Beautiful, AI-native markdown IDE and LLM wiki
TypeScript 2nd-brainagent-skillsclaudecodexcompany-braindocsknowledge-knowledge-baseknowledge-managementllmllm-wikillm-wiki-karpathymarkdownmarkdown-editormdmdxnotespkmsecond-brainskills
import { mkdtempSync, rmSync } from 'node:fs';
import { tmpdir } from 'node:os';
import { join } from 'node:path';
import { afterEach, describe, expect, test } from 'vitest';
import { getLogger } from '../logger.ts';
import { createAnchor } from './anchor.ts';
import { CommentIndex } from './comment-index.ts';
import { CommentThreadStore } from './thread-store.ts';
import type { CommentThreadMeta } from './types.ts';
const log = getLogger('comments-index-test');
let dirs: string[] = [];
function tmp(): string {
const d = mkdtempSync(join(tmpdir(), 'comments-index-test-'));
dirs.push(d);
return d;
}
afterEach(() => {
for (const d of dirs) rmSync(d, { recursive: true, force: true });
dirs = [];
});
const BODY = 'one two three four five';
const anchor = (q: string) => {
const s = BODY.indexOf(q);
return createAnchor(BODY, s, s + q.length);
};
function thread(threadId: string, docName: string, quote = 'one'): CommentThreadMeta {
return {
threadId,
docName,
anchor: anchor(quote),
state: 'anchored',
queued: false,
latestComment: `note on ${threadId}`,
createdBy: 'principal-x',
createdAt: 1,
};
}
describe('CommentIndex', () => {
test('builds from disk at boot', async () => {
const store = new CommentThreadStore(tmp(), log);
await store.init();
for (const [id, doc, quote] of [
['t1', 'a', 'one'],
['t2', 'a', 'two'],
['t3', 'b', 'three'],
] as const) {
await store.createThread({
threadId: id,
docName: doc,
anchor: anchor(quote),
createdBy: 'principal-x',
createdAt: 1,
body: 'a note',
});
await store.whenIdle(id);
}
const index = new CommentIndex();
await index.build(store);
expect(index.size).toBe(3);
expect(index.threadsForDoc('a').sort()).toEqual(['t1', 't2']);
expect(index.threadsForDoc('b')).toEqual(['t3']);
expect(index.threadsForDoc('missing')).toEqual([]);
expect(index.listForDoc('b')[0].latestComment).toBe('a note');
});
test('listAll returns every cached thread', () => {
const index = new CommentIndex();
index.upsert(thread('t1', 'a'));
index.upsert(thread('t2', 'b'));
expect(
index
.listAll()
.map((m) => m.threadId)
.sort(),
).toEqual(['t1', 't2']);
});
test('upsert replaces the cached thread, not just its doc', () => {
const index = new CommentIndex();
index.upsert(thread('t1', 'a'));
index.upsert({ ...thread('t1', 'a'), state: 'resolved', latestComment: 'revised' });
const [meta] = index.listForDoc('a');
expect(meta.state).toBe('resolved');
expect(meta.latestComment).toBe('revised');
});
test('upsert moves a thread to a new doc', () => {
const index = new CommentIndex();
index.upsert(thread('t1', 'a'));
index.upsert(thread('t1', 'b'));
expect(index.listForDoc('a')).toEqual([]);
expect(index.threadsForDoc('b')).toEqual(['t1']);
});
test('remove drops a thread and prunes an empty doc key', () => {
const index = new CommentIndex();
index.upsert(thread('t1', 'a'));
index.remove('t1');
expect(index.threadsForDoc('a')).toEqual([]);
expect(index.listAll()).toEqual([]);
expect(index.size).toBe(0);
});
test('renameDoc re-points every thread and rewrites its cached docName', () => {
const index = new CommentIndex();
index.upsert(thread('t1', 'old'));
index.upsert(thread('t2', 'old'));
index.renameDoc('old', 'new');
expect(index.threadsForDoc('old')).toEqual([]);
expect(index.threadsForDoc('new').sort()).toEqual(['t1', 't2']);
expect(index.listAll().every((m) => m.docName === 'new')).toBe(true);
});
test('renameDoc merges into an existing destination doc', () => {
const index = new CommentIndex();
index.upsert(thread('t1', 'old'));
index.upsert(thread('t2', 'new'));
index.renameDoc('old', 'new');
expect(index.threadsForDoc('new').sort()).toEqual(['t1', 't2']);
});
});
describe('CommentIndex counting', () => {
test('counts every requested doc, zero included', () => {
const index = new CommentIndex();
index.upsert(thread('t1', 'a'));
index.upsert(thread('t2', 'a'));
expect([...index.countsForDocs(['a', 'b'])]).toEqual([
['a', 2],
['b', 0],
]);
});
test('resolved threads do not count', () => {
const index = new CommentIndex();
index.upsert(thread('t1', 'a'));
index.upsert({ ...thread('t2', 'a'), state: 'resolved' });
expect(index.countForDoc('a')).toBe(1);
expect(index.countsForDocs(['a']).get('a')).toBe(1);
});
test('orphaned threads still count', () => {
const index = new CommentIndex();
index.upsert({ ...thread('t1', 'a'), state: 'orphaned' });
expect(index.countForDoc('a')).toBe(1);
});
test('prefix rollup is sparse and segment-bounded', () => {
const index = new CommentIndex();
index.upsert(thread('t1', 'docs/a'));
index.upsert(thread('t2', 'docs/nested/b'));
index.upsert(thread('t3', 'docsite/c'));
index.upsert({ ...thread('t4', 'docs/resolved-only'), state: 'resolved' });
const counts = index.countsUnderPrefix('docs');
expect([...counts].sort()).toEqual([
['docs/a', 1],
['docs/nested/b', 1],
]);
expect(counts.has('docs/resolved-only')).toBe(false);
expect(counts.has('docsite/c')).toBe(false);
});
test('a trailing slash on the prefix means the same folder', () => {
const index = new CommentIndex();
index.upsert(thread('t1', 'docs/a'));
expect([...index.countsUnderPrefix('docs/')]).toEqual([['docs/a', 1]]);
});
test('an empty prefix rolls up the whole project', () => {
const index = new CommentIndex();
index.upsert(thread('t1', 'docs/a'));
index.upsert(thread('t2', 'top'));
expect([...index.countsUnderPrefix('')].sort()).toEqual([
['docs/a', 1],
['top', 1],
]);
});
});