|
| 1 | +package com.thealgorithms.compression; |
| 2 | + |
| 3 | +import java.math.BigDecimal; |
| 4 | +import java.math.MathContext; |
| 5 | +import java.util.ArrayList; |
| 6 | +import java.util.Collections; |
| 7 | +import java.util.HashMap; |
| 8 | +import java.util.List; |
| 9 | +import java.util.Map; |
| 10 | + |
| 11 | +/** |
| 12 | + * An implementation of the Arithmetic Coding algorithm. |
| 13 | + * |
| 14 | + * <p> |
| 15 | + * Arithmetic coding is a form of entropy encoding used in lossless data |
| 16 | + * compression. It encodes an entire message into a single number, a fraction n |
| 17 | + * where (0.0 <= n < 1.0). Unlike Huffman coding, which assigns a specific |
| 18 | + * bit sequence to each symbol, arithmetic coding represents the message as a |
| 19 | + * sub-interval of the [0, 1) interval. |
| 20 | + * </p> |
| 21 | + * |
| 22 | + * <p> |
| 23 | + * This implementation uses BigDecimal for precision to handle the shrinking |
| 24 | + * intervals, making it suitable for educational purposes to demonstrate the |
| 25 | + * core logic. |
| 26 | + * </p> |
| 27 | + * |
| 28 | + * <p> |
| 29 | + * Time Complexity: O(n*m) for compression and decompression where n is the |
| 30 | + * length of the input and m is the number of unique symbols, due to the need |
| 31 | + * to calculate symbol probabilities. |
| 32 | + * </p> |
| 33 | + * |
| 34 | + * <p> |
| 35 | + * References: |
| 36 | + * <ul> |
| 37 | + * <li><a href="https://en.wikipedia.org/wiki/Arithmetic_coding">Wikipedia: |
| 38 | + * Arithmetic coding</a></li> |
| 39 | + * </ul> |
| 40 | + * </p> |
| 41 | + */ |
| 42 | +public final class ArithmeticCoding { |
| 43 | + |
| 44 | + private ArithmeticCoding() { |
| 45 | + } |
| 46 | + |
| 47 | + /** |
| 48 | + * Compresses a string using the Arithmetic Coding algorithm. |
| 49 | + * |
| 50 | + * @param uncompressed The string to be compressed. |
| 51 | + * @return The compressed representation as a BigDecimal number. |
| 52 | + * @throws IllegalArgumentException if the input string is null or empty. |
| 53 | + */ |
| 54 | + public static BigDecimal compress(String uncompressed) { |
| 55 | + if (uncompressed == null || uncompressed.isEmpty()) { |
| 56 | + throw new IllegalArgumentException("Input string cannot be null or empty."); |
| 57 | + } |
| 58 | + |
| 59 | + Map<Character, Symbol> probabilityTable = calculateProbabilities(uncompressed); |
| 60 | + |
| 61 | + BigDecimal low = BigDecimal.ZERO; |
| 62 | + BigDecimal high = BigDecimal.ONE; |
| 63 | + |
| 64 | + for (char symbol : uncompressed.toCharArray()) { |
| 65 | + BigDecimal range = high.subtract(low); |
| 66 | + Symbol sym = probabilityTable.get(symbol); |
| 67 | + |
| 68 | + high = low.add(range.multiply(sym.high())); |
| 69 | + low = low.add(range.multiply(sym.low())); |
| 70 | + } |
| 71 | + |
| 72 | + return low; // Return the lower bound of the final interval |
| 73 | + } |
| 74 | + |
| 75 | + /** |
| 76 | + * Decompresses a BigDecimal number back into the original string. |
| 77 | + * |
| 78 | + * @param compressed The compressed BigDecimal number. |
| 79 | + * @param length The length of the original uncompressed string. |
| 80 | + * @param probabilityTable The probability table used during compression. |
| 81 | + * @return The original, uncompressed string. |
| 82 | + */ |
