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, expect, it } from "vitest"
import { buildWordDiff, findUniqueTextSelection, normalizeEditableMarkdown, normalizeSelectionReplacement } from "./selection-edit"
describe("normalizeSelectionReplacement", () => {
it("removes one complete outer Markdown fence", () => {
expect(normalizeSelectionReplacement("```markdown\nreplacement\n```"))
.toBe("replacement")
})
it("preserves whitespace and incomplete or embedded fences", () => {
expect(normalizeSelectionReplacement(" replacement ")).toBe(" replacement ")
expect(normalizeSelectionReplacement("before\n```text\ninside\n```\nafter"))
.toBe("before\n```text\ninside\n```\nafter")
})
})
describe("normalizeEditableMarkdown", () => {
it("normalizes CRLF and lone CR without changing existing LF", () => {
expect(normalizeEditableMarkdown("first\r\nsecond\rthird\nfourth"))
.toBe("first\nsecond\nthird\nfourth")
})
})
describe("buildWordDiff", () => {
it("marks inserted and deleted words while preserving equal text", () => {
const parts = buildWordDiff("The old text", "The new text")
expect(parts.some((part) => part.type === "delete" && part.value.includes("old"))).toBe(true)
expect(parts.some((part) => part.type === "insert" && part.value.includes("new"))).toBe(true)
expect(parts.filter((part) => part.type === "equal").map((part) => part.value).join(""))
.toContain("The ")
})
it("keeps unchanged CJK characters instead of replacing the whole sentence", () => {
const parts = buildWordDiff("知识图谱检索", "知识图谱搜索")
expect(parts.some((part) => part.type === "equal" && part.value.includes("知识图谱"))).toBe(true)
expect(parts.some((part) => part.type === "delete" && part.value === "检")).toBe(true)
expect(parts.some((part) => part.type === "insert" && part.value === "搜")).toBe(true)
})
})
describe("findUniqueTextSelection", () => {
it("maps a unique rendered selection to its exact source range", () => {
expect(findUniqueTextSelection("before selected after", "selected")).toEqual({
prefix: "before ",
selectedText: "selected",
suffix: " after",
})
})
it("accepts browser-collapsed whitespace when the match is unique", () => {
expect(findUniqueTextSelection("before\nselected words\nafter", "selected words")?.selectedText)
.toBe("selected words")
})
it("rejects empty and ambiguous selections", () => {
expect(findUniqueTextSelection("same and same", "same")).toBeNull()
expect(findUniqueTextSelection("content", " ")).toBeNull()
})
it("returns null when the rendered text does not exist in the source", () => {
expect(findUniqueTextSelection("existing content", "missing content")).toBeNull()
})
})