|
| 1 | +/* |
| 2 | + * Copyright 2021 java-diff-utils. |
| 3 | + * |
| 4 | + * Licensed under the Apache License, Version 2.0 (the "License"); |
| 5 | + * you may not use this file except in compliance with the License. |
| 6 | + * You may obtain a copy of the License at |
| 7 | + * |
| 8 | + * http://www.apache.org/licenses/LICENSE-2.0 |
| 9 | + * |
| 10 | + * Unless required by applicable law or agreed to in writing, software |
| 11 | + * distributed under the License is distributed on an "AS IS" BASIS, |
| 12 | + * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
| 13 | + * See the License for the specific language governing permissions and |
| 14 | + * limitations under the License. |
| 15 | + */ |
| 16 | +package com.github.difflib.algorithm.myers; |
| 17 | + |
| 18 | +import com.github.difflib.algorithm.Change; |
| 19 | +import com.github.difflib.algorithm.DiffAlgorithmI; |
| 20 | +import com.github.difflib.algorithm.DiffAlgorithmListener; |
| 21 | +import com.github.difflib.patch.DeltaType; |
| 22 | +import java.util.ArrayList; |
| 23 | +import java.util.List; |
| 24 | +import java.util.Objects; |
| 25 | +import java.util.function.BiPredicate; |
| 26 | + |
| 27 | +/** |
| 28 | + * |
| 29 | + * @author tw |
| 30 | + */ |
| 31 | +public class MeyersDiffWithLinearSpace<T> implements DiffAlgorithmI<T> { |
| 32 | + |
| 33 | + private final BiPredicate<T, T> equalizer; |
| 34 | + |
| 35 | + public MeyersDiffWithLinearSpace() { |
| 36 | + equalizer = Object::equals; |
| 37 | + } |
| 38 | + |
| 39 | + public MeyersDiffWithLinearSpace(final BiPredicate<T, T> equalizer) { |
| 40 | + Objects.requireNonNull(equalizer, "equalizer must not be null"); |
| 41 | + this.equalizer = equalizer; |
| 42 | + } |
| 43 | + |
| 44 | + @Override |
| 45 | + public List<Change> computeDiff(List<T> source, List<T> target, DiffAlgorithmListener progress) { |
| 46 | + DiffData data = new DiffData(source, target); |
| 47 | + //shouldn't it be source.size() - 1? |
| 48 | + buildScript(data, 0, source.size(), 0, target.size()); |
| 49 | + return data.script; |
| 50 | + } |
| 51 | + |
| 52 | + private void buildScript(DiffData data, int start1, int end1, int start2, int end2) { |
| 53 | + final Snake middle = getMiddleSnake(data, start1, end1, start2, end2); |
| 54 | + if (middle == null |
| 55 | + || middle.start == end1 && middle.diag == end1 - end2 |
| 56 | + || middle.end == start1 && middle.diag == start1 - start2) { |
| 57 | + int i = start1; |
| 58 | + int j = start2; |
| 59 | + while (i < end1 || j < end2) { |
| 60 | + if (i < end1 && j < end2 && equalizer.test(data.source.get(i), data.target.get(j))) { |
| 61 | + //script.append(new KeepCommand<>(left.charAt(i))); |
| 62 | + ++i; |
| 63 | + ++j; |
| 64 | + } else { |
| 65 | + //TODO: compress these commands. |
| 66 | + if (end1 - start1 > end2 - start2) { |
| 67 | + //script.append(new DeleteCommand<>(left.charAt(i))); |
| 68 | + data.script.add(new Change(DeltaType.DELETE, i, i + 1, j, j)); |
| 69 | + ++i; |
| 70 | + } else { |
| 71 | + //script.append(new InsertCommand<>(right.charAt(j))); |
| 72 | + data.script.add(new Change(DeltaType.INSERT, i, i, j, j + 1)); |
| 73 | + ++j; |
| 74 | + } |
| 75 | + } |
| 76 | + } |
| 77 | + } else { |
| 78 | + buildScript(data, start1, middle.start, start2, middle.start - middle.diag); |
| 79 | +// for (int i = middle.getStart(); i < middle.getEnd(); ++i) { |
| 80 | +// script.append(new KeepCommand<>(left.charAt(i))); |
| 81 | +// } |
| 82 | + buildScript(data, middle.end, end1, middle.end - middle.diag, end2); |
| 83 | + } |
| 84 | + } |
| 85 | + |
| 86 | + private Snake getMiddleSnake(DiffData data, int start1, int end1, int start2, int end2) { |
| 87 | + final int m = end1 - start1; |
