diff --git a/docs/superpowers/plans/2026-06-22-architecture-deepening.md b/docs/superpowers/plans/2026-06-22-architecture-deepening.md new file mode 100644 index 0000000..df11142 --- /dev/null +++ b/docs/superpowers/plans/2026-06-22-architecture-deepening.md @@ -0,0 +1,1750 @@ +# Architecture Deepening Implementation Plan + +> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development (recommended) or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (`- [ ]`) syntax for tracking. + +**Goal:** Deepen the architecture of Obsidian Automatic Linker without changing user-visible behavior. + +**Architecture:** Keep Obsidian lifecycle, editor writes, notices, network requests, and command registration in adapter code. Move pure transformation sequencing, candidate scanning, settings projection, Markdown segmentation, and URL formatter orchestration into modules with narrow interfaces and unit tests. Each stage preserves current output strings and leaves the repository in a releasable state. + +**Tech Stack:** TypeScript, Obsidian plugin APIs, Vitest, `npm run test -- --reporter=dot`, `npm run tsc`, `npm run lint`. + +## Global Constraints + +- Do not add new user-facing features. +- Do not change existing settings names, defaults, command ids, or manifest metadata. +- Do not change Obsidian link output formats except where an existing test already expects that behavior. +- Do not replace the current test framework or build tooling. +- Do not introduce a Markdown parser dependency unless a later implementation step proves regex segmentation cannot preserve current behavior. +- Use `git` for version-control operations in this repository. +- Do not access GitHub URLs directly. Use the `gh` CLI for GitHub operations. +- When committing, use English Conventional Commit messages and include a clear description of Why and What. + +--- + +## File Structure + +- Create `src/replace-links/candidate-scanner.ts`: owns candidate occurrence discovery shared by link replacement and AI ambiguity resolution. +- Create `src/replace-links/__tests__/candidate-scanner.test.ts`: tests candidate occurrence discovery without AI or rendering. +- Modify `src/replace-links/replace-links.ts`: imports scanner helpers and keeps rendering replacement output. +- Modify `src/utils/resolve-ambiguities.ts`: builds AI requests from candidate scanner output instead of its own simplified trie scan. +- Create `src/formatting-run.ts`: owns pure formatting sequencing for document bodies and whole Markdown documents. +- Create `src/__tests__/formatting-run.test.ts`: tests formatting sequencing without constructing `AutomaticLinkerPlugin`. +- Modify `src/main.ts`: uses formatting run for file, vault, selection, and AI content transformation. +- Create `src/settings/settings-catalog.ts`: owns settings defaults, metadata, refresh rules, and runtime projections. +- Create `src/settings/__tests__/settings-catalog.test.ts`: verifies catalog coverage and projections. +- Modify `src/settings/settings-info.ts`: re-exports settings type and defaults from the catalog. +- Modify `src/settings/settings.ts`: renders settings from catalog metadata while preserving visible labels and descriptions. +- Create `src/markdown-segments.ts`: owns Markdown prose/protected segment splitting and prose mapping. +- Create `src/__tests__/markdown-segments.test.ts`: verifies exact round-trip and prose-only transforms. +- Modify `src/replace-url-with-title/index.ts`: maps URL title replacement over Markdown prose segments. +- Modify `src/replace-url-with-title/utils/list-up-all-urls.ts`: collects URLs only from Markdown prose segments. +- Modify `src/replace-links/replace-links.ts`: uses Markdown segment module for protected regions. +- Create `src/replace-urls/url-formatting.ts`: owns URL formatter adapter ordering and one text pass. +- Create `src/replace-urls/__tests__/url-formatting.test.ts`: tests orchestration across GitHub, Jira, and Linear. +- Modify `src/replace-urls/github.ts`: remove sibling adapter imports from GitHub adapter. +- Modify `src/replace-urls/replace-urls.ts`: keep as compatibility shim that preserves the existing `replaceURLs` interface for current tests and callers. + +--- + +### Task 1: Candidate Scanning Module + +**Files:** +- Create: `src/replace-links/candidate-scanner.ts` +- Create: `src/replace-links/__tests__/candidate-scanner.test.ts` +- Modify: `src/replace-links/replace-links.ts` +- Modify: `src/utils/resolve-ambiguities.ts` +- Test: `src/replace-links/__tests__/candidate-scanner.test.ts` +- Test: `src/utils/__tests__/resolve-ambiguities.test.ts` +- Test: `src/replace-links/__tests__/ai-disambiguation.test.ts` + +**Interfaces:** +- Consumes: `CandidateData`, `TrieNode`, and `ReplaceLinksSettings`. +- Produces: + +```ts +export type CandidateOccurrenceKind = "unlinked" | "existing-wikilink" + +export interface CandidateOccurrence { + kind: CandidateOccurrenceKind + start: number + end: number + text: string + candidateKey: string + candidateData: CandidateData + isInTable: boolean +} + +export interface ScanCandidateOccurrencesOptions { + text: string + filePath: string + trie: TrieNode + candidateMap: Map + settings?: ReplaceLinksSettings +} + +export const scanCandidateOccurrences: ( + options: ScanCandidateOccurrencesOptions, +) => CandidateOccurrence[] + +export const getOccurrenceContext: ( + text: string, + occurrence: CandidateOccurrence, + maxContext: number, +) => string +``` + +- Later tasks may import `scanCandidateOccurrences` from `src/replace-links/candidate-scanner.ts`; no later task may recreate candidate matching logic. + +- [ ] **Step 1: Write the failing scanner tests** + +Create `src/replace-links/__tests__/candidate-scanner.test.ts`: + +```ts +import { describe, expect, it } from "vitest" +import { buildCandidateTrieForTest } from "./test-helpers" +import { + getOccurrenceContext, + scanCandidateOccurrences, +} from "../candidate-scanner" + +describe("scanCandidateOccurrences", () => { + it("reports ambiguous prose candidates and skips inline code", () => { + const { candidateMap, trie } = buildCandidateTrieForTest({ + files: [ + { path: "work/meeting" }, + { path: "private/meeting" }, + ], + settings: { + scoped: false, + baseDir: undefined, + ignoreCase: true, + }, + }) + + const occurrences = scanCandidateOccurrences({ + text: "`meeting` meeting", + filePath: "notes/today", + trie, + candidateMap, + settings: { ignoreCase: true, proximityBasedLinking: true }, + }) + + expect(occurrences.map(o => ({ + kind: o.kind, + text: o.text, + start: o.start, + end: o.end, + candidates: o.candidateData.candidates.map(c => c.canonical), + }))).toEqual([ + { + kind: "unlinked", + text: "meeting", + start: 10, + end: 17, + candidates: ["work/meeting", "private/meeting"], + }, + ]) + }) + + it("reports existing wikilinks by their display alias", () => { + const { candidateMap, trie } = buildCandidateTrieForTest({ + files: [ + { path: "work/meeting" }, + { path: "private/meeting" }, + ], + settings: { + scoped: false, + baseDir: undefined, + ignoreCase: true, + }, + }) + + const occurrences = scanCandidateOccurrences({ + text: "Check [[private/meeting|meeting]] notes.", + filePath: "notes/today", + trie, + candidateMap, + settings: { ignoreCase: true, proximityBasedLinking: true }, + }) + + expect(occurrences.map(o => ({ + kind: o.kind, + text: o.text, + candidateKey: o.candidateKey, + candidates: o.candidateData.candidates.map(c => c.canonical), + }))).toEqual([ + { + kind: "existing-wikilink", + text: "[[private/meeting|meeting]]", + candidateKey: "meeting", + candidates: ["work/meeting", "private/meeting"], + }, + ]) + }) + + it("filters scoped candidates outside the current namespace", () => { + const { candidateMap, trie } = buildCandidateTrieForTest({ + files: [ + { path: "pages/team-a/internal" }, + { path: "pages/team-b/internal" }, + ], + settings: { + scoped: true, + baseDir: "pages", + ignoreCase: true, + }, + }) + + const occurrences = scanCandidateOccurrences({ + text: "internal", + filePath: "pages/team-a/today", + trie, + candidateMap, + settings: { + baseDir: "pages", + ignoreCase: true, + proximityBasedLinking: true, + }, + }) + + expect(occurrences).toHaveLength(1) + expect(occurrences[0].candidateData.candidates.map(c => c.canonical)).toEqual([ + "pages/team-a/internal", + ]) + }) +}) + +describe("getOccurrenceContext", () => { + it("returns bounded surrounding text for AI requests", () => { + const occurrence = { + kind: "unlinked" as const, + start: 10, + end: 17, + text: "meeting", + candidateKey: "meeting", + candidateData: { candidates: [] }, + isInTable: false, + } + + expect(getOccurrenceContext("before -- meeting -- after", occurrence, 4)).toBe( + " -- meeting -- ", + ) + }) +}) +``` + +- [ ] **Step 2: Run scanner tests to verify RED** + +Run: + +```bash +npx vitest run src/replace-links/__tests__/candidate-scanner.test.ts +``` + +Expected: FAIL because `src/replace-links/candidate-scanner.ts` does not exist. + +- [ ] **Step 3: Implement candidate scanner** + +Create `src/replace-links/candidate-scanner.ts` with the exported interfaces from this task. Move these private helpers from `src/replace-links/replace-links.ts` into the new module, preserving behavior, then import them back into `replace-links.ts`: + +```ts +REGEX_PATTERNS +isWordBoundary +isMonthNote +isProtectedLink +isCjkText +isCjkCandidate +isKoreanText +isSentenceStart +buildFallbackIndex +getCurrentNamespace +normalizeCanonicalPath +extractLinkParts +escapeRegExp +extractFencedCodeBlocks +isSelfLink +shouldSkipCandidate +isMarkdownTableLine +isIndexInsideMarkdownTable +findBestCandidateInSameNamespace +``` + +Implement the exported scanner around those helpers: + +```ts +export const scanCandidateOccurrences = ({ + text, + filePath, + trie, + candidateMap, + settings = {}, +}: ScanCandidateOccurrencesOptions): CandidateOccurrence[] => { + const occurrences: CandidateOccurrence[] = [] + const normalizedText = text.normalize("NFC") + const fallbackIndex = buildFallbackIndex(candidateMap, settings.ignoreCase) + const currentNamespace = getCurrentNamespace(filePath, settings.baseDir) + + collectExistingWikilinks(normalizedText, candidateMap, occurrences) + collectUnlinkedOccurrences({ + text: normalizedText, + filePath, + trie, + candidateMap, + fallbackIndex, + currentNamespace, + settings, + occurrences, + }) + + return occurrences.sort((a, b) => a.start - b.start) +} + +export const getOccurrenceContext = ( + text: string, + occurrence: CandidateOccurrence, + maxContext: number, +): string => { + const start = Math.max(0, occurrence.start - maxContext) + const end = Math.min(text.length, occurrence.end + maxContext) + return text.slice(start, end) +} +``` + +`collectUnlinkedOccurrences` must use the same trie traversal, fallback search, word-boundary, CJK, sentence-case, namespace, date, month-note, and self-link checks that `processStandardText` currently uses. Keep the scanner pure; it must not call a `LinkGenerator` or produce final wiki link strings. + +Modify `src/replace-links/replace-links.ts` to import the moved helpers from `./candidate-scanner` and delete their local duplicate definitions. Keep `replaceLinks` rendering behavior unchanged. + +- [ ] **Step 4: Run scanner tests to verify GREEN** + +Run: + +```bash +npx vitest run src/replace-links/__tests__/candidate-scanner.test.ts +``` + +Expected: PASS. + +- [ ] **Step 5: Route AI ambiguity detection through scanner** + +Modify `src/utils/resolve-ambiguities.ts` so it no longer scans the trie directly: + +```ts +import { CandidateData, TrieNode } from "../trie" +import { AutomaticLinkerSettings } from "../settings/settings-info" +import { resolveAmbiguitiesBatch, AIResolveRequest } from "./ai-client" +import { + getOccurrenceContext, + scanCandidateOccurrences, +} from "../replace-links/candidate-scanner" + +export const resolveAmbiguities = async ( + text: string, + candidateMap: Map, + trie: TrieNode, + settings: AutomaticLinkerSettings, +): Promise> => { + const occurrences = scanCandidateOccurrences({ + text, + filePath: "", + trie, + candidateMap, + settings, + }) + + const requests: AIResolveRequest[] = occurrences + .filter(occurrence => occurrence.candidateData.candidates.length > 1) + .map((occurrence) => ({ + word: occurrence.text, + text: getOccurrenceContext(text, occurrence, settings.aiMaxContext), + candidates: occurrence.candidateData.candidates.map(c => c.canonical), + })) + + const uniqueRequests = Array.from(new Map(requests.map(r => [r.word, r])).values()) + return await resolveAmbiguitiesBatch(settings, uniqueRequests) +} +``` + +For existing wikilinks, `occurrence.text` must be the full wikilink string, because `replaceLinks` already uses full wikilink text as the key for replacement. + +- [ ] **Step 6: Run focused AI and replacement tests** + +Run: + +```bash +npx vitest run src/utils/__tests__/resolve-ambiguities.test.ts src/replace-links/__tests__/ai-disambiguation.test.ts src/replace-links/__tests__/replace-links.basic.test.ts +``` + +Expected: PASS. + +- [ ] **Step 7: Run full verification for the stage** + +Run: + +```bash +npm run test -- --reporter=dot +npm run tsc +npm run lint +``` + +Expected: all commands exit 0. + +- [ ] **Step 8: Commit candidate scanning stage** + +Run: + +```bash +git add src/replace-links/candidate-scanner.ts src/replace-links/__tests__/candidate-scanner.test.ts src/replace-links/replace-links.ts src/utils/resolve-ambiguities.ts +MSG_FILE=$(mktemp) +cat > "$MSG_FILE" <<'EOF' +refactor(replace-links): centralize candidate scanning + +Why: +- Candidate detection was duplicated between link replacement and AI ambiguity resolution, which made matching behavior hard to reason about. +- A single scanner improves locality for namespace, protected-span, case, and CJK matching rules. + +What: +- Add a shared candidate scanner module for unlinked text and existing wikilinks. +- Route AI ambiguity request construction through the scanner. +- Keep link rendering behavior unchanged and covered by existing replacement tests. +EOF +git commit -F "$MSG_FILE" +rm -f "$MSG_FILE" +``` + +--- + +### Task 2: Formatting Run Module + +**Files:** +- Create: `src/formatting-run.ts` +- Create: `src/__tests__/formatting-run.test.ts` +- Modify: `src/main.ts` +- Modify: `src/__tests__/main-url-title-frontmatter.test.ts` +- Test: `src/__tests__/formatting-run.test.ts` +- Test: `src/__tests__/main-url-title-frontmatter.test.ts` +- Test: `src/__tests__/main-link-generator.test.ts` + +**Interfaces:** +- Consumes: `AutomaticLinkerSettings`, `ReplaceLinksSettings`, `TrieNode`, `CandidateData`, `LinkGenerator`, and URL title map. +- Produces: + +```ts +export interface CandidateIndex { + trie: TrieNode + candidateMap: Map +} + +export interface FormattingRunOptions { + content: string + filePath: string + contentStart?: number + frontmatter?: Record + settings: AutomaticLinkerSettings + baseDir?: string + candidateIndex?: CandidateIndex + urlTitleMap?: Map + linkGenerator?: LinkGenerator +} + +export const toReplaceLinksSettings: ( + settings: AutomaticLinkerSettings, + baseDir?: string, +) => ReplaceLinksSettings + +export const formatMarkdownDocument: (options: FormattingRunOptions) => string + +export const formatMarkdownBody: ( + options: Omit & { body: string }, +) => string +``` + +- `formatMarkdownDocument` handles frontmatter splitting. +- `formatMarkdownBody` transforms a body or selection without frontmatter splitting. + +- [ ] **Step 1: Write formatting-run tests** + +Create `src/__tests__/formatting-run.test.ts`: + +```ts +import { describe, expect, it } from "vitest" +import { formatMarkdownBody, formatMarkdownDocument, toReplaceLinksSettings } from "../formatting-run" +import { buildCandidateTrieForTest } from "../replace-links/__tests__/test-helpers" +import { DEFAULT_SETTINGS } from "../settings/settings-info" + +describe("toReplaceLinksSettings", () => { + it("projects only replacement settings and applies baseDir", () => { + expect(toReplaceLinksSettings({ + ...DEFAULT_SETTINGS, + proximityBasedLinking: false, + ignoreDateFormats: false, + ignoreCase: false, + matchSentenceCase: false, + preventSelfLinking: true, + removeAliasInDirs: ["archive"], + ignoreHeadings: true, + ignoreMarkdownTables: true, + }, "pages")).toEqual({ + proximityBasedLinking: false, + baseDir: "pages", + ignoreDateFormats: false, + ignoreCase: false, + matchSentenceCase: false, + preventSelfLinking: true, + removeAliasInDirs: ["archive"], + ignoreHeadings: true, + ignoreMarkdownTables: true, + }) + }) +}) + +describe("formatMarkdownDocument", () => { + it("preserves frontmatter and respects URL title opt-out", () => { + const result = formatMarkdownDocument({ + content: "---\nautomatic-linker-disable-url-title: true\n---\nhttps://example.com", + filePath: "current-file.md", + frontmatter: { "automatic-linker-disable-url-title": true }, + settings: { + ...DEFAULT_SETTINGS, + formatGitHubURLs: false, + formatJiraURLs: false, + formatLinearURLs: false, + replaceUrlWithTitle: true, + }, + urlTitleMap: new Map([["https://example.com", "Example Title"]]), + }) + + expect(result).toBe( + "---\nautomatic-linker-disable-url-title: true\n---\nhttps://example.com", + ) + }) + + it("runs URL formatting, URL titles, and link replacement in order", () => { + const { candidateMap, trie } = buildCandidateTrieForTest({ + files: [{ path: "notes/TypeScript" }], + settings: { + scoped: false, + baseDir: undefined, + ignoreCase: true, + }, + }) + + const result = formatMarkdownDocument({ + content: "Read TypeScript at https://example.com", + filePath: "current-file.md", + settings: { + ...DEFAULT_SETTINGS, + formatGitHubURLs: false, + formatJiraURLs: false, + formatLinearURLs: false, + replaceUrlWithTitle: true, + ignoreCase: true, + }, + candidateIndex: { candidateMap, trie }, + urlTitleMap: new Map([["https://example.com", "Example Title"]]), + }) + + expect(result).toBe("Read [[notes/TypeScript|TypeScript]] at [Example Title](https://example.com)") + }) +}) + +describe("formatMarkdownBody", () => { + it("formats selected body text without frontmatter splitting", () => { + const { candidateMap, trie } = buildCandidateTrieForTest({ + files: [{ path: "notes/TypeScript" }], + settings: { + scoped: false, + baseDir: undefined, + ignoreCase: true, + }, + }) + + const result = formatMarkdownBody({ + body: "TypeScript", + filePath: "current-file.md", + settings: { + ...DEFAULT_SETTINGS, + formatGitHubURLs: false, + formatJiraURLs: false, + formatLinearURLs: false, + replaceUrlWithTitle: false, + ignoreCase: true, + }, + candidateIndex: { candidateMap, trie }, + }) + + expect(result).toBe("[[notes/TypeScript|TypeScript]]") + }) +}) +``` + +- [ ] **Step 2: Run formatting-run tests to verify RED** + +Run: + +```bash +npx vitest run src/__tests__/formatting-run.test.ts +``` + +Expected: FAIL because `src/formatting-run.ts` does not exist. + +- [ ] **Step 3: Implement formatting-run module** + +Create `src/formatting-run.ts`: + +```ts +import { + isUrlTitleReplacementOff, +} from "./frontmatter-utils" +import { replaceLinks, LinkGenerator } from "./replace-links/replace-links" +import { replaceUrlWithTitle } from "./replace-url-with-title" +import { formatGitHubURL } from "./replace-urls/github" +import { formatJiraURL } from "./replace-urls/jira" +import { formatLinearURL } from "./replace-urls/linear" +import { replaceURLs } from "./replace-urls/replace-urls" +import { AutomaticLinkerSettings } from "./settings/settings-info" +import { CandidateData, TrieNode } from "./trie" +import { ReplaceLinksSettings } from "./replace-links/replace-links" + +export interface CandidateIndex { + trie: TrieNode + candidateMap: Map +} + +export interface FormattingRunOptions { + content: string + filePath: string + contentStart?: number + frontmatter?: Record + settings: AutomaticLinkerSettings + baseDir?: string + candidateIndex?: CandidateIndex + urlTitleMap?: Map + linkGenerator?: LinkGenerator +} + +export const toReplaceLinksSettings = ( + settings: AutomaticLinkerSettings, + baseDir?: string, +): ReplaceLinksSettings => ({ + proximityBasedLinking: settings.proximityBasedLinking, + baseDir, + ignoreDateFormats: settings.ignoreDateFormats, + ignoreCase: settings.ignoreCase, + matchSentenceCase: settings.matchSentenceCase, + preventSelfLinking: settings.preventSelfLinking, + removeAliasInDirs: settings.removeAliasInDirs, + ignoreHeadings: settings.ignoreHeadings, + ignoreMarkdownTables: settings.ignoreMarkdownTables, +}) + +export const formatMarkdownBody = ({ + body, + filePath, + frontmatter, + settings, + baseDir, + candidateIndex, + urlTitleMap = new Map(), + linkGenerator, +}: Omit & { body: string }): string => { + let updatedBody = body + + if (settings.formatGitHubURLs) { + updatedBody = replaceURLs(updatedBody, settings, formatGitHubURL) + } + if (settings.formatJiraURLs) { + updatedBody = replaceURLs(updatedBody, settings, formatJiraURL) + } + if (settings.formatLinearURLs) { + updatedBody = replaceURLs(updatedBody, settings, formatLinearURL) + } + if (settings.replaceUrlWithTitle && !isUrlTitleReplacementOff(frontmatter)) { + updatedBody = replaceUrlWithTitle({ body: updatedBody, urlTitleMap }) + } + if (candidateIndex) { + updatedBody = replaceLinks({ + body: updatedBody, + linkResolverContext: { + filePath: filePath.replace(/\.md$/, ""), + trie: candidateIndex.trie, + candidateMap: candidateIndex.candidateMap, + }, + settings: toReplaceLinksSettings(settings, baseDir), + linkGenerator, + }) + } + + return updatedBody +} + +const inferContentStart = (content: string): number => { + const frontmatter = content.match(/^---\r?\n[\s\S]*?\r?\n---\r?\n?/) + return frontmatter?.[0].length ?? 0 +} + +export const formatMarkdownDocument = ({ + content, + contentStart = inferContentStart(content), + ...options +}: FormattingRunOptions): string => { + const frontmatterText = content.slice(0, contentStart) + const body = content.slice(contentStart) + return frontmatterText + formatMarkdownBody({ ...options, body }) +} +``` + +- [ ] **Step 4: Run formatting-run tests to verify GREEN** + +Run: + +```bash +npx vitest run src/__tests__/formatting-run.test.ts +``` + +Expected: PASS. + +- [ ] **Step 5: Route `main.ts` through formatting-run** + +Modify `src/main.ts`: + +```ts +import { + formatMarkdownBody, + formatMarkdownDocument, + toReplaceLinksSettings, +} from "./formatting-run" +``` + +In `modifyLinks`, replace the URL and link transformation sequence with: + +```ts +const baseDir = this.settings.respectNewFileFolderPath ? this.app.vault.getConfig("newFileFolderPath") : undefined +const candidateIndex = this.trie && this.candidateMap + ? { trie: this.trie, candidateMap: this.candidateMap } + : undefined + +return formatMarkdownDocument({ + content: fileContent, + filePath, + contentStart: getFrontMatterInfo(fileContent).contentStart, + frontmatter, + settings: this.settings, + baseDir, + candidateIndex, + urlTitleMap: this.urlTitleMap, + linkGenerator: candidateIndex ? this.createLinkGenerator(filePath) : undefined, +}) +``` + +In `mofifyLinksSelection`, replace the direct `replaceLinks` call with: + +```ts +const updatedText = formatMarkdownBody({ + body: selectedText, + filePath: activeFile.path, + settings: this.settings, + baseDir, + candidateIndex: { + trie: this.trie, + candidateMap: this.candidateMap, + }, + linkGenerator, +}) +``` + +In the AI command, keep `resolvedAmbiguities` rendering in `replaceLinks` until Task 1 scanner output supports direct formatting-run AI rendering. Use `toReplaceLinksSettings(this.settings, baseDir)` rather than passing `this.settings` directly. + +- [ ] **Step 6: Move pure URL-title frontmatter test coverage** + +In `src/__tests__/main-url-title-frontmatter.test.ts`, keep only the test that verifies `buildUrlTitleMap` does not fetch when active-file frontmatter disables URL titles. Remove the `plugin.modifyLinks` pure transformation test because `src/__tests__/formatting-run.test.ts` now covers it without Obsidian mocks. + +- [ ] **Step 7: Run focused main and formatting tests** + +Run: + +```bash +npx vitest run src/__tests__/formatting-run.test.ts src/__tests__/main-url-title-frontmatter.test.ts src/__tests__/main-link-generator.test.ts +``` + +Expected: PASS. + +- [ ] **Step 8: Run full verification for the stage** + +Run: + +```bash +npm run test -- --reporter=dot +npm run tsc +npm run lint +``` + +Expected: all commands exit 0. + +- [ ] **Step 9: Commit formatting-run stage** + +Run: + +```bash +git add src/formatting-run.ts src/__tests__/formatting-run.test.ts src/main.ts src/__tests__/main-url-title-frontmatter.test.ts +MSG_FILE=$(mktemp) +cat > "$MSG_FILE" <<'EOF' +refactor: extract pure formatting run + +Why: +- main.ts mixed Obsidian adapter work with pure transformation sequencing, which made formatting behavior require plugin mocks to test. +- A pure formatting run increases locality for URL formatting, URL title replacement, and link replacement order. + +What: +- Add a formatting-run module with document and body formatting entry points. +- Route file and selection formatting through the new module. +- Move pure frontmatter URL-title coverage out of main adapter tests. +EOF +git commit -F "$MSG_FILE" +rm -f "$MSG_FILE" +``` + +--- + +### Task 3: Settings Catalog Module + +**Files:** +- Create: `src/settings/settings-catalog.ts` +- Create: `src/settings/__tests__/settings-catalog.test.ts` +- Modify: `src/settings/settings-info.ts` +- Modify: `src/settings/settings.ts` +- Modify: `src/formatting-run.ts` +- Test: `src/settings/__tests__/settings-catalog.test.ts` + +**Interfaces:** +- Produces: + +```ts +export type AutomaticLinkerSettings = { + formatOnSave: boolean + showNotice: boolean + respectNewFileFolderPath: boolean + includeAliases: boolean + proximityBasedLinking: boolean + ignoreDateFormats: boolean + ignoreHeadings: boolean + formatGitHubURLs: boolean + githubEnterpriseURLs: string[] + formatJiraURLs: boolean + jiraURLs: string[] + formatLinearURLs: boolean + debug: boolean + ignoreCase: boolean + matchSentenceCase: boolean + replaceUrlWithTitle: boolean + replaceUrlWithTitleIgnoreDomains: string[] + excludeDirsFromAutoLinking: string[] + preventSelfLinking: boolean + removeAliasInDirs: string[] + ignoreMarkdownTables: boolean + runLinterAfterFormatting: boolean + runPrettierAfterFormatting: boolean + formatDelayMs: number + aiEnabled: boolean + aiEndpoint: string + aiModel: string + aiMaxContext: number +} + +export type SettingControl = "toggle" | "text" | "textarea" + +export interface SettingCatalogEntry { + key: K + group: string + name: string + description: string + control: SettingControl + placeholder?: string + multiline?: boolean + refreshesIndex: boolean + runtimeOnly?: boolean +} + +export const DEFAULT_SETTINGS: AutomaticLinkerSettings +export const SETTINGS_CATALOG: readonly SettingCatalogEntry[] +export const settingRefreshesIndex: (key: keyof AutomaticLinkerSettings) => boolean +export const projectReplaceLinksSettings: ( + settings: AutomaticLinkerSettings, + baseDir?: string, +) => ReplaceLinksSettings +export const projectUrlFormattingSettings: ( + settings: AutomaticLinkerSettings, +) => Pick +``` + +- [ ] **Step 1: Write settings catalog tests** + +Create `src/settings/__tests__/settings-catalog.test.ts`: + +```ts +import { describe, expect, it } from "vitest" +import { + DEFAULT_SETTINGS, + SETTINGS_CATALOG, + projectReplaceLinksSettings, + projectUrlFormattingSettings, + settingRefreshesIndex, +} from "../settings-catalog" + +describe("SETTINGS_CATALOG", () => { + it("covers every default setting exactly once", () => { + const defaultKeys = Object.keys(DEFAULT_SETTINGS).sort() + const catalogKeys = SETTINGS_CATALOG.map(entry => entry.key).sort() + + expect(catalogKeys).toEqual(defaultKeys) + expect(new Set(catalogKeys).size).toBe(catalogKeys.length) + }) + + it("marks current index-refresh settings", () => { + expect(settingRefreshesIndex("respectNewFileFolderPath")).toBe(true) + expect(settingRefreshesIndex("includeAliases")).toBe(true) + expect(settingRefreshesIndex("proximityBasedLinking")).toBe(true) + expect(settingRefreshesIndex("ignoreDateFormats")).toBe(true) + expect(settingRefreshesIndex("ignoreCase")).toBe(true) + expect(settingRefreshesIndex("preventSelfLinking")).toBe(true) + expect(settingRefreshesIndex("excludeDirsFromAutoLinking")).toBe(true) + expect(settingRefreshesIndex("removeAliasInDirs")).toBe(true) + expect(settingRefreshesIndex("debug")).toBe(false) + }) +}) + +describe("settings