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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 | 32x 32x 6x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 2x 32x 14x 14x 14x 14x 18x 18x 4x 14x 7x 7x 14x 14x 20x 17x 34x 33x 33x 33x 33x 33x 17x 23x 7x 17x 14x 14x 3x 3x 3x 3x 14x 14x 14x 14x 14x 14x 14x 14x 14x 14x | /**
* @fileoverview Search result formatting utilities for Pagefind.
* These functions transform raw Pagefind API responses into display-ready format.
* @see https://pagefind.app/docs/api/
*/
import { isNumber } from 'lodash';
import type {
PagefindSearchFragment, PagefindSubResult, PagefindSearchAnchor, FormattedSearchResult,
} from './types.js';
/**
* Truncates an excerpt to ensure the <mark> tags are visible.
* - Shows max 15 chars before <mark>
* - Shows all content after <mark> (no limit)
* - Adds ellipsis if prefix doesn't start at word boundary
* - Handles HTML entities in the prefix
*
* @param excerpt - The raw excerpt from Pagefind
* @returns Truncated excerpt with <mark> visible
*/
function truncateExcerptToShowMark(excerpt: string): string {
const markStart = excerpt.indexOf('<mark>');
// No mark found, return as is
if (markStart === -1) return excerpt;
// If mark is at position 0, return as is
if (markStart === 0) return excerpt;
// Get up to 15 chars before <mark>
const prefix = excerpt.substring(0, markStart);
const truncatedPrefix = prefix.slice(-15); // Last 15 chars
// Check if starts at word boundary:
// - Any whitespace (space, tab, newline)
// - Any non-alphanumeric character
const firstChar = truncatedPrefix[0];
const isWordBoundary = /[\s\d_\-.,;:'"()[\]{}|\\/@#$%^&*!~`]/.test(firstChar);
if (!isWordBoundary) {
// Find the first word boundary in the prefix (within 15 chars of end)
const searchArea = prefix.slice(-15);
const wordBoundaryMatch = searchArea.match(/[\s\d_\-.,;:'"()[\]{}|\\/@#$%^&*!~`]/);
if (wordBoundaryMatch) {
const lastBoundaryIndex = prefix.lastIndexOf(wordBoundaryMatch[0], markStart - 1);
if (lastBoundaryIndex !== -1 && lastBoundaryIndex < markStart) {
return `...${prefix.substring(lastBoundaryIndex + 1)}${excerpt.substring(markStart)}`;
}
}
// Fallback: use ellipsis + truncated
return `...${truncatedPrefix}${excerpt.substring(markStart)}`;
}
// Starts at word boundary, no ellipsis needed
return truncatedPrefix + excerpt.substring(markStart);
}
/**
* Merges consecutive <mark> tags into a single <mark> tag.
* e.g., "<mark>making</mark> <mark>the</mark>" becomes "<mark>making the</mark>"
* This ensures that terms grouped together in the excerpt are displayed as a single highlighted segment.
*
* @param excerpt - The excerpt with potential consecutive <mark> tags
* @returns Excerpt with merged <mark> tags
*/
function mergeConsecutiveMarks(excerpt: string): string {
return excerpt.replace(/<\/mark>\s*<mark>/g, ' ');
}
/**
* Parses a single sub-result (heading/section) within a page into a display-ready format.
*
* This function constructs a hierarchical title (breadcrumb) by finding all anchor elements
* that appear before the current sub-result. For example, if the page has:
* - h1: "Installation"
* - h2: "Windows"
* - h3: "Troubleshooting"
*
* And the sub-result is "Troubleshooting", the title becomes "Installation > Windows > Troubleshooting".
*
* @param sub - The sub-result from Pagefind
* @param anchors - All anchor elements on the page
* @param result - The parent Pagefind result
* @returns Formatted search result with hierarchical title
* @see https://pagefind.app/docs/sub-results/
*/
function parseSubResult(
sub: PagefindSubResult,
anchors: PagefindSearchAnchor[],
result: PagefindSearchFragment,
): FormattedSearchResult {
const route = sub?.url || result?.url;
const description = mergeConsecutiveMarks(truncateExcerptToShowMark(sub?.excerpt || result?.excerpt || ''));
const title = sub.title || '';
return {
route,
meta: {
...result.meta,
title,
description,
},
result,
isSubResult: true,
isLastSubResult: false,
};
}
/**
* Formats raw Pagefind search results for display in the UI.
