ApiCore/ApiCore-Backend/test/voice-extraction.e2e-spec.ts

import * as bcrypt from 'bcryptjs';
import { Prisma } from '../generated/prisma';
import { PrismaService } from '../src/prisma/prisma.service';
import { sealTranscript } from '../src/voice/core/transcript';
import { openVoiceText, voiceAad } from '../src/voice/core/voice-crypto';
import { FakeStructuredModel } from '../src/voice/extraction/fake-structured-model';
import { VoiceExtractionService } from '../src/voice/extraction/voice-extraction.service';

const SECRET_TEXT = 'Geheimer Befund Königin gesehen vier Brutwaben';
const ANSWER = {
  findings: [
    {
      field: 'queenSeen',
      value: true,
      quote: 'Königin gesehen',
      corrected: false,
    },
  ],
  unclear: [],
};

/**
 * The extraction loop against the real database with a FakeStructuredModel
 * (no network, no key). Runs are scoped to this file's own users; every row
 * is created and removed here because other suites share the database.
 */
describe('voice extraction loop (e2e)', () => {
  const suffix = `${Date.now()}-${Math.round(Math.random() * 1e6)}`;
  const originalLimit = process.env.VOICE_DAILY_EXTRACTIONS;
  let prisma: PrismaService;
  let roleId: string;
  let userId: string;
  let apiaryId: string;
  let colonyId: string;

  const mine = Prisma.sql`"ownerUserId" IN (SELECT "id" FROM "User" WHERE "email" LIKE ${`%${suffix}%`})`;
  const run = (model: FakeStructuredModel, signal?: AbortSignal) =>
    new VoiceExtractionService(prisma, model).runOnce(signal, {
      filter: mine,
    });
  const row = (id: string) =>
    prisma.voiceCapture.findUniqueOrThrow({ where: { id } });
  const events = (id: string, type: string) =>
    prisma.domainEvent.findMany({
      where: { aggregateType: 'VoiceCapture', aggregateId: id, type },
    });

  async function seed(
    data: Partial<Prisma.VoiceCaptureUncheckedCreateInput> = {},
  ): Promise<string> {
    const created = await prisma.voiceCapture.create({
      data: {
        ownerUserId: userId,
        colonyId,
        inspectionDate: new Date('2026-10-16'),
        source: 'TEXT',
        status: 'TRANSCRIBED',
        contentExpiresAt: new Date(Date.now() + 7 * 24 * 3_600_000),
        ...data,
      },
    });
    await prisma.voiceCapture.update({
      where: { id: created.id },
      data: {
        transcriptEnc: sealTranscript(created.id, {
          v: 1,
          text: SECRET_TEXT,
        }),
      },
    });
    return created.id;
  }

  async function ledgerCleanup(): Promise<void> {
    const own = await prisma.voiceCapture.findMany({
      where: { ownerUserId: userId },
      select: { id: true },
    });
    const ledgerIds = own.map((each) => each.id);
    await prisma.voiceExtractionCall.deleteMany({
      where: {
        OR: [
          { captureId: { in: ledgerIds } },
          { captureId: { endsWith: suffix } },
          { id: { endsWith: suffix } },
        ],
      },
    });
  }

  beforeAll(async () => {
    prisma = new PrismaService();
    await prisma.$connect();
    roleId = (
      await prisma.role.create({
        data: { key: `VOICE_EX_${suffix}`, name: `Voice ex ${suffix}` },
      })
    ).id;
    userId = (
      await prisma.user.create({
        data: {
          email: `voice-ex-${suffix}@example.com`,
          passwordHash: await bcrypt.hash('SuperSecret123!', 4),
          name: 'Voice extraction',
          roleId,
        },
      })
    ).id;
    apiaryId = (
      await prisma.apiary.create({ data: { name: `Voice ex ${suffix}` } })
    ).id;
    colonyId = (
      await prisma.colony.create({
        data: {
          code: `VEX-${suffix}`,
          apiaryId,
          colonyType: 'PRODUCTION_COLONY',
        },
      })
    ).id;
  });

  beforeEach(async () => {
    process.env.VOICE_DAILY_EXTRACTIONS = '10000';
    const own = await prisma.voiceCapture.findMany({
      where: { ownerUserId: userId },
      select: { id: true },
    });
    await prisma.domainEvent.deleteMany({
      where: {
        aggregateType: 'VoiceCapture',
        aggregateId: { in: own.map((c) => c.id) },
      },
    });
    await ledgerCleanup();
    await prisma.voiceCapture.deleteMany({ where: { ownerUserId: userId } });
  });

