38021d231d
Multi-pass resolution now prioritizes filename stem (strongest) over alias over title. Removes path-based matching (e.path.endsWith). Legacy path-style targets like "person/alice" still work by extracting the last segment "alice" and matching by filename/title. This matches the flat vault convention where filename IS the note's identity — filename stem is always slugify(title). Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
61 lines
2.5 KiB
TypeScript
61 lines
2.5 KiB
TypeScript
/** Utility functions for parsing wikilink syntax: [[target|display]] */
|
|
|
|
import type { VaultEntry } from '../types'
|
|
|
|
/** Extracts the target path from a wikilink reference (strips [[ ]] and display text). */
|
|
export function wikilinkTarget(ref: string): string {
|
|
const inner = ref.replace(/^\[\[|\]\]$/g, '')
|
|
const pipeIdx = inner.indexOf('|')
|
|
return pipeIdx !== -1 ? inner.slice(0, pipeIdx) : inner
|
|
}
|
|
|
|
/** Extracts the display label from a wikilink reference. Falls back to humanised path stem. */
|
|
export function wikilinkDisplay(ref: string): string {
|
|
const inner = ref.replace(/^\[\[|\]\]$/g, '')
|
|
const pipeIdx = inner.indexOf('|')
|
|
if (pipeIdx !== -1) return inner.slice(pipeIdx + 1)
|
|
const last = inner.split('/').pop() ?? inner
|
|
return last.replace(/-/g, ' ').replace(/\b\w/g, (c) => c.toUpperCase())
|
|
}
|
|
|
|
/**
|
|
* Unified wikilink resolution: find the VaultEntry matching a wikilink target.
|
|
* Handles pipe syntax, case-insensitive matching.
|
|
* Resolution order (multi-pass, global priority):
|
|
* 1. Filename stem match (strongest — filename IS the identity in flat vault)
|
|
* 2. Alias match
|
|
* 3. Exact title match
|
|
* 4. Humanized title match (kebab-case → words)
|
|
* No path-based matching — flat vault uses title/filename only.
|
|
* Legacy path-style targets like "person/alice" are handled by extracting the last segment.
|
|
*/
|
|
export function resolveEntry(entries: VaultEntry[], rawTarget: string): VaultEntry | undefined {
|
|
const key = rawTarget.includes('|') ? rawTarget.split('|')[0] : rawTarget
|
|
const keyLower = key.toLowerCase()
|
|
// For legacy path-style targets like "person/alice", extract just the last segment
|
|
const lastSegment = key.includes('/') ? (key.split('/').pop() ?? key) : key
|
|
const lastSegmentLower = lastSegment.toLowerCase()
|
|
const asWords = lastSegmentLower.replace(/-/g, ' ')
|
|
|
|
// Pass 1: filename stem (strongest match — filename IS identity in flat vault)
|
|
for (const e of entries) {
|
|
const stem = e.filename.replace(/\.md$/, '').toLowerCase()
|
|
if (stem === keyLower || stem === lastSegmentLower) return e
|
|
}
|
|
// Pass 2: alias
|
|
for (const e of entries) {
|
|
if (e.aliases.some(a => a.toLowerCase() === keyLower)) return e
|
|
}
|
|
// Pass 3: exact title
|
|
for (const e of entries) {
|
|
if (e.title.toLowerCase() === keyLower || e.title.toLowerCase() === lastSegmentLower) return e
|
|
}
|
|
// Pass 4: humanized title (kebab-case → words)
|
|
if (asWords !== keyLower) {
|
|
for (const e of entries) {
|
|
if (e.title.toLowerCase() === asWords) return e
|
|
}
|
|
}
|
|
return undefined
|
|
}
|