*
* This function performs four key transformations:
*
* 1. **Sort by `balance_score`**: Sorts weighted_locations by their `balance_score` (descending),
* then by `weight` (descending), then position (ascending) as a tie-breaker.
* This prioritizes matches by their scores first, then
* higher-weighted sections (e.g., headings) over body text, then finally their position on the page
*
* 2. **Pick Top Sub-Results**: Iterates through sorted locations and finds
* which sub-results (headings) contain those locations. If multiple
* sub-results contain the same location, keeps the one with more context
* (more locations). Stops after collecting `count` results.
*
* 3. **Re-sort by Document Order**: Resorts the selected sub-results by their
* position in the document, so they appear in natural reading order.
*
* 4. **Deduplicate**: Removes duplicate titles that may arise from overlapping matches.
*
* @param result - Raw Pagefind result from `pagefind.search().results[i].data()`
* @param count - Maximum number of sub-results to return per page (default: 1)
* @returns Array of formatted results ready for display
* @see https://pagefind.app/docs/api-reference
* @see https://pagefind.app/docs/ranking/
* @see https://pagefind.app/docs/sub-results/
*
* @example
* ```typescript
* const search = await pagefind.search("installation");
* const results = await Promise.all(search.results.map(r => r.data()));
* const formatted = results.flatMap(r => formatPagefindResult(r, 2));
* // Returns up to 2 sub-results per page, sorted by relevance
* ```
*/
export function formatPagefindResult(
result: PagefindSearchFragment,
count = 10,
): FormattedSearchResult[] {
const { sub_results: subResults, anchors, weighted_locations: weightedLocations } = result;
// If no sub_results, return the main result as a non-sub-result
if (!subResults || subResults.length === 0) {
return [
{
route: result.url,
meta: {
...result.meta,
title: result.meta.title || '',
description: mergeConsecutiveMarks(truncateExcerptToShowMark(result.excerpt || '')),
},
result,
isSubResult: false,
isLastSubResult: false,
},
];
}
const sortedLocations = [...weightedLocations].sort((a, b) => {
Iif (b.balanced_score === a.balanced_score) {
// If equal balanced_score -> weight -> earlier position in document comes first.
Iif (a.weight === b.weight) {
return a.location - b.location;
}
return b.weight - a.weight;
}
return b.balanced_score - a.balanced_score;
});
// For each location, find matching subresults,
// Then pick the subresult with the top `count` based on weighted locations.
const subs: PagefindSubResult[] = [];
sortedLocations.forEach(({ location }) => {
if (subs.length >= count) return;
// Find sub-results that contain this weighted location
const filterData = subResults.filter((sub: PagefindSubResult) => {
if (sub.title === result.meta.title) return false; // Skip page-level match
const { locations } = sub;
const [min] = locations || [];
Iif (!isNumber(min)) return false;
const max = locations.length === 1 ? Number.POSITIVE_INFINITY : locations[locations.length - 1];
return min <= location && location <= max;
});
// Keep the sub-result with the most locations (most context)
const sub = filterData.reduce<PagefindSubResult | null>((prev, curr) => {
if (!prev) return curr;
return prev.locations.length > curr.locations.length ? prev : curr;
}, null);
// Add only unique sub-results
if (sub && !subs.some(existing => existing.title === sub.title)) {
subs.push(sub);
}
});
// Re-sort by document order (position in page).
subs.sort((a, b) => {
const [minA] = a.locations || [];
const [minB] = b.locations || [];
Iif (minA == null || minB == null) {
return 0;
}
return minA - minB;
});
// Remove duplicate entries that may occur from overlapping matches.
const filterMap = new Map<string, FormattedSearchResult>();
const formattedSubResults = subs
.map((sub: PagefindSubResult) => parseSubResult(sub, anchors, result))
.filter((v: FormattedSearchResult) => {
Iif (filterMap.has(v.meta.title)) {
return false;
}
filterMap.set(v.meta.title, v);
return true;
});
// Mark the last sub-result
formattedSubResults.forEach((sub, index) => {
// eslint-disable-next-line no-param-reassign
sub.isLastSubResult = index === formattedSubResults.length - 1;
});
// Return main result first, then sub-results
const mainResult = [
{
route: result.url,
meta: {
...result.meta,
title: result.meta.title || '',
description: mergeConsecutiveMarks(truncateExcerptToShowMark(result.excerpt || '')),
},
result,
isSubResult: false,
isLastSubResult: false,
},
];
return [...mainResult, ...formattedSubResults];
}
|