Momento/memento-note/tests/migration/chunk-indexing.test.ts

import { test, expect, describe, beforeAll, afterAll, beforeEach } from 'vitest'
import { prisma } from '../../lib/prisma'
import { ChunkIndexingService } from '../../lib/ai/services/chunk-indexing.service'
import { embeddingService } from '../../lib/ai/services/embedding.service'
import crypto from 'crypto'

const testNoteId = 'test-chunk-000001'

describe('US-CHUNK-2 : Indexation incrémentale avec dedup', () => {
  let originalEmbedText: any

  beforeAll(async () => {
    originalEmbedText = embeddingService.embedText
    embeddingService.embedText = async (text: string) => {
      const hash = crypto.createHash('md5').update(text).digest()
      return Array.from({ length: 1536 }, (_, i) => hash[i % 16] / 255)
    }

    await prisma.note.upsert({
      where: { id: testNoteId },
      create: {
        id: testNoteId,
        title: 'Test Note for Chunk Indexing',
        content: '<p>Test</p>',
      },
      update: {},
    })
  })

  afterAll(async () => {
    embeddingService.embedText = originalEmbedText
    await prisma.noteEmbeddingChunk.deleteMany({ where: { noteId: testNoteId } })
    await prisma.note.delete({ where: { id: testNoteId } }).catch(() => {})
  })

  beforeEach(async () => {
    await prisma.noteEmbeddingChunk.deleteMany({ where: { noteId: testNoteId } })
  })

  const longContent = Array.from({ length: 8 }, (_, i) =>
    `Section ${i} de la note de test. `.repeat(60).trim(),
  ).join('\n\n')

  test('première indexation → tous les fragments sont nouveaux', async () => {
    const service = new ChunkIndexingService()
    const result = await service.indexNote(testNoteId, 'Note de test', longContent)

    expect(result.newFragments).toBeGreaterThan(0)
    expect(result.deleted).toBe(0)
    expect(result.skipped).toBe(0)
  })

  test('deuxième indexation (même contenu) → tout skipped, 0 nouveau', async () => {
    const service = new ChunkIndexingService()
    await service.indexNote(testNoteId, 'Note de test', longContent)
    const result = await service.indexNote(testNoteId, 'Note de test', longContent)

    expect(result.newFragments).toBe(0)
    expect(result.skipped).toBeGreaterThan(0)
    expect(result.deleted).toBe(0)
  })

  test('modification d\'une section → 1 nouveau, reste skip', async () => {
    const service = new ChunkIndexingService()
    await service.indexNote(testNoteId, 'Note de test', longContent)

    const sections = Array.from({ length: 8 }, (_, i) =>
      `Section ${i === 3 ? 'MODIFIÉE' : i} de la note de test. `.repeat(60).trim(),
    )
    const modified = sections.join('\n\n')

    const result = await service.indexNote(testNoteId, 'Note de test', modified)

    expect(result.newFragments).toBeGreaterThanOrEqual(1)
    expect(result.deleted).toBeGreaterThanOrEqual(1)
  })

  test('suppression d\'une section → fragments stale nettoyés', async () => {
    const service = new ChunkIndexingService()
    const sections = Array.from({ length: 8 }, (_, i) =>
      `Section ${i} de la note de test. `.repeat(60).trim(),
    )
    await service.indexNote(testNoteId, 'Note de test', sections.join('\n\n'))

    const beforeCount = await prisma.noteEmbeddingChunk.count({
      where: { noteId: testNoteId },
    })

    const shorter = sections.slice(0, 4).join('\n\n')
    const result = await service.indexNote(testNoteId, 'Note de test', shorter)

    const afterCount = await prisma.noteEmbeddingChunk.count({
      where: { noteId: testNoteId },
    })

    expect(afterCount).toBeLessThan(beforeCount)
    expect(result.deleted).toBeGreaterThan(0)
  })

  test('note vide → tous les fragments supprimés', async () => {
    const service = new ChunkIndexingService()
    await service.indexNote(testNoteId, 'Note de test', longContent)
    const result = await service.indexNote(testNoteId, '', '')

    const count = await prisma.noteEmbeddingChunk.count({
      where: { noteId: testNoteId },
    })

    expect(count).toBe(0)
  })

  test('deleteNoteChunks → supprime tout', async () => {
    const service = new ChunkIndexingService()
    await service.indexNote(testNoteId, 'Note de test', longContent)
    await service.deleteNoteChunks(testNoteId)

    const count = await prisma.noteEmbeddingChunk.count({
      where: { noteId: testNoteId },
    })
    expect(count).toBe(0)
  })

  test('hasChunks → détection correcte', async () => {
    const service = new ChunkIndexingService()
    const before = await service.hasChunks(testNoteId)
    expect(before).toBe(false)

    await service.indexNote(testNoteId, 'Note de test', longContent)
    const after = await service.hasChunks(testNoteId)
    expect(after).toBe(true)
  })
})