Press n or j to go to the next uncovered block, b, p or k for the previous block.
| 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 | 9x 47x 47x 22x 22x 22x 2x 20x 20x 4x 16x 21x 16x 3x 3x 13x 8x 8x 8x 8x 5x 10x 44x 76x 76x 20x 20x 20x 20x 20x 50x 50x 50x 50x 21x 21x 21x 21x 20x 50x 32x 32x 31x 31x 31x 31x 31x 341x 31x 357x 31x 310x 3432x 3432x 31x 8x 8x 8x 8x | import { ContentMatcher, MatchResult, MatchCandidate } from './base.matcher';
/**
* Normalized matcher that handles whitespace variations.
* Normalizes tabs, trailing spaces, and multiple spaces before comparing.
*/
export class NormalizedMatcher implements ContentMatcher {
readonly name = 'NormalizedMatcher';
/**
* Confidence threshold for accepting a normalized match.
* Can be configured via environment variable.
*/
private readonly confidenceThreshold: number = parseFloat(
process.env.NORMALIZED_MATCH_THRESHOLD || '0.95',
);
match(searchBlock: string, content: string): MatchResult {
const normalizedSearch = this.normalizeForComparison(searchBlock);
const normalizedContent = this.normalizeForComparison(content);
// If normalized search is empty, fail
if (!normalizedSearch.trim()) {
return {
found: false,
unique: false,
confidence: 0,
candidates: [],
};
}
// Find all candidates with similarity scores
const candidates = this.findSimilarCandidates(
normalizedSearch,
normalizedContent,
content,
);
// No candidates found
if (candidates.length === 0) {
return {
found: false,
unique: false,
confidence: 0,
candidates: [],
};
}
// Filter by confidence threshold
const highConfidenceCandidates = candidates.filter(
(c) => c.confidence >= this.confidenceThreshold,
);
if (highConfidenceCandidates.length === 0) {
return {
found: true,
unique: false,
confidence: Math.max(...candidates.map((c) => c.confidence)),
candidates,
};
}
// Check uniqueness
if (highConfidenceCandidates.length === 1) {
const candidate = highConfidenceCandidates[0];
const before = this.extractOriginalBefore(candidate, content);
const after = this.extractOriginalAfter(candidate, content);
return {
found: true,
unique: true,
confidence: candidate.confidence,
before,
after,
};
}
// Multiple high-confidence matches
return {
found: true,
unique: false,
confidence: Math.max(
...highConfidenceCandidates.map((c) => c.confidence),
),
candidates: highConfidenceCandidates,
};
}
/**
* Normalize text for comparison by:
* - Removing leading and trailing whitespace per line
* - Converting tabs to spaces (2 spaces per tab)
* Note: We preserve internal spacing to match string literals and aligned code
*/
private normalizeForComparison(text: string): string {
return text
.split('\n')
.map((line) => line.trim()) // Remove both leading and trailing whitespace
.map((line) => line.replace(/\t/g, ' ')) // Tabs → spaces
.join('\n')
.trim();
}
/**
* Find all candidates similar to the normalized search block.
*/
private findSimilarCandidates(
normalizedSearch: string,
normalizedContent: string,
originalContent: string,
): MatchCandidate[] {
const candidates: MatchCandidate[] = [];
const normalizedLines = normalizedContent.split('\n');
const searchLines = normalizedSearch.split('\n');
const originalLines = originalContent.split('\n');
// Sliding window to find similar blocks
for (let i = 0; i <= normalizedLines.length - searchLines.length; i++) {
const windowLines = normalizedLines.slice(i, i + searchLines.length);
const windowText = windowLines.join('\n');
const similarity = this.calculateSimilarity(normalizedSearch, windowText);
// Only consider candidates with > 50% similarity
if (similarity > 0.5) {
const beforeAnchor = i > 0 ? originalLines[i - 1] : undefined;
const afterAnchor =
i + searchLines.length < originalLines.length
? originalLines[i + searchLines.length]
: undefined;
const originalBlock = originalLines
.slice(i, i + searchLines.length)
.join('\n');
candidates.push({
block: originalBlock,
startLine: i,
endLine: i + searchLines.length - 1,
confidence: similarity,
beforeAnchor,
afterAnchor,
});
}
}
// Sort by confidence descending
return candidates.sort((a, b) => b.confidence - a.confidence);
}
/**
* Calculate similarity using Levenshtein distance.
* Returns a value between 0.0 (no match) and 1.0 (perfect match).
*/
private calculateSimilarity(a: string, b: string): number {
if (a === b) return 1.0;
Iif (a.length === 0) return 0.0;
if (b.length === 0) return 0.0;
const distance = this.levenshteinDistance(a, b);
const maxLen = Math.max(a.length, b.length);
return 1.0 - distance / maxLen;
}
/**
* Calculate Levenshtein distance between two strings.
*/
private levenshteinDistance(a: string, b: string): number {
const matrix: number[][] = [];
for (let i = 0; i <= b.length; i++) {
matrix[i] = [i];
}
for (let j = 0; j <= a.length; j++) {
matrix[0][j] = j;
}
for (let i = 1; i <= b.length; i++) {
for (let j = 1; j <= a.length; j++) {
const cost = b[i - 1] === a[j - 1] ? 0 : 1;
matrix[i][j] = Math.min(
matrix[i - 1][j] + 1, // deletion
matrix[i][j - 1] + 1, // insertion
matrix[i - 1][j - 1] + cost, // substitution
);
}
}
return matrix[b.length][a.length];
}
/**
* Extract the original content before the match.
*/
private extractOriginalBefore(
candidate: MatchCandidate,
content: string,
): string {
const lines = content.split('\n');
return lines.slice(0, candidate.startLine).join('\n');
}
/**
* Extract the original content after the match.
*/
private extractOriginalAfter(
candidate: MatchCandidate,
content: string,
): string {
const lines = content.split('\n');
return lines.slice(candidate.endLine + 1).join('\n');
}
}
|