  afterAll(async () => {
    if (originalLimit === undefined) {
      delete process.env.VOICE_DAILY_EXTRACTIONS;
    } else {
      process.env.VOICE_DAILY_EXTRACTIONS = originalLimit;
    }
    const own = await prisma.voiceCapture.findMany({
      where: { ownerUserId: userId },
      select: { id: true },
    });
    await prisma.domainEvent.deleteMany({
      where: {
        aggregateType: 'VoiceCapture',
        aggregateId: { in: own.map((c) => c.id) },
      },
    });
    await ledgerCleanup();
    await prisma.voiceCapture.deleteMany({ where: { ownerUserId: userId } });
    await prisma.user.deleteMany({ where: { id: userId } });
    await prisma.role.deleteMany({ where: { id: roleId } });
    await prisma.colony.deleteMany({ where: { id: colonyId } });
    await prisma.apiary.deleteMany({ where: { id: apiaryId } });
    await prisma.$disconnect();
  });

  it('turns a transcribed note into READY with a sealed suggestion, tokens and one event', async () => {
    const id = await seed();
    const model = new FakeStructuredModel([FakeStructuredModel.ok(ANSWER)]);

    await run(model);

    const after = await row(id);
    expect(after).toMatchObject({
      status: 'READY',
      claimToken: null,
      claimedAt: null,
      model: 'claude-sonnet-5-5',
      inputTokens: 1500,
      outputTokens: 400,
    });
    expect(after.readyAt).not.toBeNull();
    expect(after.suggestionEnc).not.toContain('queenSeen');
    const stored = JSON.parse(
      openVoiceText(after.suggestionEnc ?? '', voiceAad(id, 'suggestion')),
    ) as { suggestions: { field: string }[] };
    expect(stored.suggestions.map((each) => each.field)).toEqual(['queenSeen']);
    const ready = await events(id, 'VoiceCaptureReady');
    expect(ready).toHaveLength(1);
    expect(JSON.stringify(ready[0].payload)).not.toContain('Königin');
  });

  it('parks the note with EXTRACT_DISABLED, burns no attempt and writes one operator event', async () => {
    const first = await seed();
    const second = await seed();

    await run(new FakeStructuredModel([], false));

    for (const id of [first, second]) {
      const after = await row(id);
      expect(after).toMatchObject({
        status: 'TRANSCRIBED',
        attempts: 0,
        lastErrorCode: 'EXTRACT_DISABLED',
      });
      expect(after.availableAt.getTime()).toBeGreaterThan(
        Date.now() + 3_000_000,
      );
    }
    const disabled = [
      ...(await events(first, 'VoiceCaptureFailed')),
      ...(await events(second, 'VoiceCaptureFailed')),
    ];
    expect(disabled).toHaveLength(1);
  });

  it('counts paid calls, not notes, against the daily limit', async () => {
    // Two paid calls today that never produced a READY note, plus calls that
    // never reached the model (they do not count).
    await prisma.$executeRaw`
      INSERT INTO "VoiceExtractionCall" ("id", "model", "outcome", "captureId")
      VALUES (${`ex-1-${suffix}`}, 'claude-sonnet-5-5', 'RETRY', ${`gone-${suffix}`}),
             (${`ex-2-${suffix}`}, 'claude-sonnet-5-5', 'REFUSED', ${`gone-${suffix}`}),
             (${`ex-3-${suffix}`}, 'claude-sonnet-5-5', 'AUTH', ${`gone-${suffix}`})`;
    const waiting = await seed();
    const own = await prisma.voiceExtractionCall.count({
      where: {
        captureId: `gone-${suffix}`,
        outcome: { notIn: ['AUTH', 'ABORTED'] },
      },
    });
    expect(own).toBe(2);
    const todayCounted = await prisma.$queryRaw<{ n: number }[]>`
      SELECT count(*)::int AS "n" FROM "VoiceExtractionCall"
      WHERE "outcome" NOT IN ('AUTH', 'ABORTED')
        AND "at" >= date_trunc('day', now() AT TIME ZONE 'UTC')`;
    process.env.VOICE_DAILY_EXTRACTIONS = String(todayCounted[0].n);
    const model = new FakeStructuredModel([FakeStructuredModel.ok(ANSWER)]);

    await run(model);

    expect(model.requests).toHaveLength(0);
    const after = await row(waiting);
    expect(after).toMatchObject({
      status: 'TRANSCRIBED',
      attempts: 0,
      lastErrorCode: 'DAILY_LIMIT',
    });
    expect(after.availableAt.getTime()).toBeGreaterThan(Date.now());
    // One call below the limit the note goes through and is booked.
    process.env.VOICE_DAILY_EXTRACTIONS = String(todayCounted[0].n + 1);
    await prisma.voiceCapture.update({
      where: { id: waiting },
      data: { availableAt: new Date(Date.now() - 1000) },
    });
    await run(model);
    expect(model.requests).toHaveLength(1);
    const booked = await prisma.voiceExtractionCall.findMany({
      where: { captureId: waiting },
    });
    expect(booked).toEqual([
      expect.objectContaining({
        outcome: 'OK',
        model: 'claude-sonnet-5-5',
        inputTokens: 1500,
        outputTokens: 400,
      }),
    ]);
  });

  it('retries three times with the backoff and then ends FAILED with one event', async () => {
    const id = await seed();
    const model = new FakeStructuredModel([
      { kind: 'error', code: 'RETRYABLE' },
    ]);

    await run(model);
    expect(await row(id)).toMatchObject({
      status: 'TRANSCRIBED',
      attempts: 1,
      lastErrorCode: 'EXTRACT_RETRY',
    });
    for (const attempts of [1, 2]) {
      await prisma.voiceCapture.update({
        where: { id },
        data: { availableAt: new Date(Date.now() - 1000) },
      });
      await run(model);
      expect((await row(id)).attempts).toBe(attempts + 1);
    }

    expect(await row(id)).toMatchObject({
      status: 'FAILED',
      attempts: 3,
      lastErrorCode: 'EXTRACT_FAILED',
    });
    expect(await events(id, 'VoiceCaptureFailed')).toHaveLength(1);
  });

  it('ends FAILED/REFUSED at once when the model declines', async () => {
    const id = await seed();

    await run(
      new FakeStructuredModel([
        {
          kind: 'refused',
          usage: { input: 10, output: 1, cacheRead: 0, cacheWrite: 0 },
          model: 'claude-sonnet-5-5',
        },
      ]),
    );

    expect(await row(id)).toMatchObject({
      status: 'FAILED',
      lastErrorCode: 'REFUSED',
    });
    expect(await events(id, 'VoiceCaptureFailed')).toHaveLength(1);
  });

  it('does not burn an attempt when shutdown cuts the call off', async () => {
    const id = await seed();
    const stop = new AbortController();
    const model = new FakeStructuredModel([]);
    jest.spyOn(model, 'extract').mockImplementation(() => {
      stop.abort();
      return Promise.resolve({ kind: 'error', code: 'RETRYABLE' });
    });

    await run(model, stop.signal);

    expect(await row(id)).toMatchObject({
      status: 'TRANSCRIBED',
      attempts: 0,
    });
  });

  it('a confirmed capture is not overwritten by a late extraction', async () => {
    const id = await seed();
    const model = new FakeStructuredModel([]);
    jest.spyOn(model, 'extract').mockImplementation(async () => {
      // The user saves the inspection while the model is still working.
      await prisma.voiceCapture.update({
        where: { id },
        data: { status: 'DISCARDED', transcriptEnc: null },
      });
      return FakeStructuredModel.ok(ANSWER);
    });

    await run(model);

    const after = await row(id);
    expect(after.status).toBe('DISCARDED');
    expect(after.suggestionEnc).toBeNull();
    expect(await events(id, 'VoiceCaptureReady')).toHaveLength(0);
  });

  it('puts a note whose worker died back, and fails it after the third stale release', async () => {
    const id = await seed({
      status: 'EXTRACTING',
      attempts: 2,
      claimToken: 'dead-worker',
    });
    await prisma.$executeRaw`
      UPDATE "VoiceCapture" SET "claimedAt" = now() - interval '10 minutes'
      WHERE "id" = ${id}`;

    await run(new FakeStructuredModel([], false));

    expect(await row(id)).toMatchObject({
      status: 'FAILED',
      lastErrorCode: 'EXTRACT_FAILED',
    });
    expect(await events(id, 'VoiceCaptureFailed')).toHaveLength(1);
  });
});