feat(extract): extract data using LLM

server/tests/unit/nest/booking-import/booking-import.controller.test.ts

@@ -0,0 +1,61 @@
+import { describe, it, expect, vi, beforeEach } from 'vitest';
+import { HttpException } from '@nestjs/common';
+import { BookingImportController } from '../../../../src/nest/booking-import/booking-import.controller';
+import type { BookingImportService } from '../../../../src/nest/booking-import/booking-import.service';
+import type { User } from '../../../../src/types';
+
+const user = { id: 1, role: 'user' } as User;
+const file = (name = 'a.pdf') => ({ originalname: name, buffer: Buffer.from('x') } as Express.Multer.File);
+
+function make(over: Partial<BookingImportService> = {}) {
+  const svc = {
+    verifyTripAccess: vi.fn(() => ({ user_id: 1 })),
+    canEdit: vi.fn(() => true),
+    isAvailable: vi.fn(() => true),
+    aiAvailable: vi.fn(() => true),
+    preview: vi.fn(async () => ({ items: [], warnings: [], files: [] })),
+    ...over,
+  } as unknown as BookingImportService;
+  return { c: new BookingImportController(svc), svc };
+}
+
+async function status(fn: () => Promise<unknown>): Promise<number> {
+  try { await fn(); } catch (e) { expect(e).toBeInstanceOf(HttpException); return (e as HttpException).getStatus(); }
+  throw new Error('expected throw');
+}
+
+beforeEach(() => vi.clearAllMocks());
+
+describe('BookingImportController.preview', () => {
+  it('rejects an invalid mode with 400', async () => {
+    const { c } = make();
+    expect(await status(() => c.preview(user, 't1', [file()], 'bogus'))).toBe(400);
+  });
+
+  it('returns 409 for force-ai when AI is not configured', async () => {
+    const { c } = make({ aiAvailable: vi.fn(() => false) as any });
+    expect(await status(() => c.preview(user, 't1', [file()], 'force-ai'))).toBe(409);
+  });
+
+  it('returns 503 for no-ai when the extractor is unavailable', async () => {
+    const { c } = make({ isAvailable: vi.fn(() => false) as any });
+    expect(await status(() => c.preview(user, 't1', [file()], 'no-ai'))).toBe(503);
+  });
+
+  it('returns 400 when no files are uploaded', async () => {
+    const { c } = make();
+    expect(await status(() => c.preview(user, 't1', [], 'no-ai'))).toBe(400);
+  });
+
+  it('passes the parsed mode and user id through to the service', async () => {
+    const { c, svc } = make();
+    await c.preview(user, 't1', [file()], 'fallback-on-empty');
+    expect(svc.preview).toHaveBeenCalledWith([expect.anything()], 'fallback-on-empty', 1);
+  });
+
+  it('defaults the mode to no-ai when omitted', async () => {
+    const { c, svc } = make();
+    await c.preview(user, 't1', [file()], undefined);
+    expect(svc.preview).toHaveBeenCalledWith([expect.anything()], 'no-ai', 1);
+  });
+});

server/tests/unit/nest/booking-import/booking-import.service.test.ts

@@ -0,0 +1,79 @@
+import { describe, it, expect, vi, beforeEach } from 'vitest';
+import { HttpException } from '@nestjs/common';
+
+// Mock the heavy side-effect imports so the service module loads cleanly; the
+// preview() path under test only touches the extractor + llmParse deps.
+vi.mock('../../../../src/db/database', () => ({ db: { prepare: vi.fn() }, closeDb: () => {}, reinitialize: () => {} }));
+vi.mock('../../../../src/websocket', () => ({ broadcast: vi.fn() }));
+vi.mock('../../../../src/services/permissions', () => ({ checkPermission: vi.fn(() => true) }));
+vi.mock('../../../../src/services/tripAccess', () => ({ verifyTripAccess: vi.fn() }));
+vi.mock('../../../../src/services/reservationService', () => ({ createReservation: vi.fn() }));
+vi.mock('../../../../src/services/placeService', () => ({ createPlace: vi.fn() }));
+vi.mock('../../../../src/services/mapsService', () => ({ searchNominatim: vi.fn() }));
+
+import { BookingImportService } from '../../../../src/nest/booking-import/booking-import.service';
+
+const HOTEL_KI = { '@type': 'LodgingReservation', reservationNumber: 'ABC', reservationFor: { name: 'Hotel X' }, checkinTime: '2026-06-11T15:00', checkoutTime: '2026-06-12T11:00' };
+const file = (name = 'a.pdf') => ({ buffer: Buffer.from('x'), originalname: name } as any);
+
+function make(opts: { kit?: boolean; ai?: boolean; extract?: any; parse?: any }) {
+  const extractor = { isAvailable: () => opts.kit ?? false, extract: vi.fn(opts.extract ?? (async () => [])) };
+  const llmParse = { isAvailable: () => opts.ai ?? false, parse: vi.fn(opts.parse ?? (async () => ({ kiItems: [], warnings: [] }))) };
+  return { svc: new BookingImportService(extractor as any, llmParse as any), extractor, llmParse };
+}
+
+beforeEach(() => vi.clearAllMocks());
+
+describe('BookingImportService.preview', () => {
+  it('no-ai: maps kitinerary items, does not force needs_review, reports aiUsed:false', async () => {
+    const { svc, llmParse } = make({ kit: true, ai: false, extract: async () => [HOTEL_KI] });
+    const res = await svc.preview([file()], 'no-ai', 1);
+    expect(res.items).toHaveLength(1);
+    expect(res.items[0].needs_review).toBeFalsy();
+    expect(res.files).toEqual([{ fileName: 'a.pdf', aiAvailable: false, aiUsed: false }]);
+    expect(llmParse.parse).not.toHaveBeenCalled();
+  });
+
+  it('throws 503 when neither parser is available', async () => {
+    const { svc } = make({ kit: false, ai: false });
+    try {
+      await svc.preview([file()], 'no-ai', 1);
+      throw new Error('expected throw');
+    } catch (err) {
+      expect(err).toBeInstanceOf(HttpException);
+      expect((err as HttpException).getStatus()).toBe(503);
+    }
+  });
+
+  it('fallback-on-empty: runs the LLM when kitinerary finds nothing and flags needs_review', async () => {
+    const { svc, extractor, llmParse } = make({
+      kit: true, ai: true,
+      extract: async () => [],
+      parse: async () => ({ kiItems: [