| 88 | + final int n = end2 - start2; |
| 89 | + if (m == 0 || n == 0) { |
| 90 | + return null; |
| 91 | + } |
| 92 | + |
| 93 | + final int delta = m - n; |
| 94 | + final int sum = n + m; |
| 95 | + final int offset = (sum % 2 == 0 ? sum : sum + 1) / 2; |
| 96 | + data.vDown[1 + offset] = start1; |
| 97 | + data.vUp[1 + offset] = end1 + 1; |
| 98 | + |
| 99 | + for (int d = 0; d <= offset; ++d) { |
| 100 | + // Down |
| 101 | + for (int k = -d; k <= d; k += 2) { |
| 102 | + // First step |
| 103 | + |
| 104 | + final int i = k + offset; |
| 105 | + if (k == -d || k != d && data.vDown[i - 1] < data.vDown[i + 1]) { |
| 106 | + data.vDown[i] = data.vDown[i + 1]; |
| 107 | + } else { |
| 108 | + data.vDown[i] = data.vDown[i - 1] + 1; |
| 109 | + } |
| 110 | + |
| 111 | + int x = data.vDown[i]; |
| 112 | + int y = x - start1 + start2 - k; |
| 113 | + |
| 114 | + while (x < end1 && y < end2 && equalizer.test(data.source.get(x), data.target.get(y))) { |
| 115 | + data.vDown[i] = ++x; |
| 116 | + ++y; |
| 117 | + } |
| 118 | + // Second step |
| 119 | + if (delta % 2 != 0 && delta - d <= k && k <= delta + d) { |
| 120 | + if (data.vUp[i - delta] <= data.vDown[i]) { |
| 121 | + return buildSnake(data, data.vUp[i - delta], k + start1 - start2, end1, end2); |
| 122 | + } |
| 123 | + } |
| 124 | + } |
| 125 | + |
| 126 | + // Up |
| 127 | + for (int k = delta - d; k <= delta + d; k += 2) { |
| 128 | + // First step |
| 129 | + final int i = k + offset - delta; |
| 130 | + if (k == delta - d |
| 131 | + || k != delta + d && data.vUp[i + 1] <= data.vUp[i - 1]) { |
| 132 | + data.vUp[i] = data.vUp[i + 1] - 1; |
| 133 | + } else { |
| 134 | + data.vUp[i] = data.vUp[i - 1]; |
| 135 | + } |
| 136 | + |
| 137 | + int x = data.vUp[i] - 1; |
| 138 | + int y = x - start1 + start2 - k; |
| 139 | + while (x >= start1 && y >= start2 && equalizer.test(data.source.get(x), data.target.get(y))) { |
| 140 | + data.vUp[i] = x--; |
| 141 | + y--; |
| 142 | + } |
| 143 | + // Second step |
| 144 | + if (delta % 2 == 0 && -d <= k && k <= d) { |
| 145 | + if (data.vUp[i] <= data.vDown[i + delta]) { |
| 146 | + return buildSnake(data, data.vUp[i], k + start1 - start2, end1, end2); |
| 147 | + } |
| 148 | + } |
| 149 | + } |
| 150 | + } |
| 151 | + |
| 152 | + // According to Myers, this cannot happen |
| 153 | + throw new IllegalStateException("could not find a diff path"); |
| 154 | + } |
| 155 | + |
| 156 | + private Snake buildSnake(DiffData data, final int start, final int diag, final int end1, final int end2) { |
| 157 | + int end = start; |
| 158 | + while (end - diag < end2 && end < end1 && equalizer.test(data.source.get(end), data.target.get(end - diag))) { |
| 159 | + ++end; |
| 160 | + } |
| 161 | + return new Snake(start, end, diag); |
| 162 | + } |
| 163 | + |
| 164 | + class DiffData { |
| 165 | + |
| 166 | + final int size; |
| 167 | + final int[] vDown; |
| 168 | + final int[] vUp; |
| 169 | + final List<Change> script; |
| 170 | + final List<T> source; |
| 171 | + final List<T> target; |
| 172 | + |
| 173 | + public DiffData(List<T> source, List<T> target) { |
| 174 | + this.source = source; |
| 175 | + this.target = target; |
| 176 | + size = source.size() + target.size() + 2; |
| 177 | + vDown = new int[size]; |
| 178 | + vUp = new int[size]; |
| 179 | + script = new ArrayList<>(); |
| 180 | + } |
| 181 | + } |
| 182 | + |
| 183 | + class Snake { |
| 184 | + |
| 185 | + final int start; |
| 186 | + final int end; |
| 187 | + final int diag; |
| 188 | + |
| 189 | + public Snake(final int start, final int end, final int diag) { |
| 190 | + this.start = start; |
| 191 | + this.end = end; |
| 192 | + this.diag = diag; |
| 193 | + } |
| 194 | + } |
| 195 | +} |
0 commit comments