projections", () => { + it("projects link replacement settings", () => { + expect(projectReplaceLinksSettings({ + ...DEFAULT_SETTINGS, + proximityBasedLinking: false, + ignoreDateFormats: false, + ignoreCase: false, + matchSentenceCase: false, + preventSelfLinking: true, + removeAliasInDirs: ["archive"], + ignoreHeadings: true, + ignoreMarkdownTables: true, + }, "pages")).toEqual({ + proximityBasedLinking: false, + baseDir: "pages", + ignoreDateFormats: false, + ignoreCase: false, + matchSentenceCase: false, + preventSelfLinking: true, + removeAliasInDirs: ["archive"], + ignoreHeadings: true, + ignoreMarkdownTables: true, + }) + }) + + it("projects URL formatting settings", () => { + expect(projectUrlFormattingSettings({ + ...DEFAULT_SETTINGS, + formatGitHubURLs: false, + githubEnterpriseURLs: ["github.enterprise.com"], + formatJiraURLs: false, + jiraURLs: ["jira.example.com"], + formatLinearURLs: true, + })).toEqual({ + formatGitHubURLs: false, + githubEnterpriseURLs: ["github.enterprise.com"], + formatJiraURLs: false, + jiraURLs: ["jira.example.com"], + formatLinearURLs: true, + }) + }) +}) +``` + +- [ ] **Step 2: Run catalog tests to verify RED** + +Run: + +```bash +npx vitest run src/settings/__tests__/settings-catalog.test.ts +``` + +Expected: FAIL because `src/settings/settings-catalog.ts` does not exist. + +- [ ] **Step 3: Implement settings catalog** + +Create `src/settings/settings-catalog.ts` by moving the `AutomaticLinkerSettings` type and `DEFAULT_SETTINGS` object from `src/settings/settings-info.ts` into the new file. Add `SETTINGS_CATALOG` entries for every setting. Preserve every current name, description, placeholder, default, and refresh behavior from `src/settings/settings.ts`. + +The projection functions must be: + +```ts +export const projectReplaceLinksSettings = ( + settings: AutomaticLinkerSettings, + baseDir?: string, +): ReplaceLinksSettings => ({ + proximityBasedLinking: settings.proximityBasedLinking, + baseDir, + ignoreDateFormats: settings.ignoreDateFormats, + ignoreCase: settings.ignoreCase, + matchSentenceCase: settings.matchSentenceCase, + preventSelfLinking: settings.preventSelfLinking, + removeAliasInDirs: settings.removeAliasInDirs, + ignoreHeadings: settings.ignoreHeadings, + ignoreMarkdownTables: settings.ignoreMarkdownTables, +}) + +export const projectUrlFormattingSettings = ( + settings: AutomaticLinkerSettings, +) => ({ + formatGitHubURLs: settings.formatGitHubURLs, + githubEnterpriseURLs: settings.githubEnterpriseURLs, + formatJiraURLs: settings.formatJiraURLs, + jiraURLs: settings.jiraURLs, + formatLinearURLs: settings.formatLinearURLs, +}) + +export const settingRefreshesIndex = ( + key: keyof AutomaticLinkerSettings, +): boolean => SETTINGS_CATALOG.find(entry => entry.key === key)?.refreshesIndex ?? false +``` + +Modify `src/settings/settings-info.ts` to preserve imports from existing callers: + +```ts +export type { AutomaticLinkerSettings } from "./settings-catalog" +export { DEFAULT_SETTINGS } from "./settings-catalog" +``` + +Modify `src/formatting-run.ts` to import `projectReplaceLinksSettings` and remove its local `toReplaceLinksSettings` implementation. Re-export the projection for existing imports: + +```ts +export { projectReplaceLinksSettings as toReplaceLinksSettings } from "./settings/settings-catalog" +``` + +- [ ] **Step 4: Run catalog tests to verify GREEN** + +Run: + +```bash +npx vitest run src/settings/__tests__/settings-catalog.test.ts src/__tests__/formatting-run.test.ts +``` + +Expected: PASS. + +- [ ] **Step 5: Render settings tab from catalog metadata** + +Modify `src/settings/settings.ts` so the `display()` method iterates grouped catalog entries. Keep save behavior identical: + +```ts +import { App, PluginSettingTab, Setting } from "obsidian" +import AutomaticLinkerPlugin from "../main" +import { + AutomaticLinkerSettings, + SETTINGS_CATALOG, + SettingCatalogEntry, + settingRefreshesIndex, +} from "./settings-catalog" +``` + +Add private helpers to `AutomaticLinkerPluginSettingsTab`: + +```ts +private async setSettingValue( + key: K, + value: AutomaticLinkerSettings[K], +) { + this.plugin.settings[key] = value + await this.plugin.saveData(this.plugin.settings) + if (settingRefreshesIndex(key)) { + this.plugin.refreshFileDataAndTrie() + } +} + +private renderSetting(containerEl: HTMLElement, entry: SettingCatalogEntry) { + const setting = new Setting(containerEl) + .setName(entry.name) + .setDesc(entry.description) + + const value = this.plugin.settings[entry.key] + if (entry.control === "toggle") { + setting.addToggle((toggle) => { + toggle + .setValue(Boolean(value)) + .onChange(async nextValue => { + await this.setSettingValue(entry.key, nextValue as never) + }) + }) + } + if (entry.control === "text") { + setting.addText((text) => { + text.setPlaceholder(entry.placeholder ?? "") + .setValue(String(value)) + .onChange(async nextValue => { + const parsedValue = typeof value === "number" + ? parseInt(nextValue) + : nextValue + if (typeof value === "number" && (isNaN(parsedValue as number) || (parsedValue as number) < 0)) { + return + } + await this.setSettingValue(entry.key, parsedValue as never) + }) + }) + } + if (entry.control === "textarea") { + setting.addTextArea((text) => { + text.setPlaceholder(entry.placeholder ?? "") + .setValue(Array.isArray(value) ? value.join("\n") : String(value)) + .onChange(async nextValue => { + await this.setSettingValue( + entry.key, + nextValue.split("\n").map(item => item.trim()).filter(Boolean) as never, + ) + }) + text.inputEl.rows = 4 + text.inputEl.cols = 50 + }) + } +} +``` + +In `display()`, render groups in catalog order: + +```ts +const renderedGroups = new Set() +for (const entry of SETTINGS_CATALOG) { + if (!renderedGroups.has(entry.group)) { + new Setting(containerEl).setName(entry.group).setHeading() + renderedGroups.add(entry.group) + } + this.renderSetting(containerEl, entry) +} +``` + +Keep the current special side effects by encoding them in catalog-driven rendering: + +- `replaceUrlWithTitleIgnoreDomains` on change calls `await this.plugin.buildUrlTitleMap()` after saving. +- Numeric settings reject invalid values. `formatDelayMs` accepts `0`; `aiMaxContext` accepts values greater than `0`. + +- [ ] **Step 6: Run settings and full adapter tests** + +Run: + +```bash +npx vitest run src/settings/__tests__/settings-catalog.test.ts src/__tests__/main-link-generator.test.ts src/__tests__/main-url-title-frontmatter.test.ts +``` + +Expected: PASS. + +- [ ] **Step 7: Run full verification for the stage** + +Run: + +```bash +npm run test -- --reporter=dot +npm run tsc +npm run lint +``` + +Expected: all commands exit 0. + +- [ ] **Step 8: Commit settings catalog stage** + +Run: + +```bash +git add src/settings/settings-catalog.ts src/settings/__tests__/settings-catalog.test.ts src/settings/settings-info.ts src/settings/settings.ts src/formatting-run.ts +MSG_FILE=$(mktemp) +cat > "$MSG_FILE" <<'EOF' +refactor(settings): centralize settings catalog + +Why: +- Adding a setting required coordinated edits across defaults, UI, index refresh behavior, and runtime projections. +- A catalog gives settings changes one locality point while preserving current user-facing labels and defaults. + +What: +- Move settings defaults and metadata into a catalog module. +- Add projection helpers for link replacement and URL formatting. +- Render the settings tab from catalog entries without changing visible settings. +EOF +git commit -F "$MSG_FILE" +rm -f "$MSG_FILE" +``` + +--- + +### Task 4: Markdown Segment Module + +**Files:** +- Create: `src/markdown-segments.ts` +- Create: `src/__tests__/markdown-segments.test.ts` +- Modify: `src/replace-links/replace-links.ts` +- Modify: `src/replace-url-with-title/index.ts` +- Modify: `src/replace-url-with-title/utils/list-up-all-urls.ts` +- Test: `src/__tests__/markdown-segments.test.ts` +- Test: `src/replace-links/__tests__/replace-links.code.test.ts` +- Test: `src/replace-links/__tests__/replace-links.callout.test.ts` +- Test: `src/replace-url-with-title/utils/__tests__/list-up-all-urls.test.ts` + +**Interfaces:** +- Produces: + +```ts +export type MarkdownSegmentKind = "prose" | "protected" +export type MarkdownProtectedKind = + | "inline-code" + | "fenced-code" + | "wikilink" + | "markdown-link" + | "single-bracket" + | "url" + | "heading" + | "callout" + | "table-row" + +export interface MarkdownSegment { + kind: MarkdownSegmentKind + protectedKind?: MarkdownProtectedKind + start: number + end: number + text: string +} + +export interface SegmentMarkdownOptions { + protectHeadings?: boolean + protectCallouts?: boolean + protectTableRows?: boolean + protectUrls?: boolean +} + +export const segmentMarkdown: ( + text: string, + options?: SegmentMarkdownOptions, +) => MarkdownSegment[] + +export const mapMarkdownProse: ( + text: string, + transform: (segmentText: string, segment: MarkdownSegment) => string, + options?: SegmentMarkdownOptions, +) => string +``` + +- [ ] **Step 1: Write Markdown segment tests** + +Create `src/__tests__/markdown-segments.test.ts`: + +```ts +import { describe, expect, it } from "vitest" +import { mapMarkdownProse, segmentMarkdown } from "../markdown-segments" + +describe("segmentMarkdown", () => { + it("round-trips prose and protected inline code", () => { + const text = "Use `TypeScript` with TypeScript" + const segments = segmentMarkdown(text) + + expect(segments.map(segment => ({ + kind: segment.kind, + protectedKind: segment.protectedKind, + text: segment.text, + }))).toEqual([ + { kind: "prose", protectedKind: undefined, text: "Use " }, + { kind: "protected", protectedKind: "inline-code", text: "`TypeScript`" }, + { kind: "prose", protectedKind: undefined, text: " with TypeScript" }, + ]) + expect(segments.map(segment => segment.text).join("")).toBe(text) + }) + + it("protects fenced code blocks including unclosed blocks", () => { + const text = "before\n```ts\nTypeScript" + const segments = segmentMarkdown(text) + + expect(segments.map(segment => ({ + kind: segment.kind, + protectedKind: segment.protectedKind, + text: segment.text, + }))).toEqual([ + { kind: "prose", protectedKind: undefined, text: "before\n" }, + { kind: "protected", protectedKind: "fenced-code", text: "```ts\nTypeScript" }, + ]) + }) + + it("protects headings, tables, and callouts when requested", () => { + const text = "# TypeScript\n| TypeScript |\n> [!note]\n> TypeScript\nTypeScript" + const segments = segmentMarkdown(text, { + protectHeadings: true, + protectTableRows: true, + protectCallouts: true, + }) + + expect(segments.filter(segment => segment.kind === "protected").map(segment => segment.protectedKind)).toEqual([ + "heading", + "table-row", + "callout", + ]) + expect(segments.map(segment => segment.text).join("")).toBe(text) + }) +}) + +describe("mapMarkdownProse", () => { + it("transforms only prose segments", () => { + const result = mapMarkdownProse( + "TypeScript `TypeScript` [[TypeScript]]", + text => text.replaceAll("TypeScript", "TS"), + ) + + expect(result).toBe("TS `TypeScript` [[TypeScript]]") + }) +}) +``` + +- [ ] **Step 2: Run Markdown segment tests to verify RED** + +Run: + +```bash +npx vitest run src/__tests__/markdown-segments.test.ts +``` + +Expected: FAIL because `src/markdown-segments.ts` does not exist. + +- [ ] **Step 3: Implement Markdown segment module** + +Create `src/markdown-segments.ts`. Implement `segmentMarkdown` as a single ordered scanner: + +```ts +const PROTECTED_PATTERN = /(```[\s\S]*?(?:```|$)|~~~[\s\S]*?(?:~~~|$)|`[^`]*`|\[\[([^\]]+)\]\]|\[[^\]]+\]\([^)]+\)|\[[^\]]+\])/g +``` + +Before applying `PROTECTED_PATTERN`, collect optional block ranges for headings, callouts, and table rows. Merge overlapping protected ranges, sort by start index, and emit alternating prose/protected segments. `mapMarkdownProse` must join transformed prose with original protected text: + +```ts +export const mapMarkdownProse = ( + text: string, + transform: (segmentText: string, segment: MarkdownSegment) => string, + options: SegmentMarkdownOptions = {}, +): string => { + return segmentMarkdown(text, options) + .map(segment => segment.kind === "prose" ? transform(segment.text, segment) : segment.text) + .join("") +} +``` + +- [ ] **Step 4: Run Markdown segment tests to verify GREEN** + +Run: + +```bash +npx vitest run src/__tests__/markdown-segments.test.ts +``` + +Expected: PASS. + +- [ ] **Step 5: Route URL title modules through Markdown prose mapping** + +Modify `src/replace-url-with-title/index.ts`: + +```ts +import { mapMarkdownProse } from "../markdown-segments" +``` + +Wrap the existing URL replacement loop so it only receives prose: + +```ts +return mapMarkdownProse(resultBody, replaceUrlsInProse) +``` + +Move the current loop body into: + +```ts +const replaceUrlsInProse = (prose: string): string => { + let resultBody = prose + // current sorted URL replacement loop + return resultBody +} +``` + +Modify `src/replace-url-with-title/utils/list-up-all-urls.ts` to iterate prose segments: + +```ts +import { segmentMarkdown } from "../../markdown-segments" + +for (const segment of segmentMarkdown(body)) { + if (segment.kind === "protected") continue + URL_REGEX.lastIndex = 0 + while ((match = URL_REGEX.exec(segment.text)) !== null) { + const url = match[0] + const matchIndex = segment.start + match.index + // preserve existing markdown-link, angle-bracket, inline-code, domain, and punctuation checks that still apply + } +} +``` + +- [ ] **Step 6: Route replace-links protected regions through Markdown segments** + +Modify `src/replace-links/replace-links.ts` so `replaceLinks` calls `mapMarkdownProse` instead of manually replacing code blocks, headings, callouts, protected links, and table rows with placeholders: + +```ts +import { mapMarkdownProse } from "../markdown-segments" +``` + +The final body processing should become: + +```ts +return mapMarkdownProse( + body, + processTableAwareTextSegment, + { + protectHeadings: settings.ignoreHeadings, + protectCallouts: true, + protectTableRows: settings.ignoreMarkdownTables, + protectUrls: true, + }, +) +``` + +Keep existing-link replacement behavior for `resolvedAmbiguities` by allowing wikilinks to be handled before protected mapping when `resolvedAmbiguities` has the full wikilink key. Preserve the current tests in `src/replace-links/__tests__/ai-disambiguation.test.ts`. + +- [ ] **Step 7: Run focused Markdown and transformation tests** + +Run: + +```bash +npx vitest run src/__tests__/markdown-segments.test.ts src/replace-links/__tests__/replace-links.code.test.ts src/replace-links/__tests__/replace-links.callout.test.ts src/replace-links/__tests__/replace-links.table.test.ts src/replace-url-with-title/utils/__tests__/list-up-all-urls.test.ts src/replace-url-with-title/__tests__/replace-url-with-title.test.ts +``` + +Expected: PASS. + +- [ ] **Step 8: Run full verification for the stage** + +Run: + +```bash +npm run test -- --reporter=dot +npm run tsc +npm run lint +``` + +Expected: all commands exit 0. + +- [ ] **Step 9: Commit Markdown segment stage** + +Run: + +```bash +git add src/markdown-segments.ts src/__tests__/markdown-segments.test.ts src/replace-links/replace-links.ts src/replace-url-with-title/index.ts src/replace-url-with-title/utils/list-up-all-urls.ts +MSG_FILE=$(mktemp) +cat > "$MSG_FILE" <<'EOF' +refactor(markdown): centralize protected segment handling + +Why: +- Link replacement and URL title code each carried their own Markdown context checks, which made protected text behavior hard to keep consistent. +- A shared Markdown segment module improves locality for prose-only transformations. + +What: +- Add a pure Markdown segment module with prose mapping. +- Route link replacement and URL title flows through shared protected segment handling. +- Preserve exact output for code blocks, links, tables, headings, and callouts. +EOF +git commit -F "$MSG_FILE" +rm -f "$MSG_FILE" +``` + +--- + +### Task 5: URL Formatting Module + +**Files:** +- Create: `src/replace-urls/url-formatting.ts` +- Create: `src/replace-urls/__tests__/url-formatting.test.ts` +- Modify: `src/replace-urls/github.ts` +- Modify: `src/replace-urls/replace-urls.ts` +- Modify: `src/formatting-run.ts` +- Test: `src/replace-urls/__tests__/url-formatting.test.ts` +- Test: `src/replace-urls/__tests__/replace-urls.github.test.ts` +- Test: `src/replace-urls/__tests__/replace-urls.jira.test.ts` +- Test: `src/replace-urls/__tests__/replace-urls.linear.test.ts` + +**Interfaces:** +- Produces: + +```ts +export type UrlFormatter = ( + url: string, + settings: AutomaticLinkerSettings, +) => string + +export interface FormatURLsInTextOptions { + text: string + settings: AutomaticLinkerSettings + formatters?: UrlFormatter[] +} + +export const DEFAULT_URL_FORMATTERS: readonly UrlFormatter[] +export const formatURLWithAdapters: ( + url: string, + settings: AutomaticLinkerSettings, + formatters?: readonly UrlFormatter[], +) => string +export const formatURLsInText: (options: FormatURLsInTextOptions) => string +``` + +- [ ] **Step 1: Write URL formatting orchestration tests** + +Create `src/replace-urls/__tests__/url-formatting.test.ts`: + +```ts +import { describe, expect, it } from "vitest" +import { DEFAULT_SETTINGS } from "../../settings/settings-info" +import { + formatURLsInText, + formatURLWithAdapters, + UrlFormatter, +} from "../url-formatting" + +describe("formatURLWithAdapters", () => { + it("uses the first adapter that changes the URL", () => { + const first: UrlFormatter = url => `${url}-first` + const second: UrlFormatter = url => `${url}-second` + + expect(formatURLWithAdapters("https://example.com", DEFAULT_SETTINGS, [first, second])).toBe( + "https://example.com-first", + ) + }) + + it("returns the original URL when no adapter changes it", () => { + const unchanged: UrlFormatter = url => url + + expect(formatURLWithAdapters("https://example.com", DEFAULT_SETTINGS, [unchanged])).toBe( + "https://example.com", + ) + }) +}) + +describe("formatURLsInText", () => { + it("formats GitHub, Jira, and Linear URLs in one text pass", () => { + const result = formatURLsInText({ + text: [ + "https://github.com/owner/repo/issues/123", + "https://jira.company.com/browse/ABC-456", + "https://linear.app/team/issue/BUG-789/title", + ].join("\n"), + settings: { + ...DEFAULT_SETTINGS, + githubEnterpriseURLs: [], + jiraURLs: ["jira.company.com"], + formatGitHubURLs: true, + formatJiraURLs: true, + formatLinearURLs: true, + }, + }) + + expect(result).toBe([ + "[[github/owner/repo/issues/123]] [πŸ”—](https://github.com/owner/repo/issues/123)", + "[[company/jira/ABC/456]] [πŸ”—](https://jira.company.com/browse/ABC-456)", + "[[linear/team/BUG-789]] [πŸ”—](https://linear.app/team/issue/BUG-789)", + ].join("\n")) + }) + + it("leaves disabled adapter URLs unchanged", () => { + const result = formatURLsInText({ + text: "https://linear.app/team/issue/BUG-789/title", + settings: { + ...DEFAULT_SETTINGS, + formatLinearURLs: false, + }, + }) + + expect(result).toBe("https://linear.app/team/issue/BUG-789/title") + }) +}) +``` + +- [ ] **Step 2: Run URL formatting tests to verify RED** + +Run: + +```bash +npx vitest run src/replace-urls/__tests__/url-formatting.test.ts +``` + +Expected: FAIL because `src/replace-urls/url-formatting.ts` does not exist. + +- [ ] **Step 3: Implement URL formatting module** + +Create `src/replace-urls/url-formatting.ts`: + +```ts +import { AutomaticLinkerSettings } from "../settings/settings-info" +import { formatGitHubURL } from "./github" +import { formatJiraURL } from "./jira" +import { formatLinearURL } from "./linear" +import { mapMarkdownProse } from "../markdown-segments" + +export type UrlFormatter = ( + url: string, + settings: AutomaticLinkerSettings, +) => string + +export interface FormatURLsInTextOptions { + text: string + settings: AutomaticLinkerSettings + formatters?: readonly UrlFormatter[] +} + +export const DEFAULT_URL_FORMATTERS: readonly UrlFormatter[] = [ + formatGitHubURL, + formatJiraURL, + formatLinearURL, +] + +const URL_PATTERN = /(? { + for (const formatter of formatters) { + const formatted = formatter(url, settings) + if (formatted !== url) { + return formatted + } + } + return url +} + +export const formatURLsInText = ({ + text, + settings, + formatters = DEFAULT_URL_FORMATTERS, +}: FormatURLsInTextOptions): string => { + return mapMarkdownProse(text, prose => prose.replace(URL_PATTERN, match => { + return formatURLWithAdapters(match, settings, formatters) + })) +} +``` + +Modify `src/replace-urls/github.ts` to remove imports from `./jira` and `./linear`, and delete the exported `formatURL` function from that file. `formatGitHubURL` must remain unchanged. + +Modify `src/replace-urls/replace-urls.ts` to preserve the existing public helper for tests and compatibility: + +```ts +import { AutomaticLinkerSettings } from "../settings/settings-info" + +export const replaceURLs = ( + fileContent: string, + settings: AutomaticLinkerSettings, + formatter: (url: string, settings: AutomaticLinkerSettings) => string, +) => { + const urlPattern = /(? formatter(match, settings)) +} +``` + +- [ ] **Step 4: Run URL formatting tests to verify GREEN** + +Run: + +```bash +npx vitest run src/replace-urls/__tests__/url-formatting.test.ts src/replace-urls/__tests__/replace-urls.github.test.ts src/replace-urls/__tests__/replace-urls.jira.test.ts src/replace-urls/__tests__/replace-urls.linear.test.ts +``` + +Expected: PASS. + +- [ ] **Step 5: Route formatting-run through one URL formatting call** + +Modify `src/formatting-run.ts` to import: + +```ts +import { formatURLsInText } from "./replace-urls/url-formatting" +``` + +Replace the three URL formatting passes with: + +```ts +updatedBody = formatURLsInText({ + text: updatedBody, + settings, +}) +``` + +Remove direct imports of `formatGitHubURL`, `formatJiraURL`, `formatLinearURL`, and `replaceURLs` from `src/formatting-run.ts`. + +- [ ] **Step 6: Run focused formatting tests** + +Run: + +```bash +npx vitest run src/__tests__/formatting-run.test.ts src/replace-urls/__tests__/url-formatting.test.ts src/replace-urls/__tests__/replace-urls.test.ts +``` + +Expected: PASS. + +- [ ] **Step 7: Run full verification for the stage** + +Run: + +```bash +npm run test -- --reporter=dot +npm run tsc +npm run lint +``` + +Expected: all commands exit 0. + +- [ ] **Step 8: Commit URL formatting stage** + +Run: + +```bash +git add src/replace-urls/url-formatting.ts src/replace-urls/__tests__/url-formatting.test.ts src/replace-urls/github.ts src/replace-urls/replace-urls.ts src/formatting-run.ts +MSG_FILE=$(mktemp) +cat > "$MSG_FILE" <<'EOF' +refactor(replace-urls): centralize URL formatter orchestration + +Why: +- URL formatter adapters existed, but ordering and text traversal were split across main formatting code and adapter modules. +- A central URL formatting module gives adapter ordering one locality point. + +What: +- Add one-pass URL formatting orchestration for GitHub, Jira, and Linear adapters. +- Remove sibling adapter coupling from the GitHub formatter module. +- Route formatting-run through the central URL formatting module. +EOF +git commit -F "$MSG_FILE" +rm -f "$MSG_FILE" +``` + +--- + +## Final Verification + +- [ ] **Step 1: Run the full regression suite** + +Run: + +```bash +npm run test -- --reporter=dot +npm run tsc +npm run lint +npm run build +``` + +Expected: all commands exit 0. + +- [ ] **Step 2: Inspect architecture-focused diffs** + +Run: + +```bash +git status --short +git log --oneline -5 +``` + +Expected: +- working tree is clean after all stage commits +- recent commits are the five architecture deepening stage commits + +- [ ] **Step 3: Confirm success criteria** + +Confirm these statements against the final source: + +- `src/main.ts` no longer owns pure transformation sequencing. +- `src/utils/resolve-ambiguities.ts` no longer has a direct trie scanning loop. +- `src/settings/settings-catalog.ts` owns settings defaults, metadata, refresh flags, and projections. +- `src/markdown-segments.ts` is used by link replacement and URL title modules. +- `src/replace-urls/url-formatting.ts` owns URL adapter ordering and text traversal. diff --git a/docs/superpowers/specs/2026-06-22-architecture-deepening-design.md b/docs/superpowers/specs/2026-06-22-architecture-deepening-design.md new file mode 100644 index 0000000..7329fcd --- /dev/null +++ b/docs/superpowers/specs/2026-06-22-architecture-deepening-design.md @@ -0,0 +1,277 @@ +# Architecture Deepening Design + +## Goal + +Deepen the architecture of Obsidian Automatic Linker without changing user-visible behavior. + +The work should reduce duplicated transformation logic, move Obsidian-specific code toward thin adapters, and concentrate link-formatting behavior in modules with higher locality and leverage. + +## Scope + +Implement all five architecture-review candidates in a staged sequence: + +1. Deepen candidate scanning. +2. Deepen the formatting run. +3. Deepen the settings catalog. +4. Deepen Markdown protected segments. +5. Deepen URL formatting adapters. + +Each stage must preserve the existing behavior covered by the current test suite. New tests should be added before production changes when a stage exposes behavior that is currently duplicated, underspecified, or hard to test through the existing interface. + +## Non-Goals + +- Do not add new user-facing features. +- Do not change existing settings names, defaults, command ids, or manifest metadata. +- Do not change Obsidian link output formats except where an existing test already expects that behavior. +- Do not replace the current test framework or build tooling. +- Do not introduce a Markdown parser dependency unless a later implementation step proves regex segmentation cannot preserve current behavior. + +## Architecture + +The current codebase has two large centers of gravity: + +- `src/main.ts` owns Obsidian lifecycle, command registration, settings access, indexing, URL title fetching, URL formatting, link replacement, and AI command orchestration. +- `src/replace-links/replace-links.ts` owns a deep but broad implementation for link candidate matching and content transformation. + +The target architecture keeps the transformation core pure and moves app integration to adapters: + +- Obsidian commands, editor reads/writes, notices, network requests, and plugin lifecycle remain in adapter code. +- Link candidate scanning, Markdown segment protection, URL formatting, settings projection, and formatting-run sequencing become testable modules. +- Existing modules are split only where the split concentrates complexity behind a smaller interface. + +## Candidate 1: Deepen Candidate Scanning + +### Problem + +`replaceLinks` and `resolveAmbiguities` both scan note text for possible link candidates. + +`replaceLinks` contains the full implementation: protected spans, trie traversal, CJK handling, sentence-case matching, namespace scope, table awareness, and existing-link correction. `resolveAmbiguities` contains a simplified scanner for AI requests. This weakens locality because changing candidate matching requires checking two implementations. + +### Design + +Create a candidate scanning module used by both link replacement and AI ambiguity resolution. + +The module should expose occurrences of linkable text and existing links in terms of the existing data model: + +- source text range +- matched text +- candidate key +- candidate data +- whether the occurrence is protected, replaceable, or an existing wikilink +- enough context for AI request construction + +`replaceLinks` remains responsible for rendering replacements. `resolveAmbiguities` becomes an adapter that asks the scanner for ambiguous occurrences and sends them to the AI client. + +### Testing + +Tests should prove that AI ambiguity detection respects the same candidate matching rules as replacement for: + +- protected inline code +- existing Markdown links +- case-insensitive candidates +- namespace-scoped candidates +- CJK or sentence-case behavior where currently covered by `replaceLinks` + +## Candidate 2: Deepen The Formatting Run + +### Problem + +`main.ts` currently sequences multiple pure transformations directly: + +- GitHub URL formatting +- Jira URL formatting +- Linear URL formatting +- URL title replacement +- frontmatter/body splitting +- link replacement +- settings projection for `replaceLinks` + +The file and selection commands duplicate the settings projection. The AI command passes a broader settings object into `replaceLinks`, which makes the transformation interface