e>. I'm not sure what this means. I was expecting the same output of the original file, just a smaller size.

public static void writeFile(HuffTree tree) {
    String bin = ""; // String of entire binary code
    int spot = 0; // Spot in array
    byte[] bytes = new byte[256]; // byte array
    try {
        FloatileWriter writer = new FileWriter("test(encoded).txt");

            // Gets Binary String of each Character in the file
            for(int i = 0; i < fileText.length(); i++) {
                bin += tree.findDataBinary(fileText.charAt(i));
            }

            // Takes each bit and adds to byte array 
            System.out.println(bin);
            while(bin.length() > 7) {
                String temp = bin.substring(0, 7);
                bin = bin.substring(7, bin.length());
                bytes[spot] = Byte.parseByte(temp, 2);
                spot++;
            }

            // Writes bytes to file
            for(int i = 0; i <= spot; i++) {
                writer.write(bytes[i]);
            }
            writer.close();
    } catch(IOException e) {
        System.out.println("IOException!");
    }
}
1
javahuffman-code
What is fileText? --- What is the point of map? --- byte is a primitive and doesn't have any methods, so what did you expect tempByte.array() to do? Create a 1-byte array? --- Your code seems to just map each character with a 1-byte value, so not much data compression going on. Seems you missed the entire point of Huffman. You sure you understand how Huffman works? It works with bits. so you need to do bit-manipulation. - Andreas
fileText is the the text input from the file. This I saved when reading it. I am having trouble conceptually understanding where the encoding actually happens. Once I have the correct binary values associated with each character I'm unsure what to do. The map was my attempt to pair the binary value associated with the character (in byte form) and then write the bytes into the file with the map. - billybob643
What is the return value from findDataBinary()? Presumably it's a String, since that's what parseByte() requires, but is it a bit-string like "10011"? If so, then calling parseByte() to turn it into the decimal number 10011 seems meaningless. Wouldn't you at the very least want to parse it as binary, e.g. base-2? --- And let's say the first letter maps to 111 and the second letter maps to 10011, that would mean that they combined maps to 11110011, which is 8 bits and hence a single byte, compressing 2 letters into 1 byte, but you do no merging of bits. Step back and try again - Andreas
Yes findDataBinary() returns a bit-string of a character. My byte array should also work as well now. But I'm still stuck. - billybob643
"I was expecting the same output of the original file, just a smaller size" Why on earth would you expect a text file being compression into a binary file to be the "same output"? - Andreas

1 Answers