| 83 | + public static String decompress(BigDecimal compressed, int length, Map<Character, Symbol> probabilityTable) { |
| 84 | + StringBuilder decompressed = new StringBuilder(); |
| 85 | + |
| 86 | + // Create a sorted list of symbols for deterministic decompression, matching the |
| 87 | + // order used in calculateProbabilities |
| 88 | + List<Map.Entry<Character, Symbol>> sortedSymbols = new ArrayList<>(probabilityTable.entrySet()); |
| 89 | + sortedSymbols.sort(Map.Entry.comparingByKey()); |
| 90 | + |
| 91 | + BigDecimal low = BigDecimal.ZERO; |
| 92 | + BigDecimal high = BigDecimal.ONE; |
| 93 | + |
| 94 | + for (int i = 0; i < length; i++) { |
| 95 | + BigDecimal range = high.subtract(low); |
| 96 | + |
| 97 | + // Find which symbol the compressed value falls into |
| 98 | + for (Map.Entry<Character, Symbol> entry : sortedSymbols) { |
| 99 | + Symbol sym = entry.getValue(); |
| 100 | + |
| 101 | + // Calculate the actual range for this symbol in the current interval |
| 102 | + BigDecimal symLow = low.add(range.multiply(sym.low())); |
| 103 | + BigDecimal symHigh = low.add(range.multiply(sym.high())); |
| 104 | + |
| 105 | + // Check if the compressed value falls within this symbol's range |
| 106 | + if (compressed.compareTo(symLow) >= 0 && compressed.compareTo(symHigh) < 0) { |
| 107 | + decompressed.append(entry.getKey()); |
| 108 | + |
| 109 | + // Update the interval for the next iteration |
| 110 | + low = symLow; |
| 111 | + high = symHigh; |
| 112 | + break; |
| 113 | + } |
| 114 | + } |
| 115 | + } |
| 116 | + |
| 117 | + return decompressed.toString(); |
| 118 | + } |
| 119 | + |
| 120 | + /** |
| 121 | + * Calculates the frequency and probability range for each character in the |
| 122 | + * input string in a deterministic order. |
| 123 | + * |
| 124 | + * @param text The input string. |
| 125 | + * @return A map from each character to a Symbol object containing its |
| 126 | + * probability range. |
| 127 | + */ |
| 128 | + public static Map<Character, Symbol> calculateProbabilities(String text) { |
| 129 | + Map<Character, Integer> frequencies = new HashMap<>(); |
| 130 | + for (char c : text.toCharArray()) { |
| 131 | + frequencies.put(c, frequencies.getOrDefault(c, 0) + 1); |
| 132 | + } |
| 133 | + |
| 134 | + // Sort the characters to ensure a deterministic order for the probability table |
| 135 | + List<Character> sortedKeys = new ArrayList<>(frequencies.keySet()); |
| 136 | + Collections.sort(sortedKeys); |
| 137 | + |
| 138 | + Map<Character, Symbol> probabilityTable = new HashMap<>(); |
| 139 | + BigDecimal currentLow = BigDecimal.ZERO; |
| 140 | + int total = text.length(); |
| 141 | + |
| 142 | + for (char symbol : sortedKeys) { |
| 143 | + BigDecimal probability = BigDecimal.valueOf(frequencies.get(symbol)).divide(BigDecimal.valueOf(total), MathContext.DECIMAL128); |
| 144 | + BigDecimal high = currentLow.add(probability); |
| 145 | + probabilityTable.put(symbol, new Symbol(currentLow, high)); |
| 146 | + currentLow = high; |
| 147 | + } |
| 148 | + |
| 149 | + return probabilityTable; |
| 150 | + } |
| 151 | + |
| 152 | + /** |
| 153 | + * Helper class to store the probability range [low, high) for a symbol. |
| 154 | + */ |
| 155 | + public record Symbol(BigDecimal low, BigDecimal high) { |
| 156 | + } |
| 157 | +} |
0 commit comments