inconsistent. + +### Design + +Create a formatting-run module that owns content transformation sequencing. + +The Obsidian plugin adapter should supply: + +- current file path +- raw content or selected text +- frontmatter when available +- candidate index +- URL title map +- link generator +- projected runtime settings + +The formatting-run module should return transformed text and avoid direct Obsidian dependencies. + +`main.ts` keeps editor reads/writes, command registration, notices, vault traversal, URL title fetching, and plugin integration. + +### Testing + +Tests should exercise the formatting run without constructing `AutomaticLinkerPlugin` or mocking Obsidian. Move the existing `main.ts` tests that only verify pure transformation behavior to formatting-run tests; keep `main.ts` tests only for Obsidian adapter behavior such as link generator integration. + +## Candidate 3: Deepen The Settings Catalog + +### Problem + +Adding a setting currently crosses several places: + +- `src/settings/settings-info.ts` +- `src/settings/settings.ts` +- `src/main.ts` +- `src/replace-links/replace-links.ts` + +The repo's development guide already warns about this. That warning is evidence that the settings module is shallow: the interface nearly matches the implementation, and adding a field lacks locality. + +### Design + +Create a settings catalog module that owns settings metadata and projections. + +The catalog should centralize: + +- default values +- setting groups and display metadata used by the settings tab +- whether changing a setting requires index refresh +- projection from `AutomaticLinkerSettings` to link replacement settings +- projection from `AutomaticLinkerSettings` to URL formatting adapter settings, including format flags and configured domains + +The settings tab remains an Obsidian UI adapter that renders catalog entries into `Setting` controls. + +### Testing + +Tests should cover: + +- every default setting has catalog metadata or an explicit reason it is runtime-only +- link replacement projection includes all fields expected by `ReplaceLinksSettings` +- refresh-required settings match current behavior + +## Candidate 4: Deepen Markdown Protected Segments + +### Problem + +Markdown protection rules appear in multiple modules: + +- `replaceLinks` protects code blocks, inline code, existing wikilinks, Markdown links, headings, tables, and callouts. +- URL title replacement and URL discovery perform their own context checks. + +The URL title module also documents an incomplete fenced-code check. This is locality friction and an opportunity to put Markdown segmentation behind one module. + +### Design + +Create a Markdown segment module that divides text into transformable prose and protected segments. + +The module should preserve exact text and ordering. Adapters can then transform only prose segments: + +- link replacement adapter +- URL discovery adapter +- URL title replacement adapter + +This stage should be conservative. First extract fenced code, inline code, Markdown links, and wikilinks. Then add headings, tables, and callouts to the same module after exact round-trip segment tests are passing. + +### Testing + +Tests should cover exact round-trip reconstruction and per-segment transformation for: + +- inline code +- fenced code blocks, including unclosed fences +- existing wikilinks +- Markdown links +- headings when ignored +- Markdown tables when ignored +- Obsidian callouts + +## Candidate 5: Deepen URL Formatting Adapters + +### Problem + +URL formatting has separate adapter modules for GitHub, Jira, and Linear, but orchestration is split: + +- `main.ts` chooses individual passes. +- `replaceURLs` runs a formatter over URL matches. +- `github.ts` imports sibling formatters through `formatURL`, creating adapter coupling. + +This is a shallow seam: there are already multiple adapters, but the module that should own ordering and one pass over URLs does not yet own them fully. + +### Design + +Create a URL formatting module that owns: + +- URL discovery in transformable prose +- adapter ordering +- selecting the first adapter that changes a URL +- preserving current output formats + +GitHub, Jira, and Linear modules become independent adapters. They should not import each other. + +### Testing + +Tests should prove: + +- all existing GitHub, Jira, and Linear outputs remain unchanged +- a text body can be processed in one URL formatting call +- adapter ordering is deterministic +- disabled settings prevent matching adapters from changing text + +## Data Flow + +The target formatting flow is: + +1. Obsidian adapter reads active file, selection, or vault file. +2. Obsidian adapter gathers frontmatter, candidate index, settings, URL title map, and link generator. +3. Formatting-run module splits frontmatter from body when needed. +4. Formatting-run module runs URL formatting, URL title replacement, and link replacement using pure modules. +5. Obsidian adapter writes the result back to the editor or vault. +6. Optional plugin integrations such as Prettier and Obsidian Linter remain in `main.ts`. + +## Error Handling + +Preserve current error behavior: + +- Obsidian command callbacks catch and log unexpected errors. +- URL title fetching failures are logged only when debug mode is enabled. +- AI API errors continue to surface through the AI command catch block and notice. +- Pure transformation modules should not create `Notice` instances or call Obsidian APIs. + +## Test Strategy + +Use TDD for each implementation stage. + +The expected test layers are: + +- pure module tests for candidate scanning, Markdown segmentation, formatting run, settings projection, and URL formatting orchestration +- existing integration-style tests around `AutomaticLinkerPlugin` only where Obsidian adapter behavior is the subject +- full regression suite after each stage + +The baseline regression command is: + +```bash +npm run test -- --reporter=dot +``` + +## Migration Strategy + +Each stage should be reversible and independently reviewable. + +1. Add tests for the behavior being centralized. +2. Add the new module with minimal implementation. +3. Route one existing caller through the new module. +4. Run focused tests. +5. Route the second caller through the same module. +6. Run the full test suite. +7. Commit the stage. + +## Risks + +- Candidate scanning is the highest-risk stage because it touches the most behavior in `replaceLinks`. +- Markdown segmentation can accidentally change exact output if placeholder restoration differs from current code. +- Settings catalog work can create broad churn in `settings.ts`; keep the UI adapter mechanical and avoid visual changes. +- URL formatting must preserve current output strings, including the existing link icon and Enterprise URL handling. + +## Success Criteria + +- All current tests pass after each stage. +- `main.ts` no longer owns pure transformation sequencing. +- `resolveAmbiguities` no longer has a separate simplified candidate scanner. +- Link replacement settings projection exists in one place. +- URL formatting can process all configured adapters through one module call. +- Git status shows only intentional source, test, and documentation changes for each stage. diff --git a/src/__tests__/formatting-run.test.ts b/src/__tests__/formatting-run.test.ts new file mode 100644 index 0000000..ed43131 --- /dev/null +++ b/src/__tests__/formatting-run.test.ts @@ -0,0 +1,208 @@ +import { describe, expect, it } from "vitest" +import { + formatMarkdownBody, + formatMarkdownDocument, + formatMarkdownSelection, + toReplaceLinksSettings, +} from "../formatting-run" +import { buildCandidateTrieForTest } from "../replace-links/__tests__/test-helpers" +import { DEFAULT_SETTINGS } from "../settings/settings-info" + +describe("toReplaceLinksSettings", () => { + it("projects only replacement settings and applies baseDir", () => { + expect(toReplaceLinksSettings({ + ...DEFAULT_SETTINGS, + proximityBasedLinking: false, + ignoreDateFormats: false, + ignoreCase: false, + matchSentenceCase: false, + preventSelfLinking: true, + removeAliasInDirs: ["archive"], + ignoreHeadings: true, + ignoreMarkdownTables: true, + }, "pages")).toEqual({ + proximityBasedLinking: false, + baseDir: "pages", + ignoreDateFormats: false, + ignoreCase: false, + matchSentenceCase: false, + preventSelfLinking: true, + removeAliasInDirs: ["archive"], + ignoreHeadings: true, + ignoreMarkdownTables: true, + }) + }) +}) + +describe("formatMarkdownDocument", () => { + it("preserves frontmatter and respects URL title opt-out", () => { + const result = formatMarkdownDocument({ + content: "---\nautomatic-linker-disable-url-title: true\n---\nhttps://example.com", + filePath: "current-file.md", + frontmatter: { "automatic-linker-disable-url-title": true }, + settings: { + ...DEFAULT_SETTINGS, + formatGitHubURLs: false, + formatJiraURLs: false, + formatLinearURLs: false, + replaceUrlWithTitle: true, + }, + urlTitleMap: new Map([["https://example.com", "Example Title"]]), + }) + + expect(result).toBe( + "---\nautomatic-linker-disable-url-title: true\n---\nhttps://example.com", + ) + }) + + it("formats GitHub URLs in frontmatter without moving body-only formatting into frontmatter", () => { + const { candidateMap, trie } = buildCandidateTrieForTest({ + files: [{ path: "notes/TypeScript" }], + settings: { + scoped: false, + baseDir: undefined, + ignoreCase: true, + }, + }) + + const result = formatMarkdownDocument({ + content: + "---\nautomatic-linker-disable-url-title: true\nreference: https://github.com/openai/openai/issues/1\nterm: TypeScript\n---\nTypeScript https://example.com", + filePath: "current-file.md", + frontmatter: { + "automatic-linker-disable-url-title": true, + }, + settings: { + ...DEFAULT_SETTINGS, + formatGitHubURLs: true, + formatJiraURLs: false, + formatLinearURLs: false, + replaceUrlWithTitle: true, + ignoreCase: true, + }, + candidateIndex: { candidateMap, trie }, + }) + + expect(result).toBe( + "---\nautomatic-linker-disable-url-title: true\nreference: [[github/openai/openai/issues/1]] [πŸ”—](https://github.com/openai/openai/issues/1)\nterm: TypeScript\n---\n[[notes/TypeScript|TypeScript]] https://example.com", + ) + }) + + it("runs URL formatting, URL titles, and link replacement in order", () => { + const { candidateMap, trie } = buildCandidateTrieForTest({ + files: [{ path: "notes/TypeScript" }], + settings: { + scoped: false, + baseDir: undefined, + ignoreCase: true, + }, + }) + + const result = formatMarkdownDocument({ + content: "Read TypeScript at https://example.com", + filePath: "current-file.md", + settings: { + ...DEFAULT_SETTINGS, + formatGitHubURLs: false, + formatJiraURLs: false, + formatLinearURLs: false, + replaceUrlWithTitle: true, + ignoreCase: true, + }, + candidateIndex: { candidateMap, trie }, + urlTitleMap: new Map([["https://example.com", "Example Title"]]), + }) + + expect(result).toBe("Read [[notes/TypeScript|TypeScript]] at [Example Title](https://example.com)") + }) +}) + +describe("formatMarkdownBody", () => { + it("formats selected body text without frontmatter splitting", () => { + const { candidateMap, trie } = buildCandidateTrieForTest({ + files: [{ path: "notes/TypeScript" }], + settings: { + scoped: false, + baseDir: undefined, + ignoreCase: true, + }, + }) + + const result = formatMarkdownBody({ + body: "TypeScript", + filePath: "current-file.md", + settings: { + ...DEFAULT_SETTINGS, + formatGitHubURLs: false, + formatJiraURLs: false, + formatLinearURLs: false, + replaceUrlWithTitle: false, + ignoreCase: true, + }, + candidateIndex: { candidateMap, trie }, + }) + + expect(result).toBe("[[notes/TypeScript|TypeScript]]") + }) +}) + +describe("formatMarkdownSelection", () => { + it("keeps selection formatting to link replacement only", () => { + const { candidateMap, trie } = buildCandidateTrieForTest({ + files: [{ path: "notes/TypeScript" }], + settings: { + scoped: false, + baseDir: undefined, + ignoreCase: true, + }, + }) + + const result = formatMarkdownSelection({ + body: "TypeScript https://github.com/openai/openai/issues/1", + filePath: "current-file.md", + settings: { + ...DEFAULT_SETTINGS, + formatGitHubURLs: true, + formatJiraURLs: true, + formatLinearURLs: true, + replaceUrlWithTitle: true, + ignoreCase: true, + }, + candidateIndex: { candidateMap, trie }, + }) + + expect(result).toBe( + "[[notes/TypeScript|TypeScript]] https://github.com/openai/openai/issues/1", + ) + }) + + it("leaves linear URLs unchanged when a linear note exists", () => { + const { candidateMap, trie } = buildCandidateTrieForTest({ + files: [ + { path: "notes/linear" }, + { path: "notes/TypeScript" }, + ], + settings: { + scoped: false, + baseDir: undefined, + ignoreCase: true, + }, + }) + + const result = formatMarkdownSelection({ + body: "linear://workspace/issue/ACME-123", + filePath: "current-file.md", + settings: { + ...DEFAULT_SETTINGS, + formatGitHubURLs: true, + formatJiraURLs: true, + formatLinearURLs: true, + replaceUrlWithTitle: true, + ignoreCase: true, + }, + candidateIndex: { candidateMap, trie }, + }) + + expect(result).toBe("linear://workspace/issue/ACME-123") + }) +}) diff --git a/src/__tests__/main-link-generator.test.ts b/src/__tests__/main-link-generator.test.ts index 807d9b6..2408145 100644 --- a/src/__tests__/main-link-generator.test.ts +++ b/src/__tests__/main-link-generator.test.ts @@ -76,4 +76,56 @@ describe("AutomaticLinkerPlugin link generator", () => { expect(result).toBe("| [[notes/foo\\|bar]] | x |\n| --- | --- |\n") }) + + it("keeps selection formatting to link replacement only", async () => { + const { default: AutomaticLinkerPlugin } = await import("../main") + const settings = { + scoped: false, + baseDir: undefined, + ignoreCase: true, + } + const { candidateMap, trie } = buildCandidateTrieForTest({ + files: [{ path: "notes/TypeScript" }], + settings, + }) + const replaceSelection = vi.fn() + const app = { + workspace: { + getActiveFile: vi.fn(() => ({ path: "current-file.md" })), + activeEditor: { + editor: { + getSelection: vi.fn(() => "TypeScript https://github.com/openai/openai/issues/1"), + replaceSelection, + }, + }, + }, + vault: { + getConfig: vi.fn(() => "pages"), + getAbstractFileByPath: vi.fn((path: string) => { + if (path === "notes/TypeScript.md") return { path: "notes/TypeScript.md" } + return null + }), + }, + fileManager: { + generateMarkdownLink: vi.fn(() => "[[notes/TypeScript|TypeScript]]"), + }, + } + const plugin = new AutomaticLinkerPlugin(app as never, {} as never) + plugin.settings = { + ...DEFAULT_SETTINGS, + formatGitHubURLs: true, + formatJiraURLs: true, + formatLinearURLs: true, + replaceUrlWithTitle: true, + respectNewFileFolderPath: true, + } + ;(plugin as unknown as { trie: typeof trie }).trie = trie + ;(plugin as unknown as { candidateMap: typeof candidateMap }).candidateMap = candidateMap + + await plugin.mofifyLinksSelection() + + expect(replaceSelection).toHaveBeenCalledWith( + "[[notes/TypeScript|TypeScript]] https://github.com/openai/openai/issues/1", + ) + }) }) diff --git a/src/__tests__/main-url-title-frontmatter.test.ts b/src/__tests__/main-url-title-frontmatter.test.ts index 0a8b7e4..3804c00 100644 --- a/src/__tests__/main-url-title-frontmatter.test.ts +++ b/src/__tests__/main-url-title-frontmatter.test.ts @@ -46,31 +46,6 @@ describe("AutomaticLinkerPlugin URL title frontmatter opt-out", () => { requestMock.mockReset() }) - it("keeps raw URLs when URL title replacement is disabled in frontmatter", async () => { - const { default: AutomaticLinkerPlugin } = await import("../main") - const plugin = new AutomaticLinkerPlugin({} as never, {} as never) - plugin.settings = { - ...DEFAULT_SETTINGS, - formatGitHubURLs: false, - formatJiraURLs: false, - formatLinearURLs: false, - replaceUrlWithTitle: true, - } - ;(plugin as unknown as { urlTitleMap: Map }).urlTitleMap = new Map([ - ["https://example.com", "Example Title"], - ]) - - const result = plugin.modifyLinks( - "---\nautomatic-linker-disable-url-title: true\n---\nhttps://example.com", - "current-file.md", - { "automatic-linker-disable-url-title": true }, - ) - - expect(result).toBe( - "---\nautomatic-linker-disable-url-title: true\n---\nhttps://example.com", - ) - }) - it("does not fetch URL titles when disabled in active file frontmatter", async () => { const { default: AutomaticLinkerPlugin } = await import("../main") const activeFile = new MockTFile("current-file.md") diff --git a/src/__tests__/markdown-segments.test.ts b/src/__tests__/markdown-segments.test.ts new file mode 100644 index 0000000..2a16760 --- /dev/null +++ b/src/__tests__/markdown-segments.test.ts @@ -0,0 +1,102 @@ +import { describe, expect, it } from "vitest" +import { mapMarkdownProse, segmentMarkdown } from "../markdown-segments" + +describe("segmentMarkdown", () => { + it("round-trips prose and protected inline code", () => { + const text = "Use `TypeScript` with TypeScript" + const segments = segmentMarkdown(text) + + expect(segments.map(segment => ({ + kind: segment.kind, + protectedKind: segment.protectedKind, + text: segment.text, + }))).toEqual([ + { kind: "prose", protectedKind: undefined, text: "Use " }, + { kind: "protected", protectedKind: "inline-code", text: "`TypeScript`" }, + { kind: "prose", protectedKind: undefined, text: " with TypeScript" }, + ]) + expect(segments.map(segment => segment.text).join("")).toBe(text) + }) + + it("protects fenced code blocks including unclosed blocks", () => { + const text = "before\n```ts\nTypeScript" + const segments = segmentMarkdown(text) + + expect(segments.map(segment => ({ + kind: segment.kind, + protectedKind: segment.protectedKind, + text: segment.text, + }))).toEqual([ + { kind: "prose", protectedKind: undefined, text: "before\n" }, + { kind: "protected", protectedKind: "fenced-code", text: "```ts\nTypeScript" }, + ]) + }) + + it("protects tilde fenced code blocks", () => { + const text = "before\n~~~ts\nTypeScript\n~~~\nafter" + const segments = segmentMarkdown(text) + + expect(segments.map(segment => ({ + kind: segment.kind, + protectedKind: segment.protectedKind, + text: segment.text, + }))).toEqual([ + { kind: "prose", protectedKind: undefined, text: "before\n" }, + { kind: "protected", protectedKind: "fenced-code", text: "~~~ts\nTypeScript\n~~~\n" }, + { kind: "prose", protectedKind: undefined, text: "after" }, + ]) + }) + + it("protects headings, tables, and callouts when requested", () => { + const text = "# TypeScript\n| TypeScript |\n> [!note]\n> TypeScript\nTypeScript" + const segments = segmentMarkdown(text, { + protectHeadings: true, + protectTableRows: true, + protectCallouts: true, + }) + + expect(segments.filter(segment => segment.kind === "protected").map(segment => segment.protectedKind)).toEqual([ + "heading", + "table-row", + "callout", + ]) + expect(segments.map(segment => segment.text).join("")).toBe(text) + }) + + it("protects linear URLs when requested", () => { + const text = "linear://workspace/issue/ACME-123" + const segments = segmentMarkdown(text, { + protectUrls: true, + }) + + expect(segments).toEqual([ + { + kind: "protected", + protectedKind: "url", + start: 0, + end: text.length, + text, + }, + ]) + }) +}) + +describe("mapMarkdownProse", () => { + it("transforms only prose segments", () => { + const result = mapMarkdownProse( + "TypeScript `TypeScript` [[TypeScript]]", + text => text.replace(/TypeScript/g, "TS"), + ) + + expect(result).toBe("TS `TypeScript` [[TypeScript]]") + }) + + it("does not globally protect angle-bracket autolinks", () => { + const result = mapMarkdownProse( + " https://example.com", + text => text.replace(/https:\/\/example\.com/g, "URL"), + ) + + expect(result).toBe(" URL") + }) +}) diff --git a/src/formatting-run.ts b/src/formatting-run.ts new file mode 100644 index 0000000..244b976 --- /dev/null +++ b/src/formatting-run.ts @@ -0,0 +1,113 @@ +import { isUrlTitleReplacementOff } from "./frontmatter-utils" +import { + LinkGenerator, + replaceLinks, +} from "./replace-links/replace-links" +import { replaceUrlWithTitle } from "./replace-url-with-title" +import { formatURLsInText } from "./replace-urls/url-formatting" +import { + AutomaticLinkerSettings, + projectReplaceLinksSettings, +} from "./settings/settings-catalog" +import { CandidateData, TrieNode } from "./trie" + +export interface CandidateIndex { + trie: TrieNode + candidateMap: Map +} + +export interface FormattingRunOptions { + content: string + filePath: string + contentStart?: number + frontmatter?: Record + settings: AutomaticLinkerSettings + baseDir?: string + candidateIndex?: CandidateIndex + urlTitleMap?: Map + linkGenerator?: LinkGenerator +} + +export { projectReplaceLinksSettings as toReplaceLinksSettings } from "./settings/settings-catalog" + +const formatMarkdownURLs = ( + text: string, + settings: AutomaticLinkerSettings, +): string => + formatURLsInText({ + text, + settings, + }) + +export const formatMarkdownBody = ({ + body, + filePath, + frontmatter, + settings, + baseDir, + candidateIndex, + urlTitleMap = new Map(), + linkGenerator, +}: Omit & { body: string }): string => { + let updatedBody = formatMarkdownURLs(body, settings) + + if (settings.replaceUrlWithTitle && !isUrlTitleReplacementOff(frontmatter)) { + updatedBody = replaceUrlWithTitle({ body: updatedBody, urlTitleMap }) + } + if (candidateIndex) { + updatedBody = replaceLinks({ + body: updatedBody, + linkResolverContext: { + filePath: filePath.replace(/\.md$/, ""), + trie: candidateIndex.trie, + candidateMap: candidateIndex.candidateMap, + }, + settings: projectReplaceLinksSettings(settings, baseDir), + linkGenerator, + }) + } + + return updatedBody +} + +export const formatMarkdownSelection = ({ + body, + filePath, + settings, + baseDir, + candidateIndex, + linkGenerator, +}: Omit & { body: string }): string => { + if (!candidateIndex) { + return body + } + + return replaceLinks({ + body, + linkResolverContext: { + filePath: filePath.replace(/\.md$/, ""), + trie: candidateIndex.trie, + candidateMap: candidateIndex.candidateMap, + }, + settings: projectReplaceLinksSettings(settings, baseDir), + linkGenerator, + }) +} + +const inferContentStart = (content: string): number => { + const frontmatter = content.match(/^---\r?\n[\s\S]*?\r?\n---\r?\n?/) + return frontmatter?.[0].length ?? 0 +} + +export const formatMarkdownDocument = ({ + content, + contentStart = inferContentStart(content), + ...options +}: FormattingRunOptions): string => { + const frontmatterText = formatMarkdownURLs( + content.slice(0, contentStart), + options.settings, + ) + const body = content.slice(contentStart) + return frontmatterText + formatMarkdownBody({ ...options, body }) +} diff --git a/src/main.ts b/src/main.ts index 4df9a64..48c7a81 100644 --- a/src/main.ts +++ b/src/main.ts @@ -11,6 +11,11 @@ import { TFile, } from "obsidian" import { excludeLinks } from "./exclude-links" +import { + formatMarkdownDocument, + formatMarkdownSelection, + toReplaceLinksSettings, +} from "./formatting-run" import { isLinkingOff, isLinkingExcluded, @@ -26,13 +31,8 @@ import { LinkGeneratorParams, replaceLinks, } from "./replace-links/replace-links" -import { replaceUrlWithTitle } from "./replace-url-with-title" import { getTitleFromHtml } from "./replace-url-with-title/utils/get-title-from-html" import { listupAllUrls } from "./replace-url-with-title/utils/list-up-all-urls" -import { formatGitHubURL } from "./replace-urls/github" -import { formatJiraURL } from "./replace-urls/jira" -import { formatLinearURL } from "./replace-urls/linear" -import { replaceURLs } from "./replace-urls/replace-urls" import { AutomaticLinkerPluginSettingsTab } from "./settings/settings" import { AutomaticLinkerSettings, @@ -98,35 +98,15 @@ export default class AutomaticLinkerPlugin extends Plugin { filePath: string, frontmatter?: Record, ): string { - if (this.settings.formatGitHubURLs) { - fileContent = replaceURLs(fileContent, this.settings, formatGitHubURL) - } - - if (this.settings.formatJiraURLs) { - fileContent = replaceURLs(fileContent, this.settings, formatJiraURL) - } - - if (this.settings.formatLinearURLs) { - fileContent = replaceURLs( - fileContent, - this.settings, - formatLinearURL, - ) - } - - if ( - this.settings.replaceUrlWithTitle - && !isUrlTitleReplacementOff(frontmatter) - ) { - const { contentStart } = getFrontMatterInfo(fileContent) - const frontmatterText = fileContent.slice(0, contentStart) - const body = fileContent.slice(contentStart) - const updatedBody = replaceUrlWithTitle({ body, urlTitleMap: this.urlTitleMap }) - fileContent = frontmatterText + updatedBody - } - if (!this.trie || !this.candidateMap) { - return fileContent + return formatMarkdownDocument({ + content: fileContent, + filePath, + contentStart: getFrontMatterInfo(fileContent).contentStart, + frontmatter, + settings: this.settings, + urlTitleMap: this.urlTitleMap, + }) } if (this.settings.debug) { @@ -136,31 +116,21 @@ export default class AutomaticLinkerPlugin extends Plugin { new Notice(`Automatic Linker: ${new Date().toISOString()} modifyLinks started.`) } - const { contentStart } = getFrontMatterInfo(fileContent) - const frontmatterText = fileContent.slice(0, contentStart) - const linkGenerator = this.createLinkGenerator(filePath) const baseDir = this.settings.respectNewFileFolderPath ? this.app.vault.getConfig("newFileFolderPath") : undefined - const updatedBody = replaceLinks({ - body: fileContent.slice(contentStart), - linkResolverContext: { - filePath: filePath.replace(/\.md$/, ""), - trie: this.trie, - candidateMap: this.candidateMap, - }, - settings: { - proximityBasedLinking: this.settings.proximityBasedLinking, - baseDir, - ignoreDateFormats: this.settings.ignoreDateFormats, - ignoreCase: this.settings.ignoreCase, - matchSentenceCase: this.settings.matchSentenceCase, - preventSelfLinking: this.settings.preventSelfLinking, - removeAliasInDirs: this.settings.removeAliasInDirs, - ignoreHeadings: this.settings.ignoreHeadings, - ignoreMarkdownTables: this.settings.ignoreMarkdownTables, - }, - linkGenerator, + const candidateIndex = this.trie && this.candidateMap + ? { trie: this.trie, candidateMap: this.candidateMap } + : undefined + fileContent = formatMarkdownDocument({ + content: fileContent, + filePath, + contentStart: getFrontMatterInfo(fileContent).contentStart, + frontmatter, + settings: this.settings, + baseDir, + candidateIndex, + urlTitleMap: this.urlTitleMap, + linkGenerator: candidateIndex ? this.createLinkGenerator(filePath) : undefined, }) - fileContent = frontmatterText + updatedBody if (this.settings.debug) { console.log(new Date().toISOString(), "modifyLinks finished") @@ -263,24 +233,15 @@ export default class AutomaticLinkerPlugin extends Plugin { const linkGenerator = this.createLinkGenerator(activeFile.path) const baseDir = this.settings.respectNewFileFolderPath ? this.app.vault.getConfig("newFileFolderPath") : undefined - const updatedText = replaceLinks({ + const updatedText = formatMarkdownSelection({ body: selectedText, - linkResolverContext: { - filePath: activeFile.path.replace(/\.md$/, ""), + filePath: activeFile.path, + settings: this.settings, + baseDir, + candidateIndex: { trie: this.trie, candidateMap: this.candidateMap, }, - settings: { - proximityBasedLinking: this.settings.proximityBasedLinking, - baseDir, - ignoreDateFormats: this.settings.ignoreDateFormats, - ignoreCase: this.settings.ignoreCase, - matchSentenceCase: this.settings.matchSentenceCase, - preventSelfLinking: this.settings.preventSelfLinking, - removeAliasInDirs: this.settings.removeAliasInDirs, - ignoreHeadings: this.settings.ignoreHeadings, - ignoreMarkdownTables: this.settings.ignoreMarkdownTables, - }, linkGenerator, }) cm.replaceSelection(updatedText) @@ -516,6 +477,10 @@ export default class AutomaticLinkerPlugin extends Plugin { const fileContent = await this.app.vault.read(activeFile) const { contentStart } = getFrontMatterInfo(fileContent) const body = fileContent.slice(contentStart) + const normalizedActiveFilePath = activeFile.path.replace(/\.md$/, "") + const baseDir = this.settings.respectNewFileFolderPath + ? this.app.vault.getConfig("newFileFolderPath") + : undefined if (!this.candidateMap || !this.trie) { this.refreshFileDataAndTrie() @@ -531,16 +496,21 @@ export default class AutomaticLinkerPlugin extends Plugin { this.candidateMap, this.trie, this.settings, + normalizedActiveFilePath, + baseDir, ) const resultBody = replaceLinks({ body, linkResolverContext: { - filePath: activeFile.path, + filePath: normalizedActiveFilePath, trie: this.trie, candidateMap: this.candidateMap, }, - settings: this.settings, + settings: toReplaceLinksSettings( + this.settings, + baseDir, + ), resolvedAmbiguities: resolvedAmbiguitiesResult, }) diff --git a/src/markdown-protection.ts b/src/markdown-protection.ts new file mode 100644 index 0000000..8390241 --- /dev/null +++ b/src/markdown-protection.ts @@ -0,0 +1,2 @@ +export const RAW_URL_SOURCE = "(?:https?:\\/\\/|linear:\\/\\/)[^\\s]+" +export const RAW_URL_AT_START_PATTERN = new RegExp(`^(${RAW_URL_SOURCE})`) diff --git a/src/markdown-segments.ts b/src/markdown-segments.ts new file mode 100644 index 0000000..7dda3d4 --- /dev/null +++ b/src/markdown-segments.ts @@ -0,0 +1,309 @@ +import { RAW_URL_SOURCE } from "./markdown-protection" + +export type MarkdownSegmentKind = "prose" | "protected" + +export type MarkdownProtectedKind = "inline-code" + | "fenced-code" + | "wikilink" + | "markdown-link" + | "single-bracket" + | "url" + | "heading" + | "callout" + | "table-row" + +export interface MarkdownSegment { + kind: MarkdownSegmentKind + protectedKind?: MarkdownProtectedKind + start: number + end: number + text: string +} + +export interface SegmentMarkdownOptions { + protectHeadings?: boolean + protectCallouts?: boolean + protectTableRows?: boolean + protectUrls?: boolean +} + +interface ProtectedRange { + start: number + end: number + protectedKind: MarkdownProtectedKind +} + +export const isMarkdownTableLine = (line: string): boolean => { + const trimmedLine = line.trim() + if (!trimmedLine || !trimmedLine.includes("|")) { + return false + } + + if ( + trimmedLine.startsWith("|") + && trimmedLine.endsWith("|") + && /^[|:\s-]+$/.test(trimmedLine) + ) { + return true + } + + return trimmedLine.startsWith("|") && trimmedLine.endsWith("|") +} + +const collectHeadingRanges = (text: string): ProtectedRange[] => { + const ranges: ProtectedRange[] = [] + const headingPattern = /^#{1,6}\s+.*$/gm + let match: RegExpExecArray | null + + while ((match = headingPattern.exec(text)) !== null) { + ranges.push({ + start: match.index, + end: match.index + match[0].length, + protectedKind: "heading", + }) + } + + return ranges +} + +const collectCalloutRanges = (text: string): ProtectedRange[] => { + const ranges: ProtectedRange[] = [] + const calloutPattern = /^>[ \t]*\[![\w-]+\].*?(\n>.*?)*(?=\n(?!>)|$)/gm + let match: RegExpExecArray | null + + while ((match = calloutPattern.exec(text)) !== null) { + ranges.push({ + start: match.index, + end: match.index + match[0].length, + protectedKind: "callout", + }) + } + + return ranges +} + +const collectTableRowRanges = (text: string): ProtectedRange[] => { + const ranges: ProtectedRange[] = [] + const linePattern = /[^\n]*(?:\n|$)/g + let match: RegExpExecArray | null + + while ((match = linePattern.exec(text)) !== null) { + if (match[0] === "") { + break + } + + const lineText = match[0] + const lineContent = lineText.endsWith("\n") + ? lineText.slice(0, -1).replace(/\r$/, "") + : lineText.replace(/\r$/, "") + + if (isMarkdownTableLine(lineContent)) { + ranges.push({ + start: match.index, + end: match.index + lineText.length, + protectedKind: "table-row", + }) + } + } + + return ranges +} + +const escapeRegExp = (text: string): string => + text.replace(/[.*+?^${}()|[\]\\]/g, "\\$&") + +const collectFencedCodeRanges = (text: string): ProtectedRange[] => { + if (!text.includes("```") && !text.includes("~~~")) { + return [] + } + + const ranges: ProtectedRange[] = [] + const openingFencePattern = /^ {0,3}(`{3,}|~{3,})[^\r\n]*(?:\r?\n|$)/gm + let openingMatch: RegExpExecArray | null + + while ((openingMatch = openingFencePattern.exec(text)) !== null) { + const openingFence = openingMatch[1] + const fenceChar = openingFence[0] + const fenceLength = openingFence.length + const closingFencePattern = new RegExp( + `^ {0,3}${escapeRegExp(fenceChar)}{${fenceLength},}[ \\t]*(?:\\r?\\n|$)`, + "gm", + ) + closingFencePattern.lastIndex = openingMatch.index + openingMatch[0].length + const closingMatch = closingFencePattern.exec(text) + const end = closingMatch + ? closingMatch.index + closingMatch[0].length + : text.length + + ranges.push({ + start: openingMatch.index, + end, + protectedKind: "fenced-code", + }) + openingFencePattern.lastIndex = end + } + + return ranges +} + +const buildProtectedPattern = (protectUrls: boolean): RegExp => { + const parts = [ + "`[^`]*`", + "\\[\\[[^\\]]+\\]\\]", + "\\[[^\\]]+\\]\\([^)]+\\)", + "\\[[^\\]]+\\]", + ] + + if (protectUrls) { + parts.push(RAW_URL_SOURCE) + } + + return new RegExp(`(${parts.join("|")})`, "g") +} + +const getProtectedKind = (text: string): MarkdownProtectedKind => { + if (text.startsWith("```") || text.startsWith("~~~")) { + return "fenced-code" + } + + if (text.startsWith("`")) { + return "inline-code" + } + + if (text.startsWith("[[")) { + return "wikilink" + } + + if (text.startsWith("[")) { + return text.includes("](") ? "markdown-link" : "single-bracket" + } + + return "url" +} + +const sortAndMergeRanges = (ranges: ProtectedRange[]): ProtectedRange[] => { + const sortedRanges = ranges + .slice() + .sort((a, b) => a.start - b.start || a.end - b.end) + const merged: ProtectedRange[] = [] + + for (const range of sortedRanges) { + const lastRange = merged[merged.length - 1] + if (!lastRange || range.start >= lastRange.end) { + merged.push({ ...range }) + continue + } + + lastRange.end = Math.max(lastRange.end, range.end) + } + + return merged +} + +const isInsideRanges = ( + index: number, + ranges: ProtectedRange[], +): boolean => { + return ranges.some(range => index >= range.start && index < range.end) +} + +export const segmentMarkdown = ( + text: string, + options: SegmentMarkdownOptions = {}, +): MarkdownSegment[] => { + const mayContainProtectedMarkdown = text.includes("`") + || text.includes("~") + || text.includes("[") + || (options.protectHeadings && text.includes("#")) + || (options.protectCallouts && text.includes(">")) + || (options.protectTableRows && text.includes("|")) + || (options.protectUrls + && (text.includes("http") || text.includes("linear://"))) + + if (!mayContainProtectedMarkdown) { + return [{ + kind: "prose", + start: 0, + end: text.length, + text, + }] + } + + const ranges: ProtectedRange[] = [] + + if (options.protectHeadings) { + ranges.push(...collectHeadingRanges(text)) + } + + if (options.protectCallouts) { + ranges.push(...collectCalloutRanges(text)) + } + + if (options.protectTableRows) { + ranges.push(...collectTableRowRanges(text)) + } + + ranges.push(...collectFencedCodeRanges(text)) + + const protectedPattern = buildProtectedPattern(options.protectUrls ?? false) + let match: RegExpExecArray | null + + while ((match = protectedPattern.exec(text)) !== null) { + if (isInsideRanges(match.index, ranges)) { + continue + } + + ranges.push({ + start: match.index, + end: match.index + match[0].length, + protectedKind: getProtectedKind(match[0]), + }) + } + + const mergedRanges = sortAndMergeRanges(ranges) + const segments: MarkdownSegment[] = [] + let cursor = 0 + + for (const range of mergedRanges) { + if (cursor < range.start) { + segments.push({ + kind: "prose", + start: cursor, + end: range.start, + text: text.slice(cursor, range.start), + }) + } + + segments.push({ + kind: "protected", + protectedKind: range.protectedKind, + start: range.start, + end: range.end, + text: text.slice(range.start, range.end), + }) + cursor = range.end + } + + if (cursor < text.length || segments.length === 0) { + segments.push({ + kind: "prose", + start: cursor, + end: text.length, + text: text.slice(cursor), + }) + } + + return segments +} + +export const mapMarkdownProse = ( + text: string, + transform: (segmentText: string, segment: MarkdownSegment) => string, + options: SegmentMarkdownOptions = {}, +): string => { + return segmentMarkdown(text, options) + .map(segment => segment.kind === "prose" + ? transform(segment.text, segment) + : segment.text) + .join("") +} diff --git a/src/replace-links/__tests__/ai-disambiguation.test.ts b/src/replace-links/__tests__/ai-disambiguation.test.ts index 464f609..d5b67f2 100644 --- a/src/replace-links/__tests__/ai-disambiguation.test.ts +++ b/src/replace-links/__tests__/ai-disambiguation.test.ts @@ -1,6 +1,8 @@ import { describe, it, expect } from "vitest" import { replaceLinks } from "../replace-links" import { buildCandidateTrie } from "../../trie" +import { resolveAmbiguities } from "../../utils/resolve-ambiguities" +import { DEFAULT_SETTINGS } from "../../settings/settings-info" describe("replaceLinks with AI disambiguation", () => { const allFiles = [ @@ -53,4 +55,63 @@ describe("replaceLinks with AI disambiguation", () => { expect(result).toBe("Check [[work/meeting|private/meeting]] notes.") }) + + it("should honor AI-resolved Korean suffix choices", () => { + const koreanFiles = [ + { path: "work/λ¬Έμ„œ", scoped: false, aliases: [] }, + { path: "private/λ¬Έμ„œ", scoped: false, aliases: [] }, + ] + const { candidateMap: koreanCandidateMap, trie: koreanTrie } = buildCandidateTrie( + koreanFiles, + undefined, + true, + ) + const body = "λ¬Έμ„œμ΄λ‹€." + const resolvedAmbiguities = new Map([["λ¬Έμ„œ", "private/λ¬Έμ„œ"]]) + + const result = replaceLinks({ + body, + linkResolverContext: { + filePath: "test.md", + trie: koreanTrie, + candidateMap: koreanCandidateMap, + }, + resolvedAmbiguities, + }) + + expect(result).toBe("[[private/λ¬Έμ„œ|λ¬Έμ„œ]]이닀.") + }) + + it("should keep AI command self-link prevention when the active file path includes .md", async () => { + const body = "meeting" + const activeFilePath = "work/meeting.md" + const normalizedFilePath = activeFilePath.replace(/\.md$/, "") + const settings = { + ...DEFAULT_SETTINGS, + aiEnabled: true, + preventSelfLinking: true, + } + + const resolvedAmbiguities = await resolveAmbiguities( + body, + candidateMap, + trie, + settings, + normalizedFilePath, + ) + + const result = replaceLinks({ + body, + linkResolverContext: { + filePath: normalizedFilePath, + trie, + candidateMap, + }, + settings, + resolvedAmbiguities, + }) + + expect(resolvedAmbiguities.size).toBe(0) + expect(result).toBe("meeting") + }) }) diff --git a/src/replace-links/__tests__/candidate-scanner.test.ts b/src/replace-links/__tests__/candidate-scanner.test.ts new file mode 100644 index 0000000..4d9c9a3 --- /dev/null +++ b/src/replace-links/__tests__/candidate-scanner.test.ts @@ -0,0 +1,506 @@ +import { describe, expect, it } from "vitest" +import { buildTrie, CandidateData } from "../../trie" +import { buildCandidateTrieForTest } from "./test-helpers" +import { replaceLinks } from "../replace-links" +import { + getOccurrenceContext, + scanCandidateOccurrences, +} from "../candidate-scanner" + +describe("scanCandidateOccurrences", () => { + it("reports ambiguous prose candidates and skips inline code", () => { + const { candidateMap, trie } = buildCandidateTrieForTest({ + files: [ + { path: "work/meeting" }, + { path: "private/meeting" }, + ], + settings: { + scoped: false, + baseDir: undefined, + ignoreCase: true, + }, + }) + + const occurrences = scanCandidateOccurrences({ + text: "`meeting` meeting", + filePath: "notes/today", + trie, + candidateMap, + settings: { ignoreCase: true, proximityBasedLinking: true }, + }) + + expect(occurrences.map(o => ({ + kind: o.kind, + text: o.text, + start: o.start, + end: o.end, + candidates: o.candidateData.candidates.map(c => c.canonical), + }))).toEqual([ + { + kind: "unlinked", + text: "meeting", + start: 10, + end: 17, + candidates: ["work/meeting", "private/meeting"], + }, + ]) + }) + + it("reports existing wikilinks by their display alias", () => { + const { candidateMap, trie } = buildCandidateTrieForTest({ + files: [ + { path: "work/meeting" }, + { path: "private/meeting" }, + ], + settings: { + scoped: false, + baseDir: undefined, + ignoreCase: true, + }, + }) + + const occurrences = scanCandidateOccurrences({ + text: "Check [[private/meeting|meeting]] notes.", + filePath: "notes/today", + trie, + candidateMap, + settings: { ignoreCase: true, proximityBasedLinking: true }, + }) + + expect(occurrences.map(o => ({ + kind: o.kind, + text: o.text, + candidateKey: o.candidateKey, + candidates: o.candidateData.candidates.map(c => c.canonical), + }))).toEqual([ + { + kind: "existing-wikilink", + text: "[[private/meeting|meeting]]", + candidateKey: "meeting", + candidates: ["work/meeting", "private/meeting"], + }, + ]) + }) + + it("skips existing wikilinks inside inline code", () => { + const { candidateMap, trie } = buildCandidateTrieForTest({ + files: [ + { path: "work/meeting" }, + { path: "private/meeting" }, + ], + settings: { + scoped: false, + baseDir: undefined, + ignoreCase: true, + }, + }) + + const occurrences = scanCandidateOccurrences({ + text: "`[[private/meeting|meeting]]`", + filePath: "notes/today", + trie, + candidateMap, + settings: { ignoreCase: true, proximityBasedLinking: true }, + }) + + expect(occurrences).toEqual([]) + }) + + it("preserves trie-hit candidate sets for scoped candidates in the current namespace", () => { + const { candidateMap, trie } = buildCandidateTrieForTest({ + files: [ + { path: "pages/team-a/internal" }, + { path: "pages/team-b/internal" }, + ], + settings: { + scoped: true, + baseDir: "pages", + ignoreCase: true, + }, + }) + + const occurrences = scanCandidateOccurrences({ + text: "internal", + filePath: "pages/team-a/today", + trie, + candidateMap, + settings: { + baseDir: "pages", + ignoreCase: true, + proximityBasedLinking: true, + }, + }) + + expect(occurrences).toHaveLength(1) + expect(occurrences[0].candidateData.candidates.map(c => c.canonical)).toEqual([ + "pages/team-a/internal", + "pages/team-b/internal", + ]) + }) + + it("matches replaceLinks trie-hit namespace semantics", () => { + const candidateMap = new Map([ + [ + "internal", + { + candidates: [ + { + canonical: "pages/team-b/internal", + scoped: true, + namespace: "team-b", + }, + { + canonical: "pages/team-a/internal", + scoped: true, + namespace: "team-a", + }, + ], + }, + ], + ]) + const trie = buildTrie(["internal"], true) + + const occurrences = scanCandidateOccurrences({ + text: "internal", + filePath: "pages/team-a/today", + trie, + candidateMap, + settings: { + baseDir: "pages", + ignoreCase: true, + proximityBasedLinking: true, + }, + }) + + expect(occurrences).toEqual([]) + }) + + it("matches replaceLinks scoped namespace behavior when baseDir is set", () => { + const settings = { + scoped: true, + baseDir: "pages", + ignoreCase: true, + proximityBasedLinking: true, + } + const { candidateMap, trie } = buildCandidateTrieForTest({ + files: [ + { path: "pages/team-a/internal" }, + { path: "pages/team-a/archive/internal" }, + ], + settings, + }) + + const occurrences = scanCandidateOccurrences({ + text: "internal", + filePath: "pages/team-a/today", + trie, + candidateMap, + settings, + }) + const replaced = replaceLinks({ + body: "internal", + linkResolverContext: { + filePath: "pages/team-a/today", + trie, + candidateMap, + }, + settings, + }) + + expect(occurrences.map(o => ({ + text: o.text, + candidates: o.candidateData.candidates.map(c => c.canonical), + }))).toEqual([ + { + text: "internal", + candidates: [ + "pages/team-a/internal", + "pages/team-a/archive/internal", + ], + }, + ]) + expect(replaced).toBe("[[team-a/archive/internal|internal]]") + }) + + it("matches replaceLinks by protecting raw Linear URLs", () => { + const settings = { + scoped: false, + baseDir: undefined, + ignoreCase: true, + proximityBasedLinking: true, + } + const { candidateMap, trie } = buildCandidateTrieForTest({ + files: [ + { path: "work/linear" }, + { path: "private/linear" }, + ], + settings, + }) + const body = "Open linear://issue/TEAM-123 then linear" + const expectedStart = body.lastIndexOf("linear") + + const occurrences = scanCandidateOccurrences({ + text: body, + filePath: "notes/today", + trie, + candidateMap, + settings, + }) + const replaced = replaceLinks({ + body, + linkResolverContext: { + filePath: "notes/today", + trie, + candidateMap, + }, + settings, + }) + + expect(occurrences.map(o => ({ + text: o.text, + start: o.start, + end: o.end, + }))).toEqual([ + { + text: "linear", + start: expectedStart, + end: expectedStart + "linear".length, + }, + ]) + expect(replaced).toBe( + "Open linear://issue/TEAM-123 then [[private/linear|linear]]", + ) + }) + + it("skips fenced code blocks, callouts, and ignored headings", () => { + const { candidateMap, trie } = buildCandidateTrieForTest({ + files: [{ path: "meeting" }], + settings: { + scoped: false, + baseDir: undefined, + ignoreCase: true, + }, + }) + + const occurrences = scanCandidateOccurrences({ + text: `# meeting +> [!note] +> meeting + +~~~ts +meeting +~~~ + +meeting`, + filePath: "notes/today", + trie, + candidateMap, + settings: { + ignoreCase: true, + ignoreHeadings: true, + proximityBasedLinking: true, + }, + }) + + expect(occurrences.map(o => ({ + start: o.start, + end: o.end, + text: o.text, + }))).toEqual([ + { + start: 50, + end: 57, + text: "meeting", + }, + ]) + }) + + it("does not re-enter a later callout when inline code appears earlier", () => { + const { candidateMap, trie } = buildCandidateTrieForTest({ + files: [ + { path: "work/meeting" }, + { path: "private/meeting" }, + ], + settings: { + scoped: false, + baseDir: undefined, + ignoreCase: true, + }, + }) + + const occurrences = scanCandidateOccurrences({ + text: "`meeting`\n\n> [!note]\n> meeting\n\nmeeting", + filePath: "notes/today", + trie, + candidateMap, + settings: { + ignoreCase: true, + proximityBasedLinking: true, + }, + }) + + expect(occurrences.map(o => ({ + start: o.start, + end: o.end, + text: o.text, + }))).toEqual([ + { + start: 32, + end: 39, + text: "meeting", + }, + ]) + }) + + it("skips ignored Markdown table rows but still finds prose outside the table", () => { + const { candidateMap, trie } = buildCandidateTrieForTest({ + files: [ + { path: "work/meeting" }, + { path: "private/meeting" }, + ], + settings: { + scoped: false, + baseDir: undefined, + ignoreCase: true, + }, + }) + + const text = `| Topic | +| --- | +| meeting | + +meeting` + + const occurrences = scanCandidateOccurrences({ + text, + filePath: "notes/today", + trie, + candidateMap, + settings: { + ignoreCase: true, + ignoreMarkdownTables: true, + proximityBasedLinking: true, + }, + }) + + expect(occurrences.map(o => ({ + kind: o.kind, + text: o.text, + start: o.start, + end: o.end, + }))).toEqual([ + { + kind: "unlinked", + text: "meeting", + start: text.lastIndexOf("meeting"), + end: text.lastIndexOf("meeting") + "meeting".length, + }, + ]) + }) + + it("skips existing wikilinks inside ignored Markdown table rows", () => { + const { candidateMap, trie } = buildCandidateTrieForTest({ + files: [ + { path: "work/meeting" }, + { path: "private/meeting" }, + ], + settings: { + scoped: false, + baseDir: undefined, + ignoreCase: true, + }, + }) + + const occurrences = scanCandidateOccurrences({ + text: `| Topic | +| --- | +| [[private/meeting|meeting]] |`, + filePath: "notes/today", + trie, + candidateMap, + settings: { + ignoreCase: true, + ignoreMarkdownTables: true, + proximityBasedLinking: true, + }, + }) + + expect(occurrences).toEqual([]) + }) + + it("skips Korean particle hits but still finds overlapping follow-on candidates", () => { + const isolatedFixture = buildCandidateTrieForTest({ + files: [ + { path: "work/λ¬Έμ„œ" }, + { path: "private/λ¬Έμ„œ" }, + ], + settings: { + scoped: false, + baseDir: undefined, + ignoreCase: true, + }, + }) + const isolatedOccurrences = scanCandidateOccurrences({ + text: "λ¬Έμ„œλŠ”", + filePath: "notes/today", + trie: isolatedFixture.trie, + candidateMap: isolatedFixture.candidateMap, + settings: { ignoreCase: true, proximityBasedLinking: true }, + }) + + expect(isolatedOccurrences).toEqual([]) + + const overlappingFixture = buildCandidateTrieForTest({ + files: [ + { path: "work/λ¬Έμ„œ" }, + { path: "private/λ¬Έμ„œ" }, + { path: "work/μ„œλŠ”" }, + { path: "private/μ„œλŠ”" }, + ], + settings: { + scoped: false, + baseDir: undefined, + ignoreCase: true, + }, + }) + const { candidateMap, trie } = overlappingFixture + + const overlappingOccurrences = scanCandidateOccurrences({ + text: "λ¬Έμ„œλŠ”", + filePath: "notes/today", + trie, + candidateMap, + settings: { ignoreCase: true, proximityBasedLinking: true }, + }) + + expect(overlappingOccurrences.map(o => ({ + text: o.text, + start: o.start, + end: o.end, + }))).toEqual([ + { + text: "μ„œλŠ”", + start: 1, + end: 3, + }, + ]) + }) +}) + +describe("getOccurrenceContext", () => { + it("returns bounded surrounding text for AI requests", () => { + const occurrence = { + kind: "unlinked" as const, + start: 10, + end: 17, + text: "meeting", + candidateKey: "meeting", + candidateData: { candidates: [] }, + isInTable: false, + } + + expect(getOccurrenceContext("before -- meeting -- after", occurrence, 4)).toBe( + " -- meeting -- ", + ) + }) +}) diff --git a/src/replace-links/__tests__/replace-links.code.test.ts b/src/replace-links/__tests__/replace-links.code.test.ts index 2d6f0da..a16237b 100644 --- a/src/replace-links/__tests__/replace-links.code.test.ts +++ b/src/replace-links/__tests__/replace-links.code.test.ts @@ -45,6 +45,48 @@ describe("ignore code", () => { expect(result).toBe("```typescript\nexample\n```") }) + it("tilde code block", () => { + const settings = { + scoped: false, + baseDir: undefined, + } + const { candidateMap, trie } = buildCandidateTrieForTest({ + files: [{ path: "example" }, { path: "typescript" }], + settings, + }) + const result = replaceLinks({ + body: "~~~typescript\nexample\n~~~", + linkResolverContext: { + filePath: "journals/2022-01-01", + trie, + candidateMap, + }, + settings, + }) + expect(result).toBe("~~~typescript\nexample\n~~~") + }) + + it("wide code fence with embedded triple backticks", () => { + const settings = { + scoped: false, + baseDir: undefined, + } + const { candidateMap, trie } = buildCandidateTrieForTest({ + files: [{ path: "example" }, { path: "typescript" }], + settings, + }) + const result = replaceLinks({ + body: "````\nnot end ```\nexample\n````", + linkResolverContext: { + filePath: "journals/2022-01-01", + trie, + candidateMap, + }, + settings, + }) + expect(result).toBe("````\nnot end ```\nexample\n````") + }) + it("unclosed code block", () => { const settings = { scoped: false, diff --git a/src/replace-links/__tests__/replace-links.table.test.ts b/src/replace-links/__tests__/replace-links.table.test.ts index 4d98a83..9f6ae2d 100644 --- a/src/replace-links/__tests__/replace-links.table.test.ts +++ b/src/replace-links/__tests__/replace-links.table.test.ts @@ -108,6 +108,35 @@ note1 | Test Item | | | --- | --- | | [[note1]] | | +`) + }) + + it("escapes table aliases after earlier resolved wikilinks change segment offsets", () => { + const settings = { + scoped: false, + baseDir: undefined, + proximityBasedLinking: true, + } + const { candidateMap, trie } = buildCandidateTrieForTest({ + files: [{ path: "ns/note1" }], + settings, + }) + const result = replaceLinks({ + body: `[[note1]] +| note1 | | +`, + linkResolverContext: { + filePath: "journals/2022-01-01", + trie, + candidateMap, + }, + settings, + resolvedAmbiguities: new Map([ + ["[[note1]]", "very/long/path/note1|note1"], + ]), + }) + expect(result).toBe(`[[very/long/path/note1|note1]] +| [[ns/note1\\|note1]] | | `) }) }) diff --git a/src/replace-links/candidate-scanner.ts b/src/replace-links/candidate-scanner.ts new file mode 100644 index 0000000..2752de2 --- /dev/null +++ b/src/replace-links/candidate-scanner.ts @@ -0,0 +1,788 @@ +import { CandidateData, getTopLevelDirectoryName, TrieNode } from "../trie" +import { RAW_URL_AT_START_PATTERN, RAW_URL_SOURCE } from "../markdown-protection" +import { isMarkdownTableLine, segmentMarkdown } from "../markdown-segments" +import type { ReplaceLinksSettings } from "./replace-links" + +export type CandidateOccurrenceKind = "unlinked" | "existing-wikilink" + +export interface CandidateOccurrence { + kind: CandidateOccurrenceKind + start: number + end: number + text: string + candidateKey: string + candidateData: CandidateData + replacementCandidateData?: CandidateData + isInTable: boolean +} + +export type UnlinkedCandidateScanResult + = | { action: "match", occurrence: CandidateOccurrence } + | { action: "skip", end: number } + | null + +export interface ScanCandidateOccurrencesOptions { + text: string + filePath: string + trie: TrieNode + candidateMap: Map + settings?: ReplaceLinksSettings +} + +export const REGEX_PATTERNS = { + PROTECTED: new RegExp( + `(\`\`\`[\\s\\S]*?\`\`\`|\`[^\`]*\`|\\[\\[([^\\]]+)\\]\\]|\\[[^\\]]+\\]\\([^)]+\\)|\\[[^\\]]+\\]|${RAW_URL_SOURCE})`, + "g", + ), + DATE_FORMAT: /^\d{4}-\d{2}-\d{2}$/, + MONTH_NOTE: /^[0-9]{1,2}$/, + CJK: /[\p{Script=Han}\p{Script=Hiragana}\p{Script=Katakana}]/u, + CJK_CANDIDATE: /^[\p{Script=Han}\p{Script=Hiragana}\p{Script=Katakana}\p{Script=Hangul}\s\d]+$/u, + KOREAN: /^[\p{Script=Hangul}]+$/u, + JAPANESE: /^[\p{Script=Han}\p{Script=Hiragana}\p{Script=Katakana}\s\d]+$/u, + URL: RAW_URL_AT_START_PATTERN, + PROTECTED_LINK: /^\s*(\[\[[^\]]+\]\]|\[[^\]]+\]\([^)]+\))\s*$/, + KOREAN_SUFFIX: /^(이닀\.?)/, + KOREAN_PARTICLES: /^(λŠ”|은)/, + KOREAN_PARTICLES_EXTENDED: /^(κ°€|λŠ”|을|에|μ„œ|와|λ‘œλΆ€ν„°|κΉŒμ§€|보닀|둜|의|λ‚˜|도|λ˜ν•œ)/, + WORD_BOUNDARY: /[\p{L}\p{N}_/-]/u, + WHITESPACE: /[\t\n\r ]/, +} as const + +export const isWordBoundary = (char: string | undefined): boolean => { + if (char === undefined) return true + if (REGEX_PATTERNS.CJK.test(char)) return true + return (!REGEX_PATTERNS.WORD_BOUNDARY.test(char) || REGEX_PATTERNS.WHITESPACE.test(char)) +} + +export const isMonthNote = (candidate: string): boolean => + REGEX_PATTERNS.MONTH_NOTE.test(candidate) + && parseInt(candidate, 10) >= 1 + && parseInt(candidate, 10) <= 12 + +export const isProtectedLink = (body: string): boolean => REGEX_PATTERNS.PROTECTED_LINK.test(body) + +export const isCjkText = (text: string): boolean => REGEX_PATTERNS.CJK.test(text) + +export const isCjkCandidate = (candidate: string): boolean => REGEX_PATTERNS.CJK_CANDIDATE.test(candidate) + +export const isKoreanText = (text: string): boolean => REGEX_PATTERNS.KOREAN.test(text) + +export const isSentenceStart = (text: string, index: number): boolean => { + if (index === 0) return true + if (text[index - 1] === "\n") return true + if (index >= 3 && text[index - 1] === " " && text[index - 2] === ".") { + const charBeforePeriod = text[index - 3] + if (/[a-zA-Z]/.test(charBeforePeriod)) { + return true + } + } + return false +} + +const fallbackIndexCache = new WeakMap< + Map, + Map>> +>() + +export const buildFallbackIndex = ( + candidateMap: Map, + ignoreCase?: boolean, +): Map> => { + let cacheForMap = fallbackIndexCache.get(candidateMap) + if (!cacheForMap) { + cacheForMap = new Map() + fallbackIndexCache.set(candidateMap, cacheForMap) + } + + const cacheKey = ignoreCase ? "ignoreCase" : "normal" + const cached = cacheForMap.get(cacheKey) + if (cached) return cached + + const fallbackIndex = new Map>() + + for (const [key, data] of candidateMap.entries()) { + const slashIndex = key.lastIndexOf("/") + if (slashIndex === -1) continue + + const shorthand = key.slice(slashIndex + 1) + const indexKey = ignoreCase ? shorthand.toLowerCase() : shorthand + + let arr = fallbackIndex.get(indexKey) + if (!arr) { + arr = [] + fallbackIndex.set(indexKey, arr) + } + arr.push([key, data]) + } + + cacheForMap.set(cacheKey, fallbackIndex) + return fallbackIndex +} + +export const getCurrentNamespace = (filePath: string, baseDir?: string): string => { + if (baseDir) { + return getTopLevelDirectoryName(filePath, baseDir) + } + + const segments = filePath.split("/") + return segments[0] || "" +} + +export const normalizeCanonicalPath = (linkPath: string, baseDir?: string): string => { + if (baseDir && linkPath.startsWith(baseDir + "/")) { + return linkPath.slice((baseDir + "/").length) + } + return linkPath +} + +export const extractLinkParts = ( + canonicalPath: string, +): { linkPath: string, alias: string, hasAlias: boolean } => { + const pipeIndex = canonicalPath.indexOf("|") + const hasAlias = pipeIndex !== -1 + + if (hasAlias) { + const linkPath = canonicalPath.slice(0, pipeIndex) + const alias = canonicalPath.slice(pipeIndex + 1) + return { linkPath, alias, hasAlias } + } + + return { linkPath: canonicalPath, alias: "", hasAlias } +} + +export const escapeRegExp = (text: string): string => + text.replace(/[.*+?^${}()|[\]\\]/g, "\\$&") + +const findFencedCodeBlockRanges = ( + body: string, +): Array<{ start: number, end: number }> => { + if (!body.includes("```") && !body.includes("~~~")) { + return [] + } + + const ranges: Array<{ start: number, end: number }> = [] + const openingFencePattern = /^ {0,3}(`{3,}|~{3,})[^\r\n]*(?:\r?\n|$)/gm + let openingMatch: RegExpExecArray | null + + while ((openingMatch = openingFencePattern.exec(body)) !== null) { + const openingFence = openingMatch[1] + const fenceChar = openingFence[0] + const fenceLength = openingFence.length + const closingFencePattern = new RegExp( + `^ {0,3}${escapeRegExp(fenceChar)}{${fenceLength},}[ \\t]*(?:\\r?\\n|$)`, + "gm", + ) + closingFencePattern.lastIndex = openingMatch.index + openingMatch[0].length + const closingMatch = closingFencePattern.exec(body) + const end = closingMatch + ? closingMatch.index + closingMatch[0].length + : body.length + + ranges.push({ + start: openingMatch.index, + end, + }) + openingFencePattern.lastIndex = end + } + + return ranges +} + +export const extractFencedCodeBlocks = ( + body: string, +): { body: string, codeBlocks: Array<{ placeholder: string, content: string }> } => { + if (!body.includes("```") && !body.includes("~~~")) { + return { body, codeBlocks: [] } + } + + const ranges = findFencedCodeBlockRanges(body) + const codeBlocks: Array<{ placeholder: string, content: string }> = [] + let result = "" + let cursor = 0 + for (const [codeBlockIndex, range] of ranges.entries()) { + const placeholder = `__CODE_BLOCK_${codeBlockIndex}__` + codeBlocks.push({ + placeholder, + content: body.slice(range.start, range.end), + }) + result += body.slice(cursor, range.start) + placeholder + cursor = range.end + } + + result += body.slice(cursor) + return { body: result, codeBlocks } +} + +export const isSelfLink = ( + candidateData: CandidateData, + currentFilePath: string, + settings: ReplaceLinksSettings = {}, +): boolean => { + if (!settings.preventSelfLinking || candidateData.candidates.length === 0) { + return false + } + + const { linkPath } = extractLinkParts(candidateData.candidates[0].canonical) + const normalizedLinkPath = normalizeCanonicalPath( + linkPath, + settings.baseDir, + ) + const normalizedCurrentPath = normalizeCanonicalPath( + currentFilePath, + settings.baseDir, + ) + + return normalizedLinkPath === normalizedCurrentPath +} + +export const shouldSkipCandidate = ( + candidate: string, + settings: ReplaceLinksSettings, +): boolean => { + if ( + settings.ignoreDateFormats + && REGEX_PATTERNS.DATE_FORMAT.test(candidate) + ) { + return true + } + return isMonthNote(candidate) +} + +export const isIndexInsideMarkdownTable = (text: string, index: number): boolean => { + let lineStart = text.lastIndexOf("\n", index - 1) + 1 + if (lineStart === 0 && text[0] !== "\n") { + lineStart = 0 + } + + let lineEnd = text.indexOf("\n", index) + if (lineEnd === -1) { + lineEnd = text.length + } + + const line = text.slice(lineStart, lineEnd) + return isMarkdownTableLine(line) +} + +export const findBestCandidateInSameNamespace = ( + filteredCandidates: Array<[string, CandidateData]>, + filePath: string, + settings: ReplaceLinksSettings = {}, +): [string, CandidateData] | null => { + let bestCandidate: [string, CandidateData] | null = null + let bestScore = -1 + + const filePathDir = filePath.includes("/") + ? filePath.slice(0, filePath.lastIndexOf("/")) + : "" + const filePathSegments = filePathDir ? filePathDir.split("/") : [] + + for (const [key, data] of filteredCandidates) { + const slashIndex = key.lastIndexOf("/") + const candidateDir = key.slice(0, slashIndex) + const candidateSegments = candidateDir.split("/") + let score = 0 + + for ( + let idx = 0; + idx < Math.min(candidateSegments.length, filePathSegments.length); + idx++ + ) { + if (candidateSegments[idx] === filePathSegments[idx]) { + score++ + } + else { + break + } + } + + if (score > bestScore) { + bestScore = score + bestCandidate = [key, data] + } + else if (score === bestScore && bestCandidate !== null) { + if (filePathDir === "" && settings.baseDir) { + const basePrefix = settings.baseDir + "/" + const getRelativeDepth = (k: string): number => { + if (k.startsWith(basePrefix)) { + const relativeParts = k + .slice(basePrefix.length) + .split("/") + return relativeParts.length - 1 + } + return Infinity + } + + const candidateDepth = getRelativeDepth(key) + const bestCandidateDepth = getRelativeDepth(bestCandidate[0]) + + if (candidateDepth < bestCandidateDepth + || (candidateDepth === bestCandidateDepth && key.length < bestCandidate[0].length)) { + bestCandidate = [key, data] + } + } + else { + const currentBestDir = bestCandidate[0].slice(0, bestCandidate[0].lastIndexOf("/")) + const currentBestSegments = currentBestDir.split("/") + + if ( + candidateSegments.length < currentBestSegments.length + || (candidateSegments.length === currentBestSegments.length && key.length < bestCandidate[0].length) + ) { + bestCandidate = [key, data] + } + } + } + } + + return bestCandidate +} + +const dedupeCandidates = (candidates: CandidateData["candidates"]): CandidateData => { + const unique = new Map(candidates.map(candidate => [candidate.canonical, candidate])) + return { + candidates: Array.from(unique.values()).sort((a, b) => { + if (a.canonical.length !== b.canonical.length) { + return a.canonical.length - b.canonical.length + } + return a.canonical.localeCompare(b.canonical) + }), + } +} + +const collectExistingWikilinks = ( + text: string, + segments: ReturnType, + candidateMap: Map, + occurrences: CandidateOccurrence[], + settings: ReplaceLinksSettings, +): void => { + const existingLinkRegex = /^\[\[([^|\]]+)(?:\|([^\]]+))?\]\]$/ + + for (const segment of segments) { + if (segment.kind !== "protected" || segment.protectedKind !== "wikilink") { + continue + } + + const match = segment.text.match(existingLinkRegex) + if (!match) { + continue + } + + const fullMatch = match[0] + const path = match[1] + const alias = match[2] || path + const candidateKey = settings.ignoreCase ? alias.toLowerCase() : alias + const candidateData = candidateMap.get(candidateKey) ?? candidateMap.get(alias) + + if (!candidateData) { + continue + } + + occurrences.push({ + kind: "existing-wikilink", + start: segment.start, + end: segment.end, + text: fullMatch, + candidateKey, + candidateData: dedupeCandidates(candidateData.candidates), + isInTable: isIndexInsideMarkdownTable(text, segment.start), + }) + } +} + +const collectFallbackOccurrence = ({ + text, + startIndex, + fallbackIndex, + filePath, + currentNamespace, + settings, +}: { + text: string + startIndex: number + fallbackIndex: Map> + filePath: string + currentNamespace: string + settings: ReplaceLinksSettings +}): UnlinkedCandidateScanResult => { + const prevChar = text[startIndex - 1] + if (!isWordBoundary(prevChar)) { + return null + } + + let longestMatch: { + word: string + length: number + key: string + candidateList: Array<[string, CandidateData]> + } | null = null + + const maxSearchLength = Math.min(text.length - startIndex, 100) + + let potentialMatch = "" + let searchWord = "" + + for (let length = 1; length <= maxSearchLength; length++) { + const endIndex = startIndex + length + const currentChar = text[startIndex + length - 1] + potentialMatch += currentChar + + if (settings.matchSentenceCase && !settings.ignoreCase && isSentenceStart(text, startIndex)) { + if (length === 1) { + searchWord = currentChar.toLowerCase() + } + else { + searchWord += currentChar + } + } + else { + searchWord = settings.ignoreCase + ? searchWord + currentChar.toLowerCase() + : potentialMatch + } + + const candidateList = fallbackIndex.get(searchWord) + if (!candidateList) { + continue + } + + const nextChar = text[endIndex] + if (!isWordBoundary(nextChar)) { + continue + } + + if (shouldSkipCandidate(potentialMatch, settings)) { + continue + } + + longestMatch = { + word: potentialMatch, + length, + key: searchWord, + candidateList, + } + } + + if (!longestMatch) return null + + const filteredCandidates = longestMatch.candidateList.filter(([, data]) => { + if (data.candidates.length === 0 || !settings.proximityBasedLinking) { + return true + } + const candidate = data.candidates[0] + return !(candidate.scoped && candidate.namespace !== currentNamespace) + }) + + if (filteredCandidates.length === 0) { + return null + } + + const candidateData = dedupeCandidates( + filteredCandidates.flatMap(([, data]) => data.candidates), + ) + + if (candidateData.candidates.length === 0) { + return null + } + + const bestCandidateResult = filteredCandidates.length > 1 + ? findBestCandidateInSameNamespace(filteredCandidates, filePath, settings) + : filteredCandidates[0] + if (!bestCandidateResult) { + return null + } + + if (isSelfLink(bestCandidateResult[1], filePath, settings)) { + return { + action: "skip", + end: startIndex + longestMatch.length, + } + } + + return { + action: "match", + occurrence: { + kind: "unlinked", + start: startIndex, + end: startIndex + longestMatch.length, + text: longestMatch.word, + candidateKey: longestMatch.key, + candidateData, + replacementCandidateData: bestCandidateResult[1], + isInTable: isIndexInsideMarkdownTable(text, startIndex), + }, + } +} + +const shouldSkipKoreanTrieOccurrence = ( + text: string, + startIndex: number, + candidate: string, +): boolean => { + if (!isKoreanText(candidate)) { + return false + } + + const remaining = text.slice(startIndex + candidate.length) + return REGEX_PATTERNS.KOREAN_PARTICLES.test(remaining) +} + +export const scanUnlinkedCandidateAt = ({ + text, + startIndex, + filePath, + trie, + candidateMap, + fallbackIndex, + currentNamespace, + settings, +}: { + text: string + startIndex: number + filePath: string + trie: TrieNode + candidateMap: Map + fallbackIndex: Map> + currentNamespace: string + settings: ReplaceLinksSettings +}): UnlinkedCandidateScanResult => { + if ( + (text[startIndex] === "h" && text.slice(startIndex, startIndex + 4) === "http") + || (text[startIndex] === "l" && text.slice(startIndex, startIndex + 9) === "linear://") + ) { + const urlMatch = text.slice(startIndex).match(REGEX_PATTERNS.URL) + if (urlMatch) { + return { + action: "skip", + end: startIndex + urlMatch[0].length, + } + } + } + + let node = trie + let lastCandidate: { candidate: string, length: number } | null = null + let j = startIndex + let candidateBuilder = "" + + while (j < text.length) { + const ch = text[j] + let chLower = settings.ignoreCase ? ch.toLowerCase() : ch + if (settings.matchSentenceCase && !settings.ignoreCase && j === startIndex && isSentenceStart(text, startIndex)) { + chLower = ch.toLowerCase() + } + candidateBuilder += ch + + const child = node.children.get(chLower) + if (!child) break + + node = child + if (node.candidate) { + const candidateIsCjk = isCjkCandidate(candidateBuilder) + if (candidateIsCjk || isWordBoundary(text[j + 1])) { + lastCandidate = { + candidate: node.candidate, + length: j - startIndex + 1, + } + } + } + j++ + } + + if (lastCandidate) { + const candidate = candidateBuilder.slice(0, lastCandidate.length) + + if (shouldSkipCandidate(candidate, settings)) { + return { + action: "skip", + end: startIndex + lastCandidate.length, + } + } + + const trieCandidateKey = lastCandidate.candidate + const candidateData = candidateMap.get(trieCandidateKey) + + if (candidateData) { + if (isSelfLink(candidateData, filePath, settings)) { + return { + action: "skip", + end: startIndex + candidate.length, + } + } + + if (shouldSkipKoreanTrieOccurrence(text, startIndex, candidate)) { + return { + action: "skip", + end: startIndex + 1, + } + } + + const candidateIsCjk = isCjkCandidate(candidate) + if (!candidateIsCjk) { + const left = startIndex > 0 ? text[startIndex - 1] : undefined + const right = startIndex + candidate.length < text.length + ? text[startIndex + candidate.length] + : undefined + + if (!isWordBoundary(left) || !isWordBoundary(right)) { + return { + action: "skip", + end: startIndex + 1, + } + } + } + + if ( + settings.proximityBasedLinking + && candidateData.candidates.length > 0 + && candidateData.candidates[0].scoped + && candidateData.candidates[0].namespace !== currentNamespace + ) { + return { + action: "skip", + end: startIndex + candidate.length, + } + } + + return { + action: "match", + occurrence: { + kind: "unlinked", + start: startIndex, + end: startIndex + candidate.length, + text: candidate, + candidateKey: trieCandidateKey, + candidateData: dedupeCandidates(candidateData.candidates), + replacementCandidateData: candidateData, + isInTable: isIndexInsideMarkdownTable(text, startIndex), + }, + } + } + } + + if (settings.proximityBasedLinking) { + return collectFallbackOccurrence({ + text, + startIndex, + fallbackIndex, + filePath, + currentNamespace, + settings, + }) + } + + return null +} + +const collectUnlinkedOccurrences = ({ + text, + filePath, + trie, + candidateMap, + fallbackIndex, + currentNamespace, + settings, + occurrences, +}: { + text: string + filePath: string + trie: TrieNode + candidateMap: Map + fallbackIndex: Map> + currentNamespace: string + settings: ReplaceLinksSettings + occurrences: CandidateOccurrence[] +}): void => { + let i = 0 + + while (i < text.length) { + const result = scanUnlinkedCandidateAt({ + text, + startIndex: i, + filePath, + trie, + candidateMap, + fallbackIndex, + currentNamespace, + settings, + }) + + if (result?.action === "match") { + occurrences.push(result.occurrence) + i = result.occurrence.end + continue + } + + if (result?.action === "skip") { + i = result.end + continue + } + + i++ + } +} + +export const scanCandidateOccurrences = ({ + text, + filePath, + trie, + candidateMap, + settings = {}, +}: ScanCandidateOccurrencesOptions): CandidateOccurrence[] => { + const occurrences: CandidateOccurrence[] = [] + const normalizedText = text.normalize("NFC") + const fallbackIndex = buildFallbackIndex(candidateMap, settings.ignoreCase) + const currentNamespace = getCurrentNamespace(filePath, settings.baseDir) + const markdownSegments = segmentMarkdown(normalizedText, { + protectHeadings: settings.ignoreHeadings, + protectCallouts: true, + protectTableRows: settings.ignoreMarkdownTables, + protectUrls: true, + }) + + collectExistingWikilinks( + normalizedText, + markdownSegments, + candidateMap, + occurrences, + settings, + ) + + for (const segment of markdownSegments) { + if (segment.kind !== "prose") { + continue + } + + const segmentOccurrences: CandidateOccurrence[] = [] + collectUnlinkedOccurrences({ + text: segment.text, + filePath, + trie, + candidateMap, + fallbackIndex, + currentNamespace, + settings, + occurrences: segmentOccurrences, + }) + + for (const occurrence of segmentOccurrences) { + occurrences.push({ + ...occurrence, + start: occurrence.start + segment.start, + end: occurrence.end + segment.start, + }) + } + } + + return occurrences.sort((a, b) => a.start - b.start) +} + +export const getOccurrenceContext = ( + text: string, + occurrence: CandidateOccurrence, + maxContext: number, +): string => { + const start = Math.max(0, occurrence.start - maxContext) + const end = Math.min(text.length, occurrence.end + maxContext) + return text.slice(start, end) +} diff --git a/src/replace-links/replace-links.ts b/src/replace-links/replace-links.ts index daf892b..f798f02 100644 --- a/src/replace-links/replace-links.ts +++ b/src/replace-links/replace-links.ts @@ -1,4 +1,15 @@ -import { CandidateData, getTopLevelDirectoryName, TrieNode } from "../trie" +import { CandidateData, TrieNode } from "../trie" +import { mapMarkdownProse, segmentMarkdown } from "../markdown-segments" +import { + buildFallbackIndex, + extractLinkParts, + getCurrentNamespace, + isCjkText, + isIndexInsideMarkdownTable, + isProtectedLink, + normalizeCanonicalPath, + scanUnlinkedCandidateAt, +} from "./candidate-scanner" // Types for the replaceLinks function export interface LinkResolverContext { @@ -36,214 +47,6 @@ export interface ReplaceLinksOptions { resolvedAmbiguities?: Map } -// Constants and Regular Expressions -const REGEX_PATTERNS = { - PROTECTED: /(```[\s\S]*?```|`[^`]*`|\[\[([^\]]+)\]\]|\[[^\]]+\]\([^)]+\)|\[[^\]]+\]|https?:\/\/[^\s]+)/g, - DATE_FORMAT: /^\d{4}-\d{2}-\d{2}$/, - MONTH_NOTE: /^[0-9]{1,2}$/, - CJK: /[\p{Script=Han}\p{Script=Hiragana}\p{Script=Katakana}]/u, - CJK_CANDIDATE: /^[\p{Script=Han}\p{Script=Hiragana}\p{Script=Katakana}\p{Script=Hangul}\s\d]+$/u, - KOREAN: /^[\p{Script=Hangul}]+$/u, - JAPANESE: /^[\p{Script=Han}\p{Script=Hiragana}\p{Script=Katakana}\s\d]+$/u, - URL: /^(https?:\/\/[^\s]+)/, - PROTECTED_LINK: /^\s*(\[\[[^\]]+\]\]|\[[^\]]+\]\([^)]+\))\s*$/, - KOREAN_SUFFIX: /^(이닀\.?)/, - KOREAN_PARTICLES: /^(λŠ”|은)/, - KOREAN_PARTICLES_EXTENDED: /^(κ°€|λŠ”|을|에|μ„œ|와|λ‘œλΆ€ν„°|κΉŒμ§€|보닀|둜|의|λ‚˜|도|λ˜ν•œ)/, - TABLE_SEPARATOR: /^[|:\s-]+$/, - WORD_BOUNDARY: /[\p{L}\p{N}_/-]/u, - WHITESPACE: /[\t\n\r ]/, -} as const - -// Text Analysis Utilities -const isWordBoundary = (char: string | undefined): boolean => { - if (char === undefined) return true - if (REGEX_PATTERNS.CJK.test(char)) return true - return (!REGEX_PATTERNS.WORD_BOUNDARY.test(char) || REGEX_PATTERNS.WHITESPACE.test(char)) -} - -const isMonthNote = (candidate: string): boolean => - REGEX_PATTERNS.MONTH_NOTE.test(candidate) - && parseInt(candidate, 10) >= 1 - && parseInt(candidate, 10) <= 12 - -const isProtectedLink = (body: string): boolean => REGEX_PATTERNS.PROTECTED_LINK.test(body) - -const isCjkText = (text: string): boolean => REGEX_PATTERNS.CJK.test(text) - -const isCjkCandidate = (candidate: string): boolean => REGEX_PATTERNS.CJK_CANDIDATE.test(candidate) - -const isKoreanText = (text: string): boolean => REGEX_PATTERNS.KOREAN.test(text) - -const isSentenceStart = (text: string, index: number): boolean => { - if (index === 0) return true - if (text[index - 1] === "\n") return true - // ". " pattern (after period + space) - // Avoid acronyms by checking the char before the period is a letter - if (index >= 3 && text[index - 1] === " " && text[index - 2] === ".") { - const charBeforePeriod = text[index - 3] - if (/[a-zA-Z]/.test(charBeforePeriod)) { - return true - } - } - return false -} - -// Cache for fallback index to avoid rebuilding -const fallbackIndexCache = new WeakMap< - Map, - Map>> ->() - -const buildFallbackIndex = ( - candidateMap: Map, - ignoreCase?: boolean, -): Map> => { - // Get or create cache for this candidateMap - let cacheForMap = fallbackIndexCache.get(candidateMap) - if (!cacheForMap) { - cacheForMap = new Map() - fallbackIndexCache.set(candidateMap, cacheForMap) - } - - // Check if we have cached result for this ignoreCase setting - const cacheKey = ignoreCase ? "ignoreCase" : "normal" - const cached = cacheForMap.get(cacheKey) - if (cached) return cached - - // Build new fallback index - const fallbackIndex = new Map>() - - for (const [key, data] of candidateMap.entries()) { - const slashIndex = key.lastIndexOf("/") - if (slashIndex === -1) continue - - const shorthand = key.slice(slashIndex + 1) - const indexKey = ignoreCase ? shorthand.toLowerCase() : shorthand - - let arr = fallbackIndex.get(indexKey) - if (!arr) { - arr = [] - fallbackIndex.set(indexKey, arr) - } - arr.push([key, data]) - } - - // Cache the result - cacheForMap.set(cacheKey, fallbackIndex) - return fallbackIndex -} - -const getCurrentNamespace = (filePath: string, baseDir?: string): string => { - if (baseDir) { - return getTopLevelDirectoryName(filePath, baseDir) - } - - const segments = filePath.split("/") - return segments[0] || "" -} - -const normalizeCanonicalPath = (linkPath: string, baseDir?: string): string => { - if (baseDir && linkPath.startsWith(baseDir + "/")) { - return linkPath.slice((baseDir + "/").length) - } - return linkPath -} - -const extractLinkParts = ( - canonicalPath: string, -): { linkPath: string, alias: string, hasAlias: boolean } => { - const pipeIndex = canonicalPath.indexOf("|") - const hasAlias = pipeIndex !== -1 - - if (hasAlias) { - const linkPath = canonicalPath.slice(0, pipeIndex) - const alias = canonicalPath.slice(pipeIndex + 1) - return { linkPath, alias, hasAlias } - } - - return { linkPath: canonicalPath, alias: "", hasAlias } -} - -const escapeRegExp = (text: string): string => - text.replace(/[.*+?^${}()|[\]\\]/g, "\\$&") - -const extractFencedCodeBlocks = ( - body: string, -): { body: string, codeBlocks: Array<{ placeholder: string, content: string }> } => { - if (!body.includes("```") && !body.includes("~~~")) { - return { body, codeBlocks: [] } - } - - const codeBlocks: Array<{ placeholder: string, content: string }> = [] - let result = "" - let cursor = 0 - let codeBlockIndex = 0 - - const openingFencePattern = /^ {0,3}(`{3,}|~{3,})[^\r\n]*(?:\r?\n|$)/gm - let openingMatch: RegExpExecArray | null - - while ((openingMatch = openingFencePattern.exec(body)) !== null) { - const start = openingMatch.index - if (start < cursor) { - continue - } - - const openingFence = openingMatch[1] - const fenceChar = openingFence[0] - const fenceLength = openingFence.length - const closingFencePattern = new RegExp( - `^ {0,3}${escapeRegExp(fenceChar)}{${fenceLength},}[ \\t]*(?:\\r?\\n|$)`, - "gm", - ) - closingFencePattern.lastIndex = openingMatch.index + openingMatch[0].length - const closingMatch = closingFencePattern.exec(body) - const end = closingMatch - ? closingMatch.index + closingMatch[0].length - : body.length - - const placeholder = `__CODE_BLOCK_${codeBlockIndex}__` - codeBlocks.push({ - placeholder, - content: body.slice(start, end), - }) - result += body.slice(cursor, start) + placeholder - cursor = end - codeBlockIndex++ - openingFencePattern.lastIndex = end - } - - result += body.slice(cursor) - return { body: result, codeBlocks } -} - -// Self-linking Prevention -const isSelfLink = ( - candidateData: CandidateData, - currentFilePath: string, - settings: ReplaceLinksSettings = {}, -): boolean => { - if (!settings.preventSelfLinking || candidateData.candidates.length === 0) { - return false - } - - // Extract the link path from the canonical path - const { linkPath } = extractLinkParts(candidateData.candidates[0].canonical) - - // Normalize paths for comparison - const normalizedLinkPath = normalizeCanonicalPath( - linkPath, - settings.baseDir, - ) - const normalizedCurrentPath = normalizeCanonicalPath( - currentFilePath, - settings.baseDir, - ) - - // Compare the paths - return normalizedLinkPath === normalizedCurrentPath -} - // Helper function to check if a path should have its alias removed const shouldRemoveAlias = ( normalizedPath: string, @@ -345,6 +148,24 @@ const createLinkContent = ( } } +const resolveLinkContent = ( + candidateData: CandidateData, + originalMatchedText: string, + settings: ReplaceLinksSettings, + resolvedAmbiguities?: Map, +): { linkPath: string, alias?: string } => { + if (resolvedAmbiguities?.has(originalMatchedText)) { + const resolvedPath = resolvedAmbiguities.get(originalMatchedText)! + const parts = extractLinkParts(resolvedPath) + return { + linkPath: parts.linkPath, + alias: parts.alias || originalMatchedText, + } + } + + return createLinkContent(candidateData, originalMatchedText, settings) +} + // Default link generator that creates standard Obsidian wikilinks export const escapeLinkForMarkdownTable = ( link: string, @@ -371,55 +192,6 @@ export const defaultLinkGenerator: LinkGenerator = ({ return escapeLinkForMarkdownTable(`[[${linkContent}]]`, isInTable) } -// Candidate Validation -const shouldSkipCandidate = ( - candidate: string, - settings: ReplaceLinksSettings, -): boolean => { - if ( - settings.ignoreDateFormats - && REGEX_PATTERNS.DATE_FORMAT.test(candidate) - ) { - return true - } - return isMonthNote(candidate) -} - -// Markdown Table Detection -const isMarkdownTableLine = (line: string): boolean => { - const trimmedLine = line.trim() - if (!trimmedLine || !trimmedLine.includes("|")) { - return false - } - - if (trimmedLine.startsWith("|") - && trimmedLine.endsWith("|") - && REGEX_PATTERNS.TABLE_SEPARATOR.test(trimmedLine) - ) { - return true - } - - return trimmedLine.startsWith("|") && trimmedLine.endsWith("|") -} - -const isIndexInsideMarkdownTable = (text: string, index: number): boolean => { - // Find the start of the line containing the index - let lineStart = text.lastIndexOf("\n", index - 1) + 1 - if (lineStart === 0 && text[0] !== "\n") { - lineStart = 0 - } - - // Find the end of the line containing the index - let lineEnd = text.indexOf("\n", index) - if (lineEnd === -1) { - lineEnd = text.length - } - - // Extract the line and check if it's a table line - const line = text.slice(lineStart, lineEnd) - return isMarkdownTableLine(line) -} - // Processing functions for different text types const processCjkText = ( text: string, @@ -430,6 +202,7 @@ const processCjkText = ( linkGenerator: LinkGenerator, settings: ReplaceLinksSettings = {}, resolvedAmbiguities?: Map, + forceIsInTable?: boolean, ): string => { // For CJK texts that might contain non-CJK terms like "taro-san", ensure we use a consistent approach // Pass the proper filePath to maintain correct namespace resolution @@ -443,273 +216,10 @@ const processCjkText = ( linkGenerator, settings, resolvedAmbiguities, + forceIsInTable, ) } -// Fallback Search Processing -const processFallbackSearch = ( - text: string, - startIndex: number, - fallbackIndex: Map>, - filePath: string, - currentNamespace: string, - linkGenerator: LinkGenerator, - settings: ReplaceLinksSettings, -): { result: string, newIndex: number } | null => { - // Early boundary check - if start isn't a word boundary, skip - const prevChar = text[startIndex - 1] - if (!isWordBoundary(prevChar)) { - return null - } - - let longestMatch: { - word: string - length: number - key: string - candidateList: Array<[string, CandidateData]> - } | null = null - - // Iterate through potential multi-word sequences starting from startIndex - const maxSearchLength = Math.min(text.length - startIndex, 100) // Limit search length for performance - - let potentialMatch = "" - let searchWord = "" - - for (let length = 1; length <= maxSearchLength; length++) { - const endIndex = startIndex + length - const currentChar = text[startIndex + length - 1] - potentialMatch += currentChar - if (settings.matchSentenceCase && !settings.ignoreCase && isSentenceStart(text, startIndex)) { - if (length === 1) { - searchWord = currentChar.toLowerCase() - } - else { - searchWord = searchWord + currentChar - } - } - else { - searchWord = settings.ignoreCase - ? searchWord + currentChar.toLowerCase() - : potentialMatch - } - - // Check if this potential match exists in fallback index - const candidateList = fallbackIndex.get(searchWord) - if (!candidateList) { - continue - } - - // Basic boundary check: next char should be a boundary if not end of text - const nextChar = text[endIndex] - if (!isWordBoundary(nextChar)) { - continue // This isn't a valid match end - } - - // Skip date formats and month notes - if (shouldSkipCandidate(potentialMatch, settings)) { - continue // Try longer match - } - - // Found a valid candidate - longestMatch = { - word: potentialMatch, - length: length, - key: searchWord, - candidateList: candidateList, - } - // Continue checking for even longer matches - } - - // Process the longest valid match found - if (!longestMatch) return null - - // Filter candidates based on namespace restrictions - const filteredCandidates = longestMatch.candidateList.filter( - ([, data]) => { - if (data.candidates.length === 0) return true - const candidate = data.candidates[0] - return !(candidate.scoped && candidate.namespace !== currentNamespace) - }, - ) - - let bestCandidateData: CandidateData | null = null - - if (filteredCandidates.length === 1) { - bestCandidateData = filteredCandidates[0][1] - } - else if (filteredCandidates.length > 1) { - const bestCandidateResult = findBestCandidateInSameNamespace( - filteredCandidates, - filePath, - settings, - ) - if (bestCandidateResult) { - bestCandidateData = bestCandidateResult[1] - } - } - - if (!bestCandidateData) return null - - // Check if this is a self-link and should be prevented - if (isSelfLink(bestCandidateData, filePath, settings)) { - return { - result: longestMatch.word, - newIndex: startIndex + longestMatch.length, - } - } - - // Create the link - const { linkPath, alias } = createLinkContent( - bestCandidateData, - longestMatch.word, - settings, - ) - const isInTable = isIndexInsideMarkdownTable(text, startIndex) - const finalLink = linkGenerator({ - linkPath, - sourcePath: filePath, - alias, - isInTable, - }) - - return { - result: finalLink, - newIndex: startIndex + longestMatch.length, - } -} - -// Korean Language Processing -const handleKoreanSpecialCases = ( - text: string, - i: number, - candidate: string, - candidateData: CandidateData, - filePath: string, - linkGenerator: LinkGenerator, - settings: ReplaceLinksSettings = {}, -): { result: string, newIndex: number } | null => { - const remaining = text.slice(i + candidate.length) - - // Special handling when followed by "이닀" - const suffixMatch = remaining.match(REGEX_PATTERNS.KOREAN_SUFFIX) - if (suffixMatch) { - // Check if this is a self-link and should be prevented - if (isSelfLink(candidateData, filePath, settings)) { - return { - result: candidate + suffixMatch[0], - newIndex: i + candidate.length + suffixMatch[0].length, - } - } - - const { linkPath, alias } = createLinkContent( - candidateData, - candidate, - settings, - ) - const finalLink = linkGenerator({ - linkPath, - sourcePath: filePath, - alias, - isInTable: false, - }) - - return { - result: finalLink + suffixMatch[0], - newIndex: i + candidate.length + suffixMatch[0].length, - } - } - - // Special handling when followed by particles like "λŠ”" or "은" - if (remaining.match(REGEX_PATTERNS.KOREAN_PARTICLES)) { - return { - result: text[i], - newIndex: i + 1, - } - } - - return null -} - -const findBestCandidateInSameNamespace = ( - filteredCandidates: Array<[string, CandidateData]>, - filePath: string, - settings: ReplaceLinksSettings = {}, -): [string, CandidateData] | null => { - let bestCandidate: [string, CandidateData] | null = null - let bestScore = -1 - - // Get the directory portion of the current file (if any) - const filePathDir = filePath.includes("/") - ? filePath.slice(0, filePath.lastIndexOf("/")) - : "" - const filePathSegments = filePathDir ? filePathDir.split("/") : [] - - for (const [key, data] of filteredCandidates) { - const slashIndex = key.lastIndexOf("/") - const candidateDir = key.slice(0, slashIndex) - const candidateSegments = candidateDir.split("/") - let score = 0 - - // Calculate common prefix score - for ( - let idx = 0; - idx < Math.min(candidateSegments.length, filePathSegments.length); - idx++ - ) { - if (candidateSegments[idx] === filePathSegments[idx]) { - score++ - } - else { - break - } - } - - if (score > bestScore) { - bestScore = score - bestCandidate = [key, data] - } - else if (score === bestScore && bestCandidate !== null) { - if (filePathDir === "" && settings.baseDir) { - // When the current file is in the base directory, compare candidates by relative depth - const basePrefix = settings.baseDir + "/" - const getRelativeDepth = (k: string): number => { - if (k.startsWith(basePrefix)) { - // Remove the baseDir part and count the remaining segments - const relativeParts = k - .slice(basePrefix.length) - .split("/") - return relativeParts.length - 1 - } - return Infinity - } - - const candidateDepth = getRelativeDepth(key) - const bestCandidateDepth = getRelativeDepth(bestCandidate[0]) - - // Prefer the candidate with lower depth or shorter path - if (candidateDepth < bestCandidateDepth - || (candidateDepth === bestCandidateDepth && key.length < bestCandidate[0].length)) { - bestCandidate = [key, data] - } - } - else { - // Otherwise, choose the candidate with fewer directory segments - const currentBestDir = bestCandidate[0].slice(0, bestCandidate[0].lastIndexOf("/")) - const currentBestSegments = currentBestDir.split("/") - - if ( - candidateSegments.length < currentBestSegments.length - || (candidateSegments.length === currentBestSegments.length && key.length < bestCandidate[0].length) - ) { - bestCandidate = [key, data] - } - } - } - } - - return bestCandidate -} - const processStandardText = ( text: string, trie: TrieNode, @@ -720,203 +230,50 @@ const processStandardText = ( linkGenerator: LinkGenerator, settings: ReplaceLinksSettings = {}, resolvedAmbiguities?: Map, + forceIsInTable?: boolean, ): string => { let result = "" let i = 0 - outer: while (i < text.length) { - // Check for URLs first - only if current character could start a URL - if (text[i] === "h" && text.slice(i, i + 4) === "http") { - const urlMatch = text.slice(i).match(REGEX_PATTERNS.URL) - if (urlMatch) { - result += urlMatch[0] - i += urlMatch[0].length - continue - } + while (i < text.length) { + const scanResult = scanUnlinkedCandidateAt({ + text, + startIndex: i, + filePath, + trie, + candidateMap, + fallbackIndex, + currentNamespace, + settings, + }) + + if (scanResult?.action === "skip") { + result += text.slice(i, scanResult.end) + i = scanResult.end + continue } - // Try to find a candidate using the trie - let node = trie - let lastCandidate: { candidate: string, length: number } | null = null - let j = i - let candidateBuilder = "" - - while (j < text.length) { - const ch = text[j] - let chLower = settings.ignoreCase ? ch.toLowerCase() : ch - // At sentence start, lowercase only the first character to allow matching - if (settings.matchSentenceCase && !settings.ignoreCase && j === i && isSentenceStart(text, i)) { - chLower = ch.toLowerCase() - } - candidateBuilder += ch - - const child = node.children.get(chLower) - if (!child) break - - node = child - if (node.candidate) { - const candidateIsCjk = isCjkCandidate(candidateBuilder) - - if (candidateIsCjk || isWordBoundary(text[j + 1])) { - lastCandidate = { - candidate: node.candidate, - length: j - i + 1, - } - } - } - j++ - } - - if (lastCandidate) { - const candidate = candidateBuilder.slice(0, lastCandidate.length) - - // Skip if it's a date format - if ( - settings.ignoreDateFormats - && REGEX_PATTERNS.DATE_FORMAT.test(candidate) - ) { - result += candidate - i += lastCandidate.length - continue outer - } - - // Skip month notes - if (isMonthNote(candidate)) { - result += candidate - i += lastCandidate.length - continue - } - - // Use the candidate found in the trie (lastCandidate.candidate) to look up in candidateMap - const trieCandidateKey = lastCandidate.candidate - - // candidateMap lookup should always use the exact key from the trie result. - // Case comparison happened during trie traversal if ignoreCase is true. - const candidateData = candidateMap.get(trieCandidateKey) - - if (candidateData) { - // Check if this is a self-link and should be prevented - if (isSelfLink(candidateData, filePath, settings)) { - result += candidate - i += candidate.length - continue outer - } - - // Handle Korean special cases - const isKorean = isKoreanText(candidate) - if (isKorean) { - const koreanResult = handleKoreanSpecialCases( - text, - i, - candidate, - candidateData, - filePath, - linkGenerator, - settings, - ) - if (koreanResult) { - result += koreanResult.result - i = koreanResult.newIndex - continue outer - } - } - - // Word boundary check for non-CJK text - const candidateIsCjk = isCjkCandidate(candidate) - if (!candidateIsCjk) { - const left = i > 0 ? text[i - 1] : undefined - const right = i + candidate.length < text.length - ? text[i + candidate.length] - : undefined - - if (!isWordBoundary(left) || !isWordBoundary(right)) { - result += text[i] - i++ - continue outer - } - } - - // Check for Korean particles (extended) - const isKoreanCandidate = isKoreanText(candidate) - if (isKoreanCandidate) { - const right = i + candidate.length < text.length - ? text.slice( - i + candidate.length, - i + candidate.length + 10, - ) - : "" - - // Check for Korean particles (no action needed, just a check point) - if (right.match(REGEX_PATTERNS.KOREAN_PARTICLES_EXTENDED)) { - // Skip word boundary check for Korean particles - } - } - - // Skip if namespace restriction applies - if ( - settings.proximityBasedLinking - && candidateData.candidates.length > 0 - && candidateData.candidates[0].scoped - && candidateData.candidates[0].namespace !== currentNamespace - ) { - result += candidate - i += candidate.length - continue outer - } - - // Create the link - let linkPath = "" - let alias: string | undefined - - if (resolvedAmbiguities?.has(candidate)) { - const resolvedPath = resolvedAmbiguities.get(candidate)! - const parts = extractLinkParts(resolvedPath) - linkPath = parts.linkPath - alias = parts.alias || candidate - } - else { - const content = createLinkContent( - candidateData, - candidate, - settings, - ) - linkPath = content.linkPath - alias = content.alias - } - - const isInTable = isIndexInsideMarkdownTable(text, i) - const finalLink = linkGenerator({ - linkPath, - sourcePath: filePath, - alias, - isInTable, - }) - result += finalLink - - i += candidate.length - continue outer - } - } - - // Fallback: multi-word lookup using fallback index - if (settings.proximityBasedLinking) { - const fallbackResult = processFallbackSearch( - text, - i, - fallbackIndex, - filePath, - currentNamespace, - linkGenerator, + if (scanResult?.action === "match") { + const occurrence = scanResult.occurrence + const candidateData = occurrence.replacementCandidateData + ?? occurrence.candidateData + const { linkPath, alias } = resolveLinkContent( + candidateData, + occurrence.text, settings, + resolvedAmbiguities, ) - if (fallbackResult) { - result += fallbackResult.result - i = fallbackResult.newIndex - continue outer - } + const finalLink = linkGenerator({ + linkPath, + sourcePath: filePath, + alias, + isInTable: forceIsInTable ?? occurrence.isInTable, + }) + result += finalLink + i = occurrence.end + continue } - // If no rule applies, output the current character result += text[i] i++ } @@ -949,9 +306,45 @@ export const replaceLinks = ({ // Get the current namespace const currentNamespace = getCurrentNamespace(filePath, settings.baseDir) + let bodyWithResolvedWikilinks = body + + const markdownOptions = { + protectHeadings: settings.ignoreHeadings, + protectCallouts: true, + protectTableRows: settings.ignoreMarkdownTables, + protectUrls: true, + } + + const replaceResolvedWikilink = ( + wikilink: string, + start: number, + ): string => { + if (!resolvedAmbiguities?.has(wikilink)) { + return wikilink + } + + const resolvedPath = resolvedAmbiguities.get(wikilink)! + const { linkPath, alias: resolvedAlias } = extractLinkParts(resolvedPath) + const existingLinkRegex = /\[\[([^|\]]+)(?:\|([^\]]+))?\]\]/ + const linkMatch = wikilink.match(existingLinkRegex) + const existingPath = linkMatch ? linkMatch[1] : "" + const existingAlias = linkMatch ? linkMatch[2] : undefined + const finalAlias = resolvedAlias || existingAlias || (wikilink.includes("|") ? undefined : existingPath) + + return linkGenerator({ + linkPath, + sourcePath: filePath, + alias: finalAlias, + isInTable: !settings.ignoreMarkdownTables + && isIndexInsideMarkdownTable(body, start), + }) + } // Process segments of text - const processTextSegment = (text: string): string => { + const processTextSegment = ( + text: string, + forceIsInTable?: boolean, + ): string => { // Check if the text contains CJK characters const hasCjkText = isCjkText(text) @@ -965,6 +358,7 @@ export const replaceLinks = ({ linkGenerator, settings, resolvedAmbiguities, + forceIsInTable, ) } else { @@ -978,138 +372,60 @@ export const replaceLinks = ({ linkGenerator, settings, resolvedAmbiguities, + forceIsInTable, ) } } - const processTableAwareTextSegment = (text: string): string => { - if (!settings.ignoreMarkdownTables) { - return processTextSegment(text) + const processTableAwareTextSegment = ( + text: string, + segment: { start: number }, + ): string => { + if (!text.includes("\n")) { + const isInTable = !settings.ignoreMarkdownTables + && isIndexInsideMarkdownTable(bodyWithResolvedWikilinks, segment.start) + return processTextSegment(text, isInTable) } - return text.replace(/[^\n]*(?:\n|$)/g, (line) => { + return text.replace(/[^\n]*(?:\n|$)/g, (line, offset) => { if (line === "") { return line } const lineContent = line.endsWith("\n") - ? line.slice(0, -1) + ? line.slice(0, -1).replace(/\r$/, "") : line - if (isMarkdownTableLine(lineContent)) { + if (lineContent === "") { return line } - return processTextSegment(line) + const absoluteIndex = segment.start + offset + const isInTable = !settings.ignoreMarkdownTables + && isIndexInsideMarkdownTable(bodyWithResolvedWikilinks, absoluteIndex) + + return processTextSegment(line, isInTable) }) } - // Extract and protect fenced code blocks before any other block-level rules. - const { body: bodyAfterCodeBlocks, codeBlocks } = extractFencedCodeBlocks(body) + if (resolvedAmbiguities) { + bodyWithResolvedWikilinks = segmentMarkdown(body, markdownOptions) + .map((segment) => { + if ( + segment.kind === "protected" + && segment.protectedKind === "wikilink" + ) { + return replaceResolvedWikilink(segment.text, segment.start) + } - // Extract and protect headings first - const headingPattern = /^#{1,6}\s+.*$/gm - const headings: Array<{ placeholder: string, content: string }> = [] - let headingIndex = 0 - - let bodyAfterHeadings = bodyAfterCodeBlocks - if (settings.ignoreHeadings) { - bodyAfterHeadings = bodyAfterCodeBlocks.replace(headingPattern, (match) => { - const placeholder = `__HEADING_${headingIndex}__` - headings.push({ placeholder, content: match }) - headingIndex++ - return placeholder - }) - } - - // Extract and protect callout blocks first - // Match callout blocks: starts with > [!type] and continues with lines starting with > - const calloutPattern = /^>[ \t]*\[![\w-]+\].*?(\n>.*?)*(?=\n(?!>)|$)/gm - const callouts: Array<{ placeholder: string, content: string }> = [] - let calloutIndex = 0 - - // Replace callouts with placeholders - const bodyWithPlaceholders = bodyAfterHeadings.replace(calloutPattern, (match) => { - const placeholder = `__CALLOUT_${calloutIndex}__` - callouts.push({ placeholder, content: match }) - calloutIndex++ - return placeholder - }) - - // Process the entire body while preserving protected segments - let resultBody = "" - let lastIndex = 0 - let match: RegExpExecArray | null - - // Reset the regex to start from the beginning - REGEX_PATTERNS.PROTECTED.lastIndex = 0 - - while ( - (match = REGEX_PATTERNS.PROTECTED.exec(bodyWithPlaceholders)) !== null - ) { - const mIndex = match.index - const segment = bodyWithPlaceholders.slice(lastIndex, mIndex) - resultBody += processTableAwareTextSegment(segment) - - const fullMatch = match[0] - if ( - settings.ignoreMarkdownTables - && isIndexInsideMarkdownTable(bodyWithPlaceholders, mIndex) - ) { - resultBody += fullMatch - } - else if (resolvedAmbiguities?.has(fullMatch)) { - // Existing link replacement - const resolvedPath = resolvedAmbiguities.get(fullMatch)! - const { linkPath, alias: resolvedAlias } = extractLinkParts(resolvedPath) - - // Try to extract existing alias from the matched link - const existingLinkRegex = /\[\[([^|\]]+)(?:\|([^\]]+))?\]\]/ - const linkMatch = fullMatch.match(existingLinkRegex) - const existingPath = linkMatch ? linkMatch[1] : "" - const existingAlias = linkMatch ? linkMatch[2] : undefined - - // Use resolved alias if present, otherwise use existing alias, - // otherwise use existing path (as alias if it was a simple link) - const finalAlias = resolvedAlias || existingAlias || (fullMatch.includes("|") ? undefined : existingPath) - - const isInTable = isIndexInsideMarkdownTable(bodyWithPlaceholders, mIndex) - resultBody += linkGenerator({ - linkPath, - sourcePath: filePath, - alias: finalAlias, - isInTable, + return segment.text }) - } - else { - // Append the protected segment unchanged - resultBody += fullMatch - } - lastIndex = mIndex + fullMatch.length - - // Prevent infinite loop on zero-length matches - if (fullMatch.length === 0) { - REGEX_PATTERNS.PROTECTED.lastIndex++ - } + .join("") } - // Process the remaining text - resultBody += processTableAwareTextSegment(bodyWithPlaceholders.slice(lastIndex)) - - // Restore callouts - for (const { placeholder, content } of callouts) { - resultBody = resultBody.replace(placeholder, content) - } - - // Restore headings - for (const { placeholder, content } of headings) { - resultBody = resultBody.replace(placeholder, content) - } - - // Restore fenced code blocks - for (const { placeholder, content } of codeBlocks) { - resultBody = resultBody.replace(placeholder, content) - } - - return resultBody + return mapMarkdownProse( + bodyWithResolvedWikilinks, + processTableAwareTextSegment, + markdownOptions, + ) } diff --git a/src/replace-url-with-title/__tests__/replace-url-with-title.test.ts b/src/replace-url-with-title/__tests__/replace-url-with-title.test.ts index 7455cae..85e3178 100644 --- a/src/replace-url-with-title/__tests__/replace-url-with-title.test.ts +++ b/src/replace-url-with-title/__tests__/replace-url-with-title.test.ts @@ -63,4 +63,30 @@ describe("replaceUrlWithTitle", () => { "Line 1: [Example Title](https://example.com)\nLine 2: [Another Title](https://another.com)", ) }) + + it("should ignore URLs inside tilde fenced code blocks", () => { + const body = "~~~\nhttps://example.com\n~~~" + const result = replaceUrlWithTitle({ + body, + urlTitleMap: new Map( + Object.entries({ + "https://example.com": "Example Title", + }), + ), + }) + expect(result).toBe("~~~\nhttps://example.com\n~~~") + }) + + it("should not replace angle-bracket autolinks", () => { + const body = "" + const result = replaceUrlWithTitle({ + body, + urlTitleMap: new Map( + Object.entries({ + "https://example.com": "Example Title", + }), + ), + }) + expect(result).toBe("") + }) }) diff --git a/src/replace-url-with-title/index.ts b/src/replace-url-with-title/index.ts index 831f650..93d3ac0 100644 --- a/src/replace-url-with-title/index.ts +++ b/src/replace-url-with-title/index.ts @@ -1,3 +1,5 @@ +import { mapMarkdownProse } from "../markdown-segments" + type Url = string type Title = string interface ReplaceUrlWithTitleOptions { @@ -13,103 +15,77 @@ export const replaceUrlWithTitle = ({ return body } - let resultBody = body - // Sort URLs by length descending to replace longer URLs first // This helps prevent partial replacements (e.g., replacing 'example.com' before 'sub.example.com') const sortedUrls = Array.from(urlTitleMap.keys()).sort( (a, b) => b.length - a.length, ) - for (const url of sortedUrls) { - const title = urlTitleMap.get(url) - // Should not happen with Map iteration, but good practice - if (!title) continue + const replaceUrlsInProse = (prose: string): string => { + let resultBody = prose - // Escape backslashes and special characters in title for link text safety if needed - // For now, assume title is safe. - const markdownLink = `[${title}](${url})` - let currentIndex = 0 - const newBodyParts: string[] = [] + for (const url of sortedUrls) { + const title = urlTitleMap.get(url) + if (!title) continue - // Find all occurrences of the current URL in the resultBody - // resultBody is updated in each iteration of the outer loop - while (currentIndex < resultBody.length) { - // Find the next occurrence of the URL, case-sensitive. - const nextOccurrence = resultBody.indexOf(url, currentIndex) + const markdownLink = `[${title}](${url})` + let currentIndex = 0 + const newBodyParts: string[] = [] - if (nextOccurrence === -1) { - // No more occurrences found, add the rest of the string - newBodyParts.push(resultBody.substring(currentIndex)) - break - } + while (currentIndex < resultBody.length) { + const nextOccurrence = resultBody.indexOf(url, currentIndex) - // Add the text segment before the match - newBodyParts.push( - resultBody.substring(currentIndex, nextOccurrence), - ) + if (nextOccurrence === -1) { + newBodyParts.push(resultBody.substring(currentIndex)) + break + } - // --- Context Check --- - let shouldReplace = true + newBodyParts.push( + resultBody.substring(currentIndex, nextOccurrence), + ) - // 1. Check if already part of a Markdown link: [...](url) - // Look for `](` immediately before the URL and `)` immediately after. - const precedingChars = resultBody.substring( - nextOccurrence - 2, - nextOccurrence, - ) - const followingChar = resultBody[nextOccurrence + url.length] - if (precedingChars === "](" && followingChar === ")") { - shouldReplace = false - } + let shouldReplace = true - // 2. Check if inside inline code: `... url ...` - // Count non-escaped backticks before the match. Odd count means inside code. - if (shouldReplace) { - const segmentBefore = resultBody.substring(0, nextOccurrence) - // Count non-escaped backticks `(?") + ) { + shouldReplace = false + } - if (backticksCount % 2 !== 0) { - // Odd number of backticks means we might be inside a code span. - // We need to ensure the code span doesn't close before our match. - const lastBacktickIndex = segmentBefore.lastIndexOf("`") - // Check if there's another backtick between the last one and the match. - // If not, we are inside the code span. - if ( - lastBacktickIndex !== -1 - && !segmentBefore - .substring(lastBacktickIndex + 1) - .includes("`") - ) { - shouldReplace = false + if (shouldReplace) { + const segmentBefore = resultBody.substring(0, nextOccurrence) + const backticksCount = (segmentBefore.match(/(? { expect(result).not.toContain("https://example.com") }) + it("should ignore URLs inside tilde fenced code blocks", () => { + const body = "~~~\nhttps://example.com\n~~~" + const result = listupAllUrls(body) + expect(result).not.toContain("https://example.com") + }) + it("ignore domains", () => { const body = "Check this link: https://example.com" const result = listupAllUrls(body, ["example.com"]) diff --git a/src/replace-url-with-title/utils/list-up-all-urls.ts b/src/replace-url-with-title/utils/list-up-all-urls.ts index acaf9c7..7362516 100644 --- a/src/replace-url-with-title/utils/list-up-all-urls.ts +++ b/src/replace-url-with-title/utils/list-up-all-urls.ts @@ -1,3 +1,5 @@ +import { segmentMarkdown } from "../../markdown-segments" + type Url = string // Regular expression to find URLs starting with http:// or https:// @@ -16,159 +18,79 @@ export const listupAllUrls = ( const urls = new Set() let match - // --- Pre-calculate Fenced Code Block Ranges --- - const codeBlockRanges: { start: number, end: number }[] = [] - // Regex to find fenced code blocks (handles different fence lengths and optional language specifiers) - // Matches from ``` or ~~~ at the start of a line to the next ``` or ~~~ at the start of a line - const codeBlockRegex - = /^(?:```|~~~)[^\r\n]*?\r?\n([\s\S]*?)\r?\n^(?:```|~~~)$/gm - let blockMatch - while ((blockMatch = codeBlockRegex.exec(body)) !== null) { - codeBlockRanges.push({ - start: blockMatch.index, - end: blockMatch.index + blockMatch[0].length, - }) - } - // Reset regex state if needed, though new exec calls should handle this - codeBlockRegex.lastIndex = 0 + for (const segment of segmentMarkdown(body)) { + if (segment.kind === "protected") { + continue + } - while ((match = URL_REGEX.exec(body)) !== null) { - const url = match[0] - const matchIndex = match.index + URL_REGEX.lastIndex = 0 + while ((match = URL_REGEX.exec(segment.text)) !== null) { + const url = match[0] + const matchIndex = segment.start + match.index - // --- Fenced Code Block Check --- - // Check if the match index falls within any calculated code block range - let isInCodeBlock = false - for (const range of codeBlockRanges) { - if (matchIndex >= range.start && matchIndex < range.end) { - isInCodeBlock = true - break + let isBareUrl = true + + if (matchIndex >= 2) { + const followingCharIndex = matchIndex + url.length + if (followingCharIndex < body.length && body[followingCharIndex] === ")") { + const precedingChars = body.substring(matchIndex - 2, matchIndex) + if (precedingChars === "](") { + isBareUrl = false + } + } } - } - if (isInCodeBlock) { - continue // Skip this URL if it's inside a fenced code block - } - // --- Context Check --- - let isBareUrl = true - - // 1. Check if already part of a Markdown link: [...](url) - if (matchIndex >= 2) { // Need space for "](" - // Check if the URL is potentially followed by ')' - const followingCharIndex = matchIndex + url.length - if (followingCharIndex < body.length && body[followingCharIndex] === ")") { - // If followed by ')', check if preceded by "](" - const precedingChars = body.substring(matchIndex - 2, matchIndex) - if (precedingChars === "](") { - // Only if both conditions are met, it's a Markdown link + if (isBareUrl && matchIndex >= 1) { + const precedingChar = body[matchIndex - 1] + const followingChar = body[matchIndex + url.length] + if (precedingChar === "<" && followingChar === ">") { isBareUrl = false } } - } - // 2. Check if enclosed in angle brackets: - if (isBareUrl && matchIndex >= 1) { - const precedingChar = body[matchIndex - 1] - const followingChar = body[matchIndex + url.length] - if (precedingChar === "<" && followingChar === ">") { - isBareUrl = false - } - } + if (isBareUrl) { + let finalUrl = url + let shouldAdd = true - // 3. Check if inside inline code: `... url ...` - // Count non-escaped backticks before the match. Odd count means inside code. - if (isBareUrl) { - const segmentBefore = body.substring(0, matchIndex) - // Count non-escaped backticks `(? 0) { - try { - const parsedUrl = new URL(url) - const hostname = parsedUrl.hostname - if ( - ignoredDomains.some( - domain => - hostname === domain - || hostname.endsWith(`.${domain}`), + if (ignoredDomains && ignoredDomains.length > 0) { + try { + const parsedUrl = new URL(url) + const hostname = parsedUrl.hostname + if ( + ignoredDomains.some( + domain => + hostname === domain + || hostname.endsWith(`.${domain}`), + ) + ) { + shouldAdd = false + } + } + catch (e) { + console.warn( + `Failed to parse URL for domain check: ${url}`, + e, ) - ) { shouldAdd = false } } - catch (e) { - // If URL parsing fails, it's likely not a valid URL to add anyway - console.warn( - `Failed to parse URL for domain check: ${url}`, - e, - ) - shouldAdd = false - } - } - // 6. Clean Trailing Punctuation (only if not ignored) - // We only clean punctuation if the URL wasn't already excluded by context checks. - if (shouldAdd) { - const cleanedUrl = url.replace(TRAILING_PUNCTUATION_REGEX, "") - // Ensure cleaning didn't make it invalid (e.g., just "http://") - // or remove essential parts if the regex was too broad. - if (cleanedUrl.includes("://")) { - finalUrl = cleanedUrl // Use the cleaned URL only if it's still valid-looking + if (shouldAdd) { + const cleanedUrl = url.replace(TRAILING_PUNCTUATION_REGEX, "") + if (cleanedUrl.includes("://")) { + finalUrl = cleanedUrl + } + else { + finalUrl = url + console.warn(`URL cleaning potentially broke the URL: ${url} -> ${cleanedUrl}`) + } } - else { - // If cleaning resulted in an invalid URL, maybe don't add it, - // or reconsider the TRAILING_PUNCTUATION_REGEX. - // For now, let's stick with the original URL if cleaning fails badly. - // This case shouldn't happen often with the current regex. - finalUrl = url // Revert to original if cleaning broke it - console.warn(`URL cleaning potentially broke the URL: ${url} -> ${cleanedUrl}`) - // Decide if we should still add the original potentially dirty url - // shouldAdd = false; // Option: Don't add if cleaning failed - } - } - // --- Add URL if context checks pass and not ignored --- - if (shouldAdd) { - urls.add(finalUrl) + if (shouldAdd) { + urls.add(finalUrl) + } } } - - // Reset regex lastIndex to avoid issues with overlapping matches or zero-length matches - // Although our URL regex shouldn't produce zero-length matches. - // If the regex finds a match at index `i`, the next search starts at `i + 1`. - // If the match was length `l`, `exec` updates `lastIndex` to `i + l`. - // We need to ensure progress even if `l` is 0, but `URL_REGEX` won't match empty. - // If `isBareUrl` logic modified the string or indices, care would be needed. } return urls diff --git a/src/replace-urls/__tests__/replace-urls.test.ts b/src/replace-urls/__tests__/replace-urls.test.ts index d8a96c9..e6684b5 100644 --- a/src/replace-urls/__tests__/replace-urls.test.ts +++ b/src/replace-urls/__tests__/replace-urls.test.ts @@ -4,6 +4,7 @@ import { DEFAULT_SETTINGS, } from "../../settings/settings-info" import { formatGitHubURL } from "../github" +import { formatLinearURL } from "../linear" import { replaceURLs } from "../replace-urls" describe("replace-urls", () => { @@ -13,12 +14,71 @@ describe("replace-urls", () => { } it("should replace GitHub URLs", () => { - const input - = "[[github/kdnk/obsidian-automatic-linker]] [πŸ”—](https://github.com/kdnk/obsidian-automatic-linker)" + const input = "https://github.com/kdnk/obsidian-automatic-linker" const expected = "[[github/kdnk/obsidian-automatic-linker]] [πŸ”—](https://github.com/kdnk/obsidian-automatic-linker)" expect(replaceURLs(input, baseSettings, formatGitHubURL)).toBe( expected, ) }) + + it("does not rewrite https URLs inside wikilinks", () => { + const input = "[[https://github.com/kdnk/obsidian-automatic-linker]]" + + expect( + replaceURLs(input, baseSettings, formatGitHubURL), + ).toBe(input) + }) + + it("should replace linear protocol URLs", () => { + const input = "linear://workspace/issue/ACME-123" + const expected + = "[[linear/workspace/ACME-123]] [πŸ”—](https://linear.app/workspace/issue/ACME-123)" + + expect( + replaceURLs( + input, + { + ...baseSettings, + formatLinearURLs: true, + }, + formatLinearURL, + ), + ).toBe(expected) + }) + + it("does not rewrite linear protocol URLs inside wikilinks", () => { + const input = "[[linear://workspace/issue/ACME-123]]" + + expect( + replaceURLs( + input, + { + ...baseSettings, + formatLinearURLs: true, + }, + formatLinearURL, + ), + ).toBe(input) + }) + + it("keeps raw helper semantics inside Markdown links and angle autolinks", () => { + const input = [ + "[label](https://github.com/kdnk/obsidian-automatic-linker)", + "", + ].join(" ") + + const expected = [ + "[label](converted:https://github.com/kdnk/obsidian-automatic-linker)", + "", + ].join(" ") + + expect( + replaceURLs( + input, + baseSettings, + url => `converted:${url}`, + ), + ).toBe(expected) + }) }) diff --git a/src/replace-urls/__tests__/url-formatting.test.ts b/src/replace-urls/__tests__/url-formatting.test.ts new file mode 100644 index 0000000..f72ec0c --- /dev/null +++ b/src/replace-urls/__tests__/url-formatting.test.ts @@ -0,0 +1,145 @@ +import { describe, expect, it } from "vitest" +import { DEFAULT_SETTINGS } from "../../settings/settings-info" +import { + formatURLsInText, + formatURLWithAdapters, + UrlFormatter, +} from "../url-formatting" + +describe("formatURLWithAdapters", () => { + it("uses the first adapter that changes the URL", () => { + const first: UrlFormatter = url => `${url}-first` + const second: UrlFormatter = url => `${url}-second` + + expect( + formatURLWithAdapters( + "https://example.com", + DEFAULT_SETTINGS, + [first, second], + ), + ).toBe("https://example.com-first") + }) + + it("returns the original URL when no adapter changes it", () => { + const unchanged: UrlFormatter = url => url + + expect( + formatURLWithAdapters( + "https://example.com", + DEFAULT_SETTINGS, + [unchanged], + ), + ).toBe("https://example.com") + }) +}) + +describe("formatURLsInText", () => { + it("formats GitHub, Jira, and Linear URLs in one text pass", () => { + const result = formatURLsInText({ + text: [ + "https://github.com/owner/repo/issues/123", + "https://jira.company.com/browse/ABC-456", + "https://linear.app/team/issue/BUG-789/title", + "linear://team/issue/BUG-790", + ].join("\n"), + settings: { + ...DEFAULT_SETTINGS, + githubEnterpriseURLs: [], + jiraURLs: ["jira.company.com"], + formatGitHubURLs: true, + formatJiraURLs: true, + formatLinearURLs: true, + }, + }) + + expect(result).toBe([ + "[[github/owner/repo/issues/123]] [πŸ”—](https://github.com/owner/repo/issues/123)", + "[[company/jira/ABC/456]] [πŸ”—](https://jira.company.com/browse/ABC-456)", + "[[linear/team/BUG-789]] [πŸ”—](https://linear.app/team/issue/BUG-789)", + "[[linear/team/BUG-790]] [πŸ”—](https://linear.app/team/issue/BUG-790)", + ].join("\n")) + }) + + it("does not format GitHub URLs wrapped in angle brackets", () => { + const result = formatURLsInText({ + text: "", + settings: { + ...DEFAULT_SETTINGS, + formatGitHubURLs: true, + }, + }) + + expect(result).toBe("") + }) + + it("leaves disabled adapter URLs unchanged", () => { + const result = formatURLsInText({ + text: "https://linear.app/team/issue/BUG-789/title", + settings: { + ...DEFAULT_SETTINGS, + formatLinearURLs: false, + }, + }) + + expect(result).toBe("https://linear.app/team/issue/BUG-789/title") + }) + + it("leaves Jira URLs with query strings unchanged", () => { + const result = formatURLsInText({ + text: "https://jira.company.com/browse/ABC-456?focusedCommentId=12345", + settings: { + ...DEFAULT_SETTINGS, + jiraURLs: ["jira.company.com"], + formatJiraURLs: true, + }, + }) + + expect(result).toBe( + "https://jira.company.com/browse/ABC-456?focusedCommentId=12345", + ) + }) + + it("keeps closing parens and trailing punctuation outside formatted URLs", () => { + const result = formatURLsInText({ + text: "See (https://github.com/owner/repo/issues/123).", + settings: { + ...DEFAULT_SETTINGS, + formatGitHubURLs: true, + }, + }) + + expect(result).toBe( + "See ([[github/owner/repo/issues/123]] [πŸ”—](https://github.com/owner/repo/issues/123)).", + ) + }) + + it("does not format URLs inside protected Markdown segments", () => { + const result = formatURLsInText({ + text: [ + "`https://github.com/owner/repo/issues/123`", + "```md", + "https://github.com/owner/repo/issues/123", + "```", + "[label](https://github.com/owner/repo/issues/123)", + "[[https://github.com/owner/repo/issues/123]]", + "", + "https://github.com/owner/repo/issues/123", + ].join("\n"), + settings: { + ...DEFAULT_SETTINGS, + formatGitHubURLs: true, + }, + }) + + expect(result).toBe([ + "`https://github.com/owner/repo/issues/123`", + "```md", + "https://github.com/owner/repo/issues/123", + "```", + "[label](https://github.com/owner/repo/issues/123)", + "[[https://github.com/owner/repo/issues/123]]", + "", + "[[github/owner/repo/issues/123]] [πŸ”—](https://github.com/owner/repo/issues/123)", + ].join("\n")) + }) +}) diff --git a/src/replace-urls/github.ts b/src/replace-urls/github.ts index 015fc13..6f1484f 100644 --- a/src/replace-urls/github.ts +++ b/src/replace-urls/github.ts @@ -1,6 +1,4 @@ import { AutomaticLinkerSettings } from "../settings/settings-info" -import { formatJiraURL } from "./jira" -import { formatLinearURL } from "./linear" type GitHubURLInfo = { owner: string @@ -124,31 +122,3 @@ function isPullRequestOrIssueURL(url: URL): boolean { const path = url.pathname.toLowerCase() return path.includes("/pull/") || path.includes("/issues/") } - -export function formatURL( - url: string, - settings: AutomaticLinkerSettings, -): string { - if (settings.formatGitHubURLs) { - const formattedGitHubURL = formatGitHubURL(url, settings) - if (formattedGitHubURL !== url) { - return formattedGitHubURL - } - } - - if (settings.formatJiraURLs) { - const formattedJiraURL = formatJiraURL(url, settings) - if (formattedJiraURL !== url) { - return formattedJiraURL - } - } - - if (settings.formatLinearURLs) { - const formattedLinearURL = formatLinearURL(url, settings) - if (formattedLinearURL !== url) { - return formattedLinearURL - } - } - - return url -} diff --git a/src/replace-urls/replace-urls.ts b/src/replace-urls/replace-urls.ts index fabf103..5587d41 100644 --- a/src/replace-urls/replace-urls.ts +++ b/src/replace-urls/replace-urls.ts @@ -1,13 +1,39 @@ import { AutomaticLinkerSettings } from "../settings/settings-info" +const URL_PATTERN = /(?:https?:\/\/|linear:\/\/)[^\s<>\]]+/g +const WIKILINK_PATTERN = /\[\[[\s\S]*?\]\]/g + +const collectWikilinkRanges = (text: string): Array<{ start: number, end: number }> => { + const ranges: Array<{ start: number, end: number }> = [] + let match: RegExpExecArray | null + + while ((match = WIKILINK_PATTERN.exec(text)) !== null) { + ranges.push({ + start: match.index, + end: match.index + match[0].length, + }) + } + + return ranges +} + +const isInsideRange = ( + index: number, + ranges: Array<{ start: number, end: number }>, +): boolean => ranges.some(range => index >= range.start && index < range.end) + export const replaceURLs = ( fileContent: string, settings: AutomaticLinkerSettings, formatter: (url: string, settings: AutomaticLinkerSettings) => string, ) => { - const githubUrlPattern = /(? { + const wikilinkRanges = collectWikilinkRanges(fileContent) + + return fileContent.replace(URL_PATTERN, (match, offset) => { + if (isInsideRange(offset, wikilinkRanges)) { + return match + } + return formatter(match, settings) }) - return fileContent } diff --git a/src/replace-urls/url-formatting.ts b/src/replace-urls/url-formatting.ts new file mode 100644 index 0000000..e652547 --- /dev/null +++ b/src/replace-urls/url-formatting.ts @@ -0,0 +1,114 @@ +import { mapMarkdownProse } from "../markdown-segments" +import { AutomaticLinkerSettings } from "../settings/settings-info" +import { formatGitHubURL } from "./github" +import { formatJiraURL } from "./jira" +import { formatLinearURL } from "./linear" + +export type UrlFormatter = ( + url: string, + settings: AutomaticLinkerSettings, +) => string + +export interface FormatURLsInTextOptions { + text: string + settings: AutomaticLinkerSettings + formatters?: readonly UrlFormatter[] +} + +const URL_PATTERN = /(?:https?:\/\/|linear:\/\/)[^\s<>\]]+/g + +const formatGitHubURLIfEnabled: UrlFormatter = (url, settings) => + settings.formatGitHubURLs ? formatGitHubURL(url, settings) : url + +const formatJiraURLIfEnabled: UrlFormatter = (url, settings) => + settings.formatJiraURLs ? formatJiraURL(url, settings) : url + +const formatLinearURLIfEnabled: UrlFormatter = (url, settings) => + settings.formatLinearURLs ? formatLinearURL(url, settings) : url + +const TRAILING_PUNCTUATION = new Set([".", ",", ";", "!", "?", "]", "}"]) + +const countCharacter = (text: string, character: string): number => { + let count = 0 + + for (const currentCharacter of text) { + if (currentCharacter === character) { + count += 1 + } + } + + return count +} + +const splitTrailingBoundary = (match: string): { url: string, suffix: string } => { + let url = match + let suffix = "" + + while (url.length > 0) { + const lastCharacter = url[url.length - 1] + + if (TRAILING_PUNCTUATION.has(lastCharacter)) { + suffix = lastCharacter + suffix + url = url.slice(0, -1) + continue + } + + if ( + lastCharacter === ")" + && countCharacter(url, "(") < countCharacter(url, ")") + ) { + suffix = lastCharacter + suffix + url = url.slice(0, -1) + continue + } + + break + } + + return { url, suffix } +} + +export const DEFAULT_URL_FORMATTERS: readonly UrlFormatter[] = [ + formatGitHubURLIfEnabled, + formatJiraURLIfEnabled, + formatLinearURLIfEnabled, +] + +export const formatURLWithAdapters = ( + url: string, + settings: AutomaticLinkerSettings, + formatters: readonly UrlFormatter[] = DEFAULT_URL_FORMATTERS, +): string => { + for (const formatter of formatters) { + const formatted = formatter(url, settings) + if (formatted !== url) { + return formatted + } + } + + return url +} + +export const formatURLsInText = ({ + text, + settings, + formatters = DEFAULT_URL_FORMATTERS, +}: FormatURLsInTextOptions): string => + mapMarkdownProse( + text, + (prose, segment) => + prose.replace(URL_PATTERN, (match, offset) => { + const precedingCharacter = offset > 0 + ? segment.text[offset - 1] + : undefined + const followingCharacter = segment.text[offset + match.length] + + if (precedingCharacter === "<" && followingCharacter === ">") { + return match + } + + const { url, suffix } = splitTrailingBoundary(match) + + return formatURLWithAdapters(url, settings, formatters) + suffix + }), + ) diff --git a/src/settings/__tests__/settings-catalog.test.ts b/src/settings/__tests__/settings-catalog.test.ts new file mode 100644 index 0000000..cc4ae22 --- /dev/null +++ b/src/settings/__tests__/settings-catalog.test.ts @@ -0,0 +1,87 @@ +import { describe, expect, it } from "vitest" +import { + DEFAULT_SETTINGS, + SETTINGS_CATALOG, + projectReplaceLinksSettings, + projectUrlFormattingSettings, + settingRefreshesIndex, +} from "../settings-catalog" + +describe("SETTINGS_CATALOG", () => { + it("covers every default setting exactly once", () => { + const defaultKeys = Object.keys(DEFAULT_SETTINGS).sort() + const catalogKeys = SETTINGS_CATALOG.map(entry => entry.key).sort() + + expect(catalogKeys).toEqual(defaultKeys) + expect(new Set(catalogKeys).size).toBe(catalogKeys.length) + }) + + it("marks only the URL textarea settings for explicit sizing", () => { + const sizedTextareaKeys = SETTINGS_CATALOG + .filter(entry => entry.control === "textarea") + .filter(entry => (entry as { rows?: number }).rows === 4) + .filter(entry => (entry as { cols?: number }).cols === 50) + .map(entry => entry.key) + + expect(sizedTextareaKeys).toEqual([ + "replaceUrlWithTitleIgnoreDomains", + "githubEnterpriseURLs", + "jiraURLs", + ]) + }) + + it("marks current index-refresh settings", () => { + expect(settingRefreshesIndex("respectNewFileFolderPath")).toBe(true) + expect(settingRefreshesIndex("includeAliases")).toBe(true) + expect(settingRefreshesIndex("proximityBasedLinking")).toBe(true) + expect(settingRefreshesIndex("ignoreDateFormats")).toBe(true) + expect(settingRefreshesIndex("ignoreCase")).toBe(true) + expect(settingRefreshesIndex("preventSelfLinking")).toBe(true) + expect(settingRefreshesIndex("excludeDirsFromAutoLinking")).toBe(true) + expect(settingRefreshesIndex("removeAliasInDirs")).toBe(true) + expect(settingRefreshesIndex("debug")).toBe(false) + }) +}) + +describe("settings projections", () => { + it("projects link replacement settings", () => { + expect(projectReplaceLinksSettings({ + ...DEFAULT_SETTINGS, + proximityBasedLinking: false, + ignoreDateFormats: false, + ignoreCase: false, + matchSentenceCase: false, + preventSelfLinking: true, + removeAliasInDirs: ["archive"], + ignoreHeadings: true, + ignoreMarkdownTables: true, + }, "pages")).toEqual({ + proximityBasedLinking: false, + baseDir: "pages", + ignoreDateFormats: false, + ignoreCase: false, + matchSentenceCase: false, + preventSelfLinking: true, + removeAliasInDirs: ["archive"], + ignoreHeadings: true, + ignoreMarkdownTables: true, + }) + }) + + it("projects URL formatting settings", () => { + expect(projectUrlFormattingSettings({ + ...DEFAULT_SETTINGS, + formatGitHubURLs: false, + githubEnterpriseURLs: ["github.enterprise.com"], + formatJiraURLs: false, + jiraURLs: ["jira.example.com"], + formatLinearURLs: true, + })).toEqual({ + formatGitHubURLs: false, + githubEnterpriseURLs: ["github.enterprise.com"], + formatJiraURLs: false, + jiraURLs: ["jira.example.com"], + formatLinearURLs: true, + }) + }) +}) diff --git a/src/settings/settings-catalog.ts b/src/settings/settings-catalog.ts new file mode 100644 index 0000000..d8fe916 --- /dev/null +++ b/src/settings/settings-catalog.ts @@ -0,0 +1,388 @@ +import { ReplaceLinksSettings } from "../replace-links/replace-links" + +export type AutomaticLinkerSettings = { + formatOnSave: boolean + showNotice: boolean + respectNewFileFolderPath: boolean + includeAliases: boolean + proximityBasedLinking: boolean + ignoreDateFormats: boolean + ignoreHeadings: boolean + formatGitHubURLs: boolean + githubEnterpriseURLs: string[] + formatJiraURLs: boolean + jiraURLs: string[] + formatLinearURLs: boolean + debug: boolean + ignoreCase: boolean + matchSentenceCase: boolean + replaceUrlWithTitle: boolean + replaceUrlWithTitleIgnoreDomains: string[] + excludeDirsFromAutoLinking: string[] + preventSelfLinking: boolean + removeAliasInDirs: string[] + ignoreMarkdownTables: boolean + runLinterAfterFormatting: boolean + runPrettierAfterFormatting: boolean + formatDelayMs: number + aiEnabled: boolean + aiEndpoint: string + aiModel: string + aiMaxContext: number +} + +export type SettingControl = "toggle" | "text" | "textarea" + +export interface SettingCatalogEntry { + key: K + group: string + name: string + description: string + control: SettingControl + placeholder?: string + multiline?: boolean + rows?: number + cols?: number + refreshesIndex: boolean + runtimeOnly?: boolean +} + +export const DEFAULT_SETTINGS: AutomaticLinkerSettings = { + formatOnSave: false, + showNotice: false, + respectNewFileFolderPath: true, + includeAliases: true, + proximityBasedLinking: true, + ignoreDateFormats: true, + ignoreHeadings: false, + formatGitHubURLs: true, + githubEnterpriseURLs: [], + formatJiraURLs: true, + jiraURLs: [], + formatLinearURLs: false, + debug: false, + ignoreCase: true, + matchSentenceCase: true, + replaceUrlWithTitle: true, + replaceUrlWithTitleIgnoreDomains: [], + excludeDirsFromAutoLinking: [], + preventSelfLinking: false, + removeAliasInDirs: [], + ignoreMarkdownTables: false, + runLinterAfterFormatting: false, + runPrettierAfterFormatting: false, + formatDelayMs: 1, + aiEnabled: false, + aiEndpoint: "http://localhost:1234/v1", + aiModel: "gemma-4-7b", + aiMaxContext: 500, +} + +export const SETTINGS_CATALOG = [ + { + key: "formatOnSave", + group: "Formatting", + name: "Format on save", + description: + "When enabled, the file will be automatically formatted (links replaced) when saving.", + control: "toggle", + refreshesIndex: false, + }, + { + key: "respectNewFileFolderPath", + group: "Formatting", + name: "Respect 'Folder to create new notes in' setting", + description: + "When enabled, the plugin will use Obsidian's 'Folder to create new notes in' setting as the base directory for omitting folder prefixes in links.", + control: "toggle", + refreshesIndex: true, + }, + { + key: "includeAliases", + group: "Formatting", + name: "Include aliases", + description: + "When enabled, aliases will be included when processing links. Note: A restart is required for changes to take effect.", + control: "toggle", + refreshesIndex: true, + }, + { + key: "proximityBasedLinking", + group: "Formatting", + name: "Proximity-based linking", + description: + "When enabled, the plugin will automatically resolve namespaces for shorthand links. If multiple candidates share the same shorthand, the candidate with the most common path segments relative to the current file will be selected.", + control: "toggle", + refreshesIndex: true, + }, + { + key: "ignoreDateFormats", + group: "Formatting", + name: "Ignore date formats", + description: + "When enabled, links that match date formats (e.g. 2025-02-10) will be ignored. This helps maintain compatibility with Obsidian Tasks.", + control: "toggle", + refreshesIndex: true, + }, + { + key: "ignoreHeadings", + group: "Formatting", + name: "Ignore headings", + description: + "When enabled, headings (lines starting with #) will not have links added to them.", + control: "toggle", + refreshesIndex: false, + }, + { + key: "ignoreMarkdownTables", + group: "Formatting", + name: "Ignore Markdown tables", + description: + "When enabled, Markdown table rows will not have links added to them.", + control: "toggle", + refreshesIndex: false, + }, + { + key: "ignoreCase", + group: "Formatting", + name: "Ignore case", + description: + "When enabled, link matching will be case-insensitive. The original case of the text will be preserved in the link.", + control: "toggle", + refreshesIndex: true, + }, + { + key: "matchSentenceCase", + group: "Formatting", + name: "Match sentence case", + description: + "When 'Ignore case' is OFF, match text that is capitalized at the start of a sentence. For example, 'My name' at a sentence start will match the file 'my name'. This setting is ignored when 'Ignore case' is ON.", + control: "toggle", + refreshesIndex: false, + }, + { + key: "preventSelfLinking", + group: "Formatting", + name: "Prevent self-linking", + description: + "When enabled, text will not be linked to its own file. For example, the word 'VPN' inside VPN.md will not be converted to a link.", + control: "toggle", + refreshesIndex: true, + }, + { + key: "excludeDirsFromAutoLinking", + group: "Formatting", + name: "Exclude directories from automatic linking", + description: + "Directories to be excluded from automatic linking, one per line (e.g. 'Templates')", + control: "textarea", + placeholder: "Templates\nArchive", + multiline: true, + refreshesIndex: true, + }, + { + key: "removeAliasInDirs", + group: "Formatting", + name: "Remove aliases in directories", + description: + "Directories where link aliases should be removed, one per line (e.g. 'dir' will convert [[dir/xxx|yyy]] to [[dir/xxx]]). This affects both auto-generated aliases and frontmatter aliases.", + control: "textarea", + placeholder: "dir1\ndir2/subdir", + multiline: true, + refreshesIndex: true, + }, + { + key: "runLinterAfterFormatting", + group: "Integrations", + name: "Run Obsidian Linter after formatting", + description: + "When enabled, Obsidian Linter will be executed after Automatic Linker formatting. This requires the Obsidian Linter plugin to be installed.", + control: "toggle", + refreshesIndex: false, + }, + { + key: "runPrettierAfterFormatting", + group: "Integrations", + name: "Run Prettier after formatting", + description: + "When enabled, Prettier will be executed after Automatic Linker formatting. This requires prettier-format plugin to be installed. https://github.com/dylanarmstrong/obsidian-prettier-plugin", + control: "toggle", + refreshesIndex: false, + }, + { + key: "formatDelayMs", + group: "Integrations", + name: "Format delay (ms)", + description: + "Delay in milliseconds before formatting. Increase this value if the linter/prettier runs before the file is fully saved.", + control: "text", + placeholder: "e.g. 100", + refreshesIndex: false, + }, + { + key: "replaceUrlWithTitle", + group: "URL Replacement with Title", + name: "Replace URL with title", + description: + "When enabled, raw URLs will be replaced with [Page Title](URL). Requires fetching the URL content.", + control: "toggle", + refreshesIndex: false, + }, + { + key: "replaceUrlWithTitleIgnoreDomains", + group: "URL Replacement with Title", + name: "Ignore domains", + description: + "Ignore domains for replacing URLs with titles, one per line (e.g. x.com)", + control: "textarea", + placeholder: "", + multiline: true, + rows: 4, + cols: 50, + refreshesIndex: false, + }, + { + key: "formatGitHubURLs", + group: "URL Formatting for GitHub", + name: "Format GitHub URLs on save", + description: + "When enabled, GitHub URLs will be formatted when saving the file.", + control: "toggle", + refreshesIndex: false, + }, + { + key: "githubEnterpriseURLs", + group: "URL Formatting for GitHub", + name: "GitHub Enterprise URLs", + description: + "Add your GitHub Enterprise URLs, one per line (e.g. github.enterprise.com)", + control: "textarea", + placeholder: "github.enterprise.com\ngithub.company.com", + multiline: true, + rows: 4, + cols: 50, + refreshesIndex: false, + }, + { + key: "formatJiraURLs", + group: "URL Formatting for Jira", + name: "Format JIRA URLs on save", + description: + "When enabled, JIRA URLs will be formatted when saving the file.", + control: "toggle", + refreshesIndex: false, + }, + { + key: "jiraURLs", + group: "URL Formatting for Jira", + name: "JIRA URLs", + description: + "Add your JIRA URLs, one per line (e.g. jira.enterprise.com)", + control: "textarea", + placeholder: "jira.enterprise.com\njira.company.com", + multiline: true, + rows: 4, + cols: 50, + refreshesIndex: false, + }, + { + key: "formatLinearURLs", + group: "URL Formatting for Linear", + name: "Format Linear URLs on save", + description: + "When enabled, Linear URLs will be formatted when saving the file.", + control: "toggle", + refreshesIndex: false, + }, + { + key: "showNotice", + group: "Debug", + name: "Show load notice", + description: "Display a notice when markdown files are loaded.", + control: "toggle", + refreshesIndex: false, + }, + { + key: "debug", + group: "Debug", + name: "Debug mode", + description: + "When enabled, debug information will be logged to the console.", + control: "toggle", + refreshesIndex: false, + }, + { + key: "aiEnabled", + group: "AI Link Enhancement (Beta)", + name: "Enable AI Link Enhancement", + description: + "When enabled, an AI-powered link enhancer command will be available. It uses a local LLM to resolve ambiguous links and correct existing ones.", + control: "toggle", + refreshesIndex: false, + }, + { + key: "aiEndpoint", + group: "AI Link Enhancement (Beta)", + name: "AI API Endpoint", + description: + "The URL of your OpenAI-compatible AI server (e.g. LM Studio, Ollama).", + control: "text", + placeholder: "http://localhost:1234/v1", + refreshesIndex: false, + }, + { + key: "aiModel", + group: "AI Link Enhancement (Beta)", + name: "AI Model", + description: "The name of the model to use (e.g. gemma-4-7b).", + control: "text", + placeholder: "gemma-4-7b", + refreshesIndex: false, + }, + { + key: "aiMaxContext", + group: "AI Link Enhancement (Beta)", + name: "Max Context Length", + description: + "Number of characters around the link to provide as context to the AI.", + control: "text", + placeholder: "500", + refreshesIndex: false, + }, +] as const satisfies readonly SettingCatalogEntry[] + +export const settingRefreshesIndex = ( + key: keyof AutomaticLinkerSettings, +): boolean => SETTINGS_CATALOG.find(entry => entry.key === key)?.refreshesIndex ?? false + +export const projectReplaceLinksSettings = ( + settings: AutomaticLinkerSettings, + baseDir?: string, +): ReplaceLinksSettings => ({ + proximityBasedLinking: settings.proximityBasedLinking, + baseDir, + ignoreDateFormats: settings.ignoreDateFormats, + ignoreCase: settings.ignoreCase, + matchSentenceCase: settings.matchSentenceCase, + preventSelfLinking: settings.preventSelfLinking, + removeAliasInDirs: settings.removeAliasInDirs, + ignoreHeadings: settings.ignoreHeadings, + ignoreMarkdownTables: settings.ignoreMarkdownTables, +}) + +export const projectUrlFormattingSettings = ( + settings: AutomaticLinkerSettings, +): Pick< + AutomaticLinkerSettings, + | "formatGitHubURLs" + | "githubEnterpriseURLs" + | "formatJiraURLs" + | "jiraURLs" + | "formatLinearURLs" +> => ({ + formatGitHubURLs: settings.formatGitHubURLs, + githubEnterpriseURLs: settings.githubEnterpriseURLs, + formatJiraURLs: settings.formatJiraURLs, + jiraURLs: settings.jiraURLs, + formatLinearURLs: settings.formatLinearURLs, +}) diff --git a/src/settings/settings-info.ts b/src/settings/settings-info.ts index dd119e3..9748b54 100644 --- a/src/settings/settings-info.ts +++ b/src/settings/settings-info.ts @@ -1,61 +1,2 @@ -export type AutomaticLinkerSettings = { - formatOnSave: boolean - showNotice: boolean - respectNewFileFolderPath: boolean // Respect Obsidian's "Folder to create new notes in" setting as base directory - includeAliases: boolean // Include aliases when linking - proximityBasedLinking: boolean // Automatically resolve namespaces for shorthand links - ignoreDateFormats: boolean // Ignore date formatted links (e.g. 2025-02-10) - ignoreHeadings: boolean // Ignore headings (lines starting with #) when adding links - formatGitHubURLs: boolean // Format GitHub URLs on save - githubEnterpriseURLs: string[] // List of GitHub Enterprise URLs - formatJiraURLs: boolean // Format Jira URLs on save - jiraURLs: string[] // List of Jira URLs (domains) - formatLinearURLs: boolean // Format Linear URLs on save - debug: boolean // Enable debug logging - ignoreCase: boolean // Ignore case when matching links - matchSentenceCase: boolean // Match sentence-start capitalized text when ignoreCase is off - replaceUrlWithTitle: boolean // Replace raw URLs with [Title](URL) - replaceUrlWithTitleIgnoreDomains: string[] // List of domains to ignore when replacing URLs with titles - excludeDirsFromAutoLinking: string[] // Optional: List of directories to exclude from auto-linking - preventSelfLinking: boolean // Prevent linking text to its own file - removeAliasInDirs: string[] // Remove aliases for links in specified directories - ignoreMarkdownTables: boolean // Ignore Markdown table rows when adding links - runLinterAfterFormatting: boolean // Run Obsidian Linter after automatic linker formatting - runPrettierAfterFormatting: boolean // Run Prettier after automatic linker formatting - formatDelayMs: number // Delay in milliseconds before running linter after formatting - aiEnabled: boolean // Enable AI link enhancement - aiEndpoint: string // API endpoint for AI (OpenAI compatible) - aiModel: string // Model name for AI - aiMaxContext: number // Maximum context length for AI -} - -export const DEFAULT_SETTINGS: AutomaticLinkerSettings = { - formatOnSave: false, - showNotice: false, - respectNewFileFolderPath: true, // Default: do not respect Obsidian's "Folder to create new notes in" setting - includeAliases: true, // Default: include aliases - proximityBasedLinking: true, // Default: enable automatic namespace resolution - ignoreDateFormats: true, // Default: ignore date formats (e.g. 2025-02-10) - ignoreHeadings: false, // Default: do not ignore headings - formatGitHubURLs: true, // Default: format GitHub URLs - githubEnterpriseURLs: [], // Default: empty list - formatJiraURLs: true, // Default: format Jira URLs - jiraURLs: [], // Default: empty list - formatLinearURLs: false, // Default: format Linear URLs - debug: false, // Default: disable debug logging - ignoreCase: true, // Default: case-sensitive matching - matchSentenceCase: true, // Default: match sentence-start capitalized text - replaceUrlWithTitle: true, // Default: enable replacing URLs with titles - replaceUrlWithTitleIgnoreDomains: [], - excludeDirsFromAutoLinking: [], // Default: no excluded directories - preventSelfLinking: false, // Default: allow self-linking (backward compatibility) - removeAliasInDirs: [], // Default: no directories for alias removal - ignoreMarkdownTables: false, // Default: add links inside Markdown tables - runLinterAfterFormatting: false, // Default: do not run linter after formatting - runPrettierAfterFormatting: false, // Run Prettier after automatic linker formatting - formatDelayMs: 1, // Default: 1ms delay before running linter - aiEnabled: false, // Default: disable AI enhancement - aiEndpoint: "http://localhost:1234/v1", // Default: LM Studio - aiModel: "gemma-4-7b", // Default: Gemma 4 - aiMaxContext: 500, // Default: 500 characters -} +export type { AutomaticLinkerSettings } from "./settings-catalog" +export { DEFAULT_SETTINGS } from "./settings-catalog" diff --git a/src/settings/settings.ts b/src/settings/settings.ts index d414c9b..50a5238 100644 --- a/src/settings/settings.ts +++ b/src/settings/settings.ts @@ -1,5 +1,11 @@ import { App, PluginSettingTab, Setting } from "obsidian" import AutomaticLinkerPlugin from "../main" +import { + AutomaticLinkerSettings, + SETTINGS_CATALOG, + SettingCatalogEntry, + settingRefreshesIndex, +} from "./settings-catalog" export class AutomaticLinkerPluginSettingsTab extends PluginSettingTab { plugin: AutomaticLinkerPlugin @@ -8,500 +14,111 @@ export class AutomaticLinkerPluginSettingsTab extends PluginSettingTab { this.plugin = plugin } + private async setSettingValue( + key: K, + value: AutomaticLinkerSettings[K], + ) { + this.plugin.settings[key] = value + await this.plugin.saveData(this.plugin.settings) + if (key === "replaceUrlWithTitleIgnoreDomains") { + await this.plugin.buildUrlTitleMap() + } + if (settingRefreshesIndex(key)) { + this.plugin.refreshFileDataAndTrie() + } + } + + private parseTextValue( + key: K, + currentValue: AutomaticLinkerSettings[K], + nextValue: string, + ): AutomaticLinkerSettings[K] | null { + if (typeof currentValue !== "number") { + return nextValue as AutomaticLinkerSettings[K] + } + + const parsedValue = parseInt(nextValue) + if (isNaN(parsedValue)) { + return null + } + if (key === "formatDelayMs" && parsedValue < 0) { + return null + } + if (key === "aiMaxContext" && parsedValue <= 0) { + return null + } + + return parsedValue as AutomaticLinkerSettings[K] + } + + private renderSetting(containerEl: HTMLElement, entry: SettingCatalogEntry) { + const setting = new Setting(containerEl) + .setName(entry.name) + .setDesc(entry.description) + const value = this.plugin.settings[entry.key] + + if (entry.control === "toggle") { + setting.addToggle((toggle) => { + toggle + .setValue(Boolean(value)) + .onChange(async (nextValue) => { + await this.setSettingValue(entry.key, nextValue as never) + }) + }) + return + } + + if (entry.control === "text") { + setting.addText((text) => { + text.setPlaceholder(entry.placeholder ?? "") + .setValue(String(value)) + .onChange(async (nextValue) => { + const parsedValue = this.parseTextValue( + entry.key, + value, + nextValue, + ) + if (parsedValue === null) { + return + } + await this.setSettingValue(entry.key, parsedValue) + }) + }) + return + } + + setting.addTextArea((text) => { + text.setPlaceholder(entry.placeholder ?? "") + .setValue(Array.isArray(value) ? value.join("\n") : String(value)) + .onChange(async (nextValue) => { + await this.setSettingValue( + entry.key, + nextValue + .split("\n") + .map(item => item.trim()) + .filter(Boolean) as never, + ) + }) + if (entry.rows !== undefined) { + text.inputEl.rows = entry.rows + } + if (entry.cols !== undefined) { + text.inputEl.cols = entry.cols + } + }) + } + display(): void { const { containerEl } = this containerEl.empty() - new Setting(containerEl).setName("Formatting").setHeading() - - // Toggle for "Format on Save" setting. - new Setting(containerEl) - .setName("Format on save") - .setDesc( - "When enabled, the file will be automatically formatted (links replaced) when saving.", - ) - .addToggle((toggle) => { - toggle - .setValue(this.plugin.settings.formatOnSave) - .onChange(async (value) => { - this.plugin.settings.formatOnSave = value - await this.plugin.saveData(this.plugin.settings) - }) - }) - - // Toggle for respecting Obsidian's "Folder to create new notes in" setting - new Setting(containerEl) - .setName("Respect 'Folder to create new notes in' setting") - .setDesc( - "When enabled, the plugin will use Obsidian's 'Folder to create new notes in' setting as the base directory for omitting folder prefixes in links.", - ) - .addToggle((toggle) => { - toggle - .setValue(this.plugin.settings.respectNewFileFolderPath) - .onChange(async (value) => { - this.plugin.settings.respectNewFileFolderPath = value - await this.plugin.saveData(this.plugin.settings) - this.plugin.refreshFileDataAndTrie() - }) - }) - - // Toggle for including aliases. - new Setting(containerEl) - .setName("Include aliases") - .setDesc( - "When enabled, aliases will be included when processing links. Note: A restart is required for changes to take effect.", - ) - .addToggle((toggle) => { - toggle - .setValue(this.plugin.settings.includeAliases) - .onChange(async (value) => { - this.plugin.settings.includeAliases = value - await this.plugin.saveData(this.plugin.settings) - this.plugin.refreshFileDataAndTrie() - }) - }) - - // Toggle for proximity-based linking. - new Setting(containerEl) - .setName("Proximity-based linking") - .setDesc( - "When enabled, the plugin will automatically resolve namespaces for shorthand links. If multiple candidates share the same shorthand, the candidate with the most common path segments relative to the current file will be selected.", - ) - .addToggle((toggle) => { - toggle - .setValue(this.plugin.settings.proximityBasedLinking) - .onChange(async (value) => { - this.plugin.settings.proximityBasedLinking = value - await this.plugin.saveData(this.plugin.settings) - this.plugin.refreshFileDataAndTrie() - }) - }) - - // Toggle for ignoring date formats. - new Setting(containerEl) - .setName("Ignore date formats") - .setDesc( - "When enabled, links that match date formats (e.g. 2025-02-10) will be ignored. This helps maintain compatibility with Obsidian Tasks.", - ) - .addToggle((toggle) => { - toggle - .setValue(this.plugin.settings.ignoreDateFormats) - .onChange(async (value) => { - this.plugin.settings.ignoreDateFormats = value - await this.plugin.saveData(this.plugin.settings) - this.plugin.refreshFileDataAndTrie() - }) - }) - - // Toggle for ignoring headings - new Setting(containerEl) - .setName("Ignore headings") - .setDesc( - "When enabled, headings (lines starting with #) will not have links added to them.", - ) - .addToggle((toggle) => { - toggle - .setValue(this.plugin.settings.ignoreHeadings) - .onChange(async (value) => { - this.plugin.settings.ignoreHeadings = value - await this.plugin.saveData(this.plugin.settings) - }) - }) - - // Toggle for ignoring Markdown tables - new Setting(containerEl) - .setName("Ignore Markdown tables") - .setDesc( - "When enabled, Markdown table rows will not have links added to them.", - ) - .addToggle((toggle) => { - toggle - .setValue(this.plugin.settings.ignoreMarkdownTables) - .onChange(async (value) => { - this.plugin.settings.ignoreMarkdownTables = value - await this.plugin.saveData(this.plugin.settings) - }) - }) - - // Toggle for ignoring case in link matching - new Setting(containerEl) - .setName("Ignore case") - .setDesc( - "When enabled, link matching will be case-insensitive. The original case of the text will be preserved in the link.", - ) - .addToggle((toggle) => { - toggle - .setValue(this.plugin.settings.ignoreCase) - .onChange(async (value) => { - this.plugin.settings.ignoreCase = value - await this.plugin.saveData(this.plugin.settings) - this.plugin.refreshFileDataAndTrie() - }) - }) - - // Toggle for matching sentence-case text - new Setting(containerEl) - .setName("Match sentence case") - .setDesc( - "When 'Ignore case' is OFF, match text that is capitalized at the start of a sentence. For example, 'My name' at a sentence start will match the file 'my name'. This setting is ignored when 'Ignore case' is ON.", - ) - .addToggle((toggle) => { - toggle - .setValue(this.plugin.settings.matchSentenceCase) - .onChange(async (value) => { - this.plugin.settings.matchSentenceCase = value - await this.plugin.saveData(this.plugin.settings) - }) - }) - - // Toggle for preventing self-linking - new Setting(containerEl) - .setName("Prevent self-linking") - .setDesc( - "When enabled, text will not be linked to its own file. For example, the word 'VPN' inside VPN.md will not be converted to a link.", - ) - .addToggle((toggle) => { - toggle - .setValue(this.plugin.settings.preventSelfLinking) - .onChange(async (value) => { - this.plugin.settings.preventSelfLinking = value - await this.plugin.saveData(this.plugin.settings) - this.plugin.refreshFileDataAndTrie() - }) - }) - - // Add excluding dirs that you wish to exclude from the automatic linking - new Setting(containerEl) - .setName("Exclude directories from automatic linking") - .setDesc( - "Directories to be excluded from automatic linking, one per line (e.g. 'Templates')", - ) - .addTextArea((text) => { - text.setPlaceholder("Templates\nArchive") - .setValue( - this.plugin.settings.excludeDirsFromAutoLinking.join( - "\n", - ), - ) - .onChange(async (value) => { - // Split by newlines and filter out empty lines - const dirs = value.split("\n").map(dir => dir.trim()).filter(Boolean) - this.plugin.settings.excludeDirsFromAutoLinking = dirs - await this.plugin.saveData(this.plugin.settings) - this.plugin.refreshFileDataAndTrie() - }) - }) - - // Remove aliases for links in specified directories - new Setting(containerEl) - .setName("Remove aliases in directories") - .setDesc( - "Directories where link aliases should be removed, one per line (e.g. 'dir' will convert [[dir/xxx|yyy]] to [[dir/xxx]]). This affects both auto-generated aliases and frontmatter aliases.", - ) - .addTextArea((text) => { - text.setPlaceholder("dir1\ndir2/subdir") - .setValue(this.plugin.settings.removeAliasInDirs.join("\n")) - .onChange(async (value) => { - // Split by newlines and filter out empty lines - const dirs = value - .split("\n") - .map(dir => dir.trim()) - .filter(Boolean) - this.plugin.settings.removeAliasInDirs = dirs - await this.plugin.saveData(this.plugin.settings) - this.plugin.refreshFileDataAndTrie() - }) - }) - - new Setting(containerEl) - .setName("Integrations") - .setHeading() - - // Toggle for running linter after formatting - new Setting(containerEl) - .setName("Run Obsidian Linter after formatting") - .setDesc( - "When enabled, Obsidian Linter will be executed after Automatic Linker formatting. This requires the Obsidian Linter plugin to be installed.", - ) - .addToggle((toggle) => { - toggle - .setValue(this.plugin.settings.runLinterAfterFormatting) - .onChange(async (value) => { - this.plugin.settings.runLinterAfterFormatting = value - await this.plugin.saveData(this.plugin.settings) - }) - }) - - // Toggle for running prettier after formatting - new Setting(containerEl) - .setName("Run Prettier after formatting") - .setDesc( - "When enabled, Prettier will be executed after Automatic Linker formatting. This requires prettier-format plugin to be installed. https://github.com/dylanarmstrong/obsidian-prettier-plugin", - ) - .addToggle((toggle) => { - toggle - .setValue(this.plugin.settings.runPrettierAfterFormatting) - .onChange(async (value) => { - this.plugin.settings.runPrettierAfterFormatting = value - await this.plugin.saveData(this.plugin.settings) - }) - }) - - // Setting for linter delay - new Setting(containerEl) - .setName("Format delay (ms)") - .setDesc( - "Delay in milliseconds before formatting. Increase this value if the linter/prettier runs before the file is fully saved.", - ) - .addText((text) => { - text.setPlaceholder("e.g. 100") - .setValue(this.plugin.settings.formatDelayMs.toString()) - .onChange(async (value) => { - const parsedValue = parseInt(value) - if (!isNaN(parsedValue) && parsedValue >= 0) { - this.plugin.settings.formatDelayMs = parsedValue - await this.plugin.saveData(this.plugin.settings) - } - }) - }) - - new Setting(containerEl) - .setName("URL Replacement with Title") - .setHeading() - - // Toggle for replacing URLs with titles - new Setting(containerEl) - .setName("Replace URL with title") - .setDesc( - "When enabled, raw URLs will be replaced with [Page Title](URL). Requires fetching the URL content.", - ) - .addToggle((toggle) => { - toggle - .setValue(this.plugin.settings.replaceUrlWithTitle) - .onChange(async (value) => { - this.plugin.settings.replaceUrlWithTitle = value - await this.plugin.saveData(this.plugin.settings) - }) - }) - - new Setting(containerEl) - .setName("Ignore domains") - .setDesc( - "Ignore domains for replacing URLs with titles, one per line (e.g. x.com)", - ) - .addTextArea((text) => { - text.setPlaceholder("") - .setValue( - this.plugin.settings.replaceUrlWithTitleIgnoreDomains.join( - "\n", - ), - ) - .onChange(async (value) => { - // Split by newlines and filter out empty lines - const urls = value - .split("\n") - .map(url => url.trim()) - .filter(Boolean) - this.plugin.settings.replaceUrlWithTitleIgnoreDomains - = urls - await this.plugin.saveData(this.plugin.settings) - await this.plugin.buildUrlTitleMap() - }) - // Make the text area taller - text.inputEl.rows = 4 - text.inputEl.cols = 50 - }) - - new Setting(containerEl) - .setName("URL Formatting for GitHub") - .setHeading() - - // Toggle for formatting GitHub URLs on save - new Setting(containerEl) - .setName("Format GitHub URLs on save") - .setDesc( - "When enabled, GitHub URLs will be formatted when saving the file.", - ) - .addToggle((toggle) => { - toggle - .setValue(this.plugin.settings.formatGitHubURLs) - .onChange(async (value) => { - this.plugin.settings.formatGitHubURLs = value - await this.plugin.saveData(this.plugin.settings) - }) - }) - - // Add GitHub Enterprise URLs setting - new Setting(containerEl) - .setName("GitHub Enterprise URLs") - .setDesc( - "Add your GitHub Enterprise URLs, one per line (e.g. github.enterprise.com)", - ) - .addTextArea((text) => { - text.setPlaceholder("github.enterprise.com\ngithub.company.com") - .setValue( - this.plugin.settings.githubEnterpriseURLs.join("\n"), - ) - .onChange(async (value) => { - // Split by newlines and filter out empty lines - const urls = value - .split("\n") - .map(url => url.trim()) - .filter(Boolean) - this.plugin.settings.githubEnterpriseURLs = urls - await this.plugin.saveData(this.plugin.settings) - }) - // Make the text area taller - text.inputEl.rows = 4 - text.inputEl.cols = 50 - }) - - new Setting(containerEl) - .setName("URL Formatting for Jira") - .setHeading() - - // Toggle for formatting JIRA URLs on save - new Setting(containerEl) - .setName("Format JIRA URLs on save") - .setDesc( - "When enabled, JIRA URLs will be formatted when saving the file.", - ) - .addToggle((toggle) => { - toggle - .setValue(this.plugin.settings.formatJiraURLs) - .onChange(async (value) => { - this.plugin.settings.formatJiraURLs = value - await this.plugin.saveData(this.plugin.settings) - }) - }) - - // Add JIRA URLs setting - new Setting(containerEl) - .setName("JIRA URLs") - .setDesc( - "Add your JIRA URLs, one per line (e.g. jira.enterprise.com)", - ) - .addTextArea((text) => { - text.setPlaceholder("jira.enterprise.com\njira.company.com") - .setValue(this.plugin.settings.jiraURLs.join("\n")) - .onChange(async (value) => { - // Split by newlines and filter out empty lines - const urls = value - .split("\n") - .map(url => url.trim()) - .filter(Boolean) - this.plugin.settings.jiraURLs = urls - await this.plugin.saveData(this.plugin.settings) - }) - // Make the text area taller - text.inputEl.rows = 4 - text.inputEl.cols = 50 - }) - - new Setting(containerEl) - .setName("URL Formatting for Linear") - .setHeading() - - // Toggle for formatting Linear URLs on save - new Setting(containerEl) - .setName("Format Linear URLs on save") - .setDesc( - "When enabled, Linear URLs will be formatted when saving the file.", - ) - .addToggle((toggle) => { - toggle - .setValue(this.plugin.settings.formatLinearURLs) - .onChange(async (value) => { - this.plugin.settings.formatLinearURLs = value - await this.plugin.saveData(this.plugin.settings) - }) - }) - - new Setting(containerEl).setName("Debug").setHeading() - - // Toggle for showing the load notice. - new Setting(containerEl) - .setName("Show load notice") - .setDesc("Display a notice when markdown files are loaded.") - .addToggle((toggle) => { - toggle - .setValue(this.plugin.settings.showNotice) - .onChange(async (value) => { - this.plugin.settings.showNotice = value - await this.plugin.saveData(this.plugin.settings) - }) - }) - - // Toggle for debug logging - new Setting(containerEl) - .setName("Debug mode") - .setDesc( - "When enabled, debug information will be logged to the console.", - ) - .addToggle((toggle) => { - toggle - .setValue(this.plugin.settings.debug) - .onChange(async (value) => { - this.plugin.settings.debug = value - await this.plugin.saveData(this.plugin.settings) - }) - }) - - new Setting(containerEl) - .setName("AI Link Enhancement (Beta)") - .setHeading() - - new Setting(containerEl) - .setName("Enable AI Link Enhancement") - .setDesc( - "When enabled, an AI-powered link enhancer command will be available. It uses a local LLM to resolve ambiguous links and correct existing ones.", - ) - .addToggle((toggle) => { - toggle - .setValue(this.plugin.settings.aiEnabled) - .onChange(async (value) => { - this.plugin.settings.aiEnabled = value - await this.plugin.saveData(this.plugin.settings) - }) - }) - - new Setting(containerEl) - .setName("AI API Endpoint") - .setDesc("The URL of your OpenAI-compatible AI server (e.g. LM Studio, Ollama).") - .addText((text) => { - text.setPlaceholder("http://localhost:1234/v1") - .setValue(this.plugin.settings.aiEndpoint) - .onChange(async (value) => { - this.plugin.settings.aiEndpoint = value - await this.plugin.saveData(this.plugin.settings) - }) - }) - - new Setting(containerEl) - .setName("AI Model") - .setDesc("The name of the model to use (e.g. gemma-4-7b).") - .addText((text) => { - text.setPlaceholder("gemma-4-7b") - .setValue(this.plugin.settings.aiModel) - .onChange(async (value) => { - this.plugin.settings.aiModel = value - await this.plugin.saveData(this.plugin.settings) - }) - }) - - new Setting(containerEl) - .setName("Max Context Length") - .setDesc("Number of characters around the link to provide as context to the AI.") - .addText((text) => { - text.setPlaceholder("500") - .setValue(this.plugin.settings.aiMaxContext.toString()) - .onChange(async (value) => { - const parsedValue = parseInt(value) - if (!isNaN(parsedValue) && parsedValue > 0) { - this.plugin.settings.aiMaxContext = parsedValue - await this.plugin.saveData(this.plugin.settings) - } - }) - }) + const renderedGroups = new Set() + for (const entry of SETTINGS_CATALOG) { + if (!renderedGroups.has(entry.group)) { + new Setting(containerEl).setName(entry.group).setHeading() + renderedGroups.add(entry.group) + } + this.renderSetting(containerEl, entry) + } } } diff --git a/src/utils/__tests__/resolve-ambiguities.test.ts b/src/utils/__tests__/resolve-ambiguities.test.ts index 36e7196..53ac13a 100644 --- a/src/utils/__tests__/resolve-ambiguities.test.ts +++ b/src/utils/__tests__/resolve-ambiguities.test.ts @@ -73,4 +73,219 @@ describe("resolveAmbiguities", () => { ) expect(result.get("[[private/meeting|meeting]]")).toBe("work/meeting") }) + + it("should not request AI for existing links inside inline code", async () => { + const text = "Check the `[[private/meeting|meeting]]` notes." + vi.mocked(aiClient.resolveAmbiguitiesBatch).mockClear() + vi.mocked(aiClient.resolveAmbiguitiesBatch).mockResolvedValue(new Map()) + + const result = await resolveAmbiguities(text, candidateMap, trie, mockSettings) + + expect(aiClient.resolveAmbiguitiesBatch).toHaveBeenCalledWith( + mockSettings, + [], + ) + expect(result.size).toBe(0) + }) + + it("should skip fenced code blocks, callouts, and ignored headings", async () => { + const text = `# meeting +> [!note] +> meeting + +~~~ts +meeting +~~~ + +meeting` + vi.mocked(aiClient.resolveAmbiguitiesBatch).mockClear() + vi.mocked(aiClient.resolveAmbiguitiesBatch).mockResolvedValue( + new Map([["meeting", "work/meeting"]]), + ) + + await resolveAmbiguities(text, candidateMap, trie, { + ...mockSettings, + ignoreHeadings: true, + proximityBasedLinking: true, + }) + + expect(aiClient.resolveAmbiguitiesBatch).toHaveBeenCalledTimes(1) + expect(aiClient.resolveAmbiguitiesBatch).toHaveBeenCalledWith( + expect.objectContaining({ + ignoreHeadings: true, + proximityBasedLinking: true, + }), + [ + expect.objectContaining({ + word: "meeting", + candidates: ["work/meeting", "private/meeting"], + }), + ], + ) + }) + + it("should not request AI for Korean particle forms that replacement skips", async () => { + const koreanCandidateMap = new Map([ + [ + "λ¬Έμ„œ", + { + candidates: [ + { canonical: "work/λ¬Έμ„œ", scoped: false, namespace: "work" }, + { canonical: "private/λ¬Έμ„œ", scoped: false, namespace: "private" }, + ], + }, + ], + ]) + const koreanTrie: TrieNode = buildTrie(["λ¬Έμ„œ"], true) + vi.mocked(aiClient.resolveAmbiguitiesBatch).mockClear() + vi.mocked(aiClient.resolveAmbiguitiesBatch).mockResolvedValue(new Map()) + + const result = await resolveAmbiguities( + "λ¬Έμ„œλŠ” λ¬Έμ„œμ€", + koreanCandidateMap, + koreanTrie, + mockSettings, + ) + + expect(aiClient.resolveAmbiguitiesBatch).toHaveBeenCalledWith( + mockSettings, + [], + ) + expect(result.size).toBe(0) + }) + + it("should not request AI for self-link candidates when the file path matches", async () => { + const selfLinkCandidateMap = new Map([ + [ + "meeting", + { + candidates: [ + { canonical: "work/meeting", scoped: false, namespace: "work" }, + { canonical: "private/meeting", scoped: false, namespace: "private" }, + ], + }, + ], + ]) + const selfLinkTrie: TrieNode = buildTrie(["meeting"], true) + vi.mocked(aiClient.resolveAmbiguitiesBatch).mockClear() + vi.mocked(aiClient.resolveAmbiguitiesBatch).mockResolvedValue(new Map()) + + const result = await resolveAmbiguities( + "meeting", + selfLinkCandidateMap, + selfLinkTrie, + { + ...mockSettings, + preventSelfLinking: true, + }, + "work/meeting", + ) + + expect(aiClient.resolveAmbiguitiesBatch).toHaveBeenCalledWith( + expect.objectContaining({ + preventSelfLinking: true, + }), + [], + ) + expect(result.size).toBe(0) + }) + + it("uses baseDir when filtering scoped candidates for AI requests", async () => { + const scopedCandidateMap = new Map([ + [ + "internal", + { + candidates: [ + { canonical: "pages/team-a/internal", scoped: true, namespace: "team-a" }, + { canonical: "pages/team-a/archive/internal", scoped: true, namespace: "team-a" }, + ], + }, + ], + ]) + const scopedTrie: TrieNode = buildTrie(["internal"], true) + vi.mocked(aiClient.resolveAmbiguitiesBatch).mockClear() + vi.mocked(aiClient.resolveAmbiguitiesBatch).mockResolvedValue(new Map()) + + await resolveAmbiguities( + "internal", + scopedCandidateMap, + scopedTrie, + { + ...mockSettings, + proximityBasedLinking: true, + }, + "pages/team-a/today", + "pages", + ) + + const [settingsArg, requestsArg] = vi.mocked(aiClient.resolveAmbiguitiesBatch).mock.calls[0] + expect(settingsArg).toEqual(expect.objectContaining({ + proximityBasedLinking: true, + })) + expect(settingsArg).not.toHaveProperty("baseDir") + expect(requestsArg).toEqual([ + expect.objectContaining({ + word: "internal", + candidates: [ + "pages/team-a/internal", + "pages/team-a/archive/internal", + ], + }), + ]) + }) + + it("does not request AI for candidates inside raw Linear URLs", async () => { + const linearCandidateMap = new Map([ + [ + "linear", + { + candidates: [ + { canonical: "work/linear", scoped: false, namespace: "work" }, + { canonical: "private/linear", scoped: false, namespace: "private" }, + ], + }, + ], + ]) + const linearTrie: TrieNode = buildTrie(["linear"], true) + vi.mocked(aiClient.resolveAmbiguitiesBatch).mockClear() + vi.mocked(aiClient.resolveAmbiguitiesBatch).mockResolvedValue(new Map()) + + const result = await resolveAmbiguities( + "Open linear://issue/TEAM-123", + linearCandidateMap, + linearTrie, + mockSettings, + ) + + expect(aiClient.resolveAmbiguitiesBatch).toHaveBeenCalledWith( + mockSettings, + [], + ) + expect(result.size).toBe(0) + }) + + it("does not request AI for candidates inside ignored Markdown table rows", async () => { + vi.mocked(aiClient.resolveAmbiguitiesBatch).mockClear() + vi.mocked(aiClient.resolveAmbiguitiesBatch).mockResolvedValue(new Map()) + + const result = await resolveAmbiguities( + `| Topic | +| --- | +| meeting |`, + candidateMap, + trie, + { + ...mockSettings, + ignoreMarkdownTables: true, + }, + ) + + expect(aiClient.resolveAmbiguitiesBatch).toHaveBeenCalledWith( + expect.objectContaining({ + ignoreMarkdownTables: true, + }), + [], + ) + expect(result.size).toBe(0) + }) }) diff --git a/src/utils/resolve-ambiguities.ts b/src/utils/resolve-ambiguities.ts index 831a947..10a17fc 100644 --- a/src/utils/resolve-ambiguities.ts +++ b/src/utils/resolve-ambiguities.ts @@ -1,79 +1,38 @@ import { CandidateData, TrieNode } from "../trie" import { AutomaticLinkerSettings } from "../settings/settings-info" import { resolveAmbiguitiesBatch, AIResolveRequest } from "./ai-client" +import { + getOccurrenceContext, + scanCandidateOccurrences, +} from "../replace-links/candidate-scanner" -// A simplified scanner to find candidates that have multiple options or existing links to verify. export const resolveAmbiguities = async ( text: string, candidateMap: Map, trie: TrieNode, settings: AutomaticLinkerSettings, + filePath = "", + baseDir?: string, ): Promise> => { - const requests: AIResolveRequest[] = [] + const scannerSettings = baseDir === undefined + ? settings + : { ...settings, baseDir } + const occurrences = scanCandidateOccurrences({ + text, + filePath, + trie, + candidateMap, + settings: scannerSettings, + }) - // 1. Scan for existing links [[Path|Alias]] or [[Path]] - const existingLinkRegex = /\[\[([^|\]]+)(?:\|([^\]]+))?\]\]/g - let match: RegExpExecArray | null - while ((match = existingLinkRegex.exec(text)) !== null) { - const fullMatch = match[0] - const path = match[1] - const alias = match[2] || path + const requests: AIResolveRequest[] = occurrences + .filter(occurrence => occurrence.candidateData.candidates.length > 1) + .map(occurrence => ({ + word: occurrence.text, + text: getOccurrenceContext(text, occurrence, settings.aiMaxContext), + candidates: occurrence.candidateData.candidates.map(c => c.canonical), + })) - const candidateData = candidateMap.get(alias) - if (candidateData && candidateData.candidates.length > 1) { - const start = Math.max(0, match.index - settings.aiMaxContext) - const end = Math.min(text.length, match.index + fullMatch.length + settings.aiMaxContext) - - requests.push({ - word: fullMatch, // Using the full link as the "word" key for replacement - text: text.slice(start, end), - candidates: candidateData.candidates.map(c => c.canonical), - }) - } - } - - // 2. Scan for unlinked words with multiple candidates using the Trie - // (This is a simplified version of the trie traversal logic) - let i = 0 - while (i < text.length) { - let node = trie - let j = i - let lastMatch: { word: string, data: CandidateData, index: number } | null = null - - while (j < text.length && node.children.has(text[j].toLowerCase())) { - node = node.children.get(text[j].toLowerCase())! - if (node.candidate) { - const data = candidateMap.get(node.candidate) - if (data) { - lastMatch = { word: node.candidate, data, index: i } - } - } - j++ - } - - if (lastMatch && lastMatch.data.candidates.length > 1) { - // Check if it's already inside a link (simple check) - const isInsideLink = text.slice(Math.max(0, i - 2), i) === "[[" - || text.slice(i + lastMatch.word.length, i + lastMatch.word.length + 2) === "]]" - - if (!isInsideLink) { - const start = Math.max(0, i - settings.aiMaxContext) - const end = Math.min(text.length, i + lastMatch.word.length + settings.aiMaxContext) - - requests.push({ - word: lastMatch.word, - text: text.slice(start, end), - candidates: lastMatch.data.candidates.map(c => c.canonical), - }) - } - i += lastMatch.word.length - } - else { - i++ - } - } - - // Deduplicate requests by word to avoid redundant AI calls const uniqueRequests = Array.from(new Map(requests.map(r => [r.word, r])).values()) return await resolveAmbiguitiesBatch(settings, uniqueRequests)