LLM Wiki is a cross-platform desktop application that turns your documents into an organized, interlinked knowledge base — automatically. Instead of traditional RAG (retrieve-and-answer from scratch every time), the LLM incrementally builds and maintains a persistent wiki from your sources。
TypeScript
import { describe, it, expect } from "vitest"
import { transformImageEmbeds, transformWikilinks } from "./wikilink-transform"
describe("transformWikilinks", () => {
it("returns input unchanged when there are no wikilinks", () => {
expect(transformWikilinks("just plain text")).toBe("just plain text")
expect(transformWikilinks("# Heading\n\nparagraph")).toBe("# Heading\n\nparagraph")
})
it("converts a bare [[slug]] to a standard markdown link", () => {
expect(transformWikilinks("see [[foo]] for details")).toBe(
"see [foo](#foo) for details",
)
})
it("uses the alias as label and target as href for [[slug|alias]]", () => {
expect(transformWikilinks("see [[foo|the foo page]]")).toBe(
"see [the foo page](#foo)",
)
})
it("trims whitespace inside the wikilink", () => {
expect(transformWikilinks("[[ foo | the alias ]]")).toBe(
"[the alias](#foo)",
)
})
it("converts multiple wikilinks on the same line", () => {
expect(transformWikilinks("[[a]] and [[b|B]] and [[c]]")).toBe(
"[a](#a) and [B](#b) and [c](#c)",
)
})
it("encodes special characters in the href", () => {
expect(transformWikilinks("[[hello world]]")).toBe(
"[hello world](#hello%20world)",
)
})
it("does not touch wikilinks inside fenced code blocks", () => {
const input = "before [[a]]\n```md\ncode [[b]] block\n```\nafter [[c]]"
expect(transformWikilinks(input)).toBe(
"before [a](#a)\n```md\ncode [[b]] block\n```\nafter [c](#c)",
)
})
it("does not touch wikilinks inside inline code spans", () => {
expect(transformWikilinks("text `[[skip]]` and [[keep]]")).toBe(
"text `[[skip]]` and [keep](#keep)",
)
})
it("handles multiple inline-code spans correctly", () => {
expect(transformWikilinks("[[a]] `[[b]]` [[c]] `[[d]]` [[e]]")).toBe(
"[a](#a) `[[b]]` [c](#c) `[[d]]` [e](#e)",
)
})
it("preserves [[empty alias|]] by falling back to target as label", () => {
expect(transformWikilinks("[[foo|]]")).toBe("[foo](#foo)")
})
it("matches the real DPAO body wikilink density", () => {
const input =
"DPAOs differ from [[paos|Polyphosphate-Accumulating Organisms (PAOs)]] " +
"and store [[pha|polyhydroxyalkanoates (PHAs)]] from [[vfa|volatile fatty acids (VFAs)]]. " +
"[[accumulibacter]] is the most-characterized genus."
const out = transformWikilinks(input)
expect(out).toContain("[Polyphosphate-Accumulating Organisms (PAOs)](#paos)")
expect(out).toContain("[polyhydroxyalkanoates (PHAs)](#pha)")
expect(out).toContain("[volatile fatty acids (VFAs)](#vfa)")
expect(out).toContain("[accumulibacter](#accumulibacter)")
expect(out).not.toContain("[[")
})
it("does not mangle existing standard markdown links", () => {
const input = "see [foo](https://example.com) and [[bar]]"
expect(transformWikilinks(input)).toBe(
"see [foo](https://example.com) and [bar](#bar)",
)
})
it("leaves dangling brackets untouched", () => {
expect(transformWikilinks("[[broken")).toBe("[[broken")
expect(transformWikilinks("broken]]")).toBe("broken]]")
})
})
describe("transformImageEmbeds", () => {
it("returns input unchanged when there are no embeds", () => {
expect(transformImageEmbeds("plain text")).toBe("plain text")
expect(transformImageEmbeds("a [[link]] but no embed")).toBe(
"a [[link]] but no embed",
)
})
it("converts an Obsidian image embed to standard markdown", () => {
expect(transformImageEmbeds("![[../assets/abc123.png]]")).toBe(
"",
)
})
it("converts a whiteboard-screenshot embed (jpg under boards/)", () => {
expect(
transformImageEmbeds("![[../assets/boards/WB1.jpg]]"),
).toBe("")
})
it("uses the alias as alt text for ![[target|alias]]", () => {
expect(transformImageEmbeds("![[img.png|A diagram]]")).toBe(
"",
)
})
it("wraps the target in <…> so spaces/non-ASCII survive the parser", () => {
expect(transformImageEmbeds("![[../assets/变与不变 图.png]]")).toBe(
"",
)
})
it("converts multiple embeds and keeps surrounding text", () => {
expect(
transformImageEmbeds("before ![[a.png]] middle ![[b.png|B]] after"),
).toBe("before  middle  after")
})
it("does not touch embeds inside fenced code blocks", () => {
const input = "![[a.png]]\n```\n![[b.png]]\n```\n![[c.png]]"
expect(transformImageEmbeds(input)).toBe(
"\n```\n![[b.png]]\n```\n",
)
})
it("does not touch embeds inside inline code spans", () => {
expect(transformImageEmbeds("`![[skip.png]]` and ![[keep.png]]")).toBe(
"`![[skip.png]]` and ",
)
})
it("leaves an embed whose alias contains `]` untouched (safe degradation)", () => {
// A `]` inside an alias is not valid Obsidian embed syntax; the
// regex declines to match rather than producing a malformed
// image. The literal text is preserved.
expect(transformImageEmbeds("![[x.png|a ] b]]")).toBe("![[x.png|a ] b]]")
})
})
describe("composition: embeds then wikilinks", () => {
it("an image embed survives the wikilink pass unmangled", () => {
// This is the real pipeline order used by the renderers.
const input = "See ![[../assets/x.png]] and the [[concept]] page."
const out = transformWikilinks(transformImageEmbeds(input))
expect(out).toBe("See  and the [concept](#concept) page.")
})
it("REGRESSION: wikilinks-only would have mangled an image embed", () => {
// Documents the bug this fix addresses: running the generic
// wikilink transform alone turns `![[x.png]]` into a broken
// image whose src is a `#fragment`. The embed pass must run first.
const mangled = transformWikilinks("![[x.png]]")
expect(mangled).toBe("")
// …whereas the correct pipeline yields a real image reference:
const correct = transformWikilinks(transformImageEmbeds("![[x.png]]"))
expect(correct).toBe("")
})
})