Compresses the input text using the LZ77 algorithm. @param text The input string to compress. Must not be null. @param windowSize The size of the sliding window (search buffer). Must be positive. @param lookaheadBufferSize The size of the lookahead buffer. Must be positive. @return A list of {@link
(String text, int windowSize, int lookaheadBufferSize)
| 52 | * @throws IllegalArgumentException if windowSize or lookaheadBufferSize are not positive. |
| 53 | */ |
| 54 | public static List<Token> compress(String text, int windowSize, int lookaheadBufferSize) { |
| 55 | if (text == null) { |
| 56 | return new ArrayList<>(); |
| 57 | } |
| 58 | if (windowSize <= 0 || lookaheadBufferSize <= 0) { |
| 59 | throw new IllegalArgumentException("Window size and lookahead buffer size must be positive."); |
| 60 | } |
| 61 | |
| 62 | List<Token> compressedOutput = new ArrayList<>(); |
| 63 | int currentPosition = 0; |
| 64 | |
| 65 | while (currentPosition < text.length()) { |
| 66 | int bestMatchDistance = 0; |
| 67 | int bestMatchLength = 0; |
| 68 | |
| 69 | // Define the start of the search window |
| 70 | int searchBufferStart = Math.max(0, currentPosition - windowSize); |
| 71 | // Define the end of the lookahead buffer (don't go past text length) |
| 72 | int lookaheadEnd = Math.min(currentPosition + lookaheadBufferSize, text.length()); |
| 73 | |
| 74 | // Search for the longest match in the window |
| 75 | for (int i = searchBufferStart; i < currentPosition; i++) { |
| 76 | int currentMatchLength = 0; |
| 77 | |
| 78 | // Check how far the match extends into the lookahead buffer |
| 79 | // This allows for overlapping matches (e.g., "aaa" can match with offset 1) |
| 80 | while (currentPosition + currentMatchLength < lookaheadEnd) { |
| 81 | int sourceIndex = i + currentMatchLength; |
| 82 | |
| 83 | // Handle overlapping matches (run-length encoding within LZ77) |
| 84 | // When we've matched beyond our starting position, wrap around using modulo |
| 85 | if (sourceIndex >= currentPosition) { |
| 86 | int offset = currentPosition - i; |
| 87 | sourceIndex = i + (currentMatchLength % offset); |
| 88 | } |
| 89 | |
| 90 | if (text.charAt(sourceIndex) == text.charAt(currentPosition + currentMatchLength)) { |
| 91 | currentMatchLength++; |
| 92 | } else { |
| 93 | break; |
| 94 | } |
| 95 | } |
| 96 | |
| 97 | // If this match is longer than the best found so far |
| 98 | if (currentMatchLength > bestMatchLength) { |
| 99 | bestMatchLength = currentMatchLength; |
| 100 | bestMatchDistance = currentPosition - i; // Calculate offset from current position |
| 101 | } |
| 102 | } |
| 103 | |
| 104 | char nextChar; |
| 105 | if (currentPosition + bestMatchLength < text.length()) { |
| 106 | nextChar = text.charAt(currentPosition + bestMatchLength); |
| 107 | } else { |
| 108 | nextChar = END_OF_STREAM; |
| 109 | } |
| 110 | |
| 111 | // Add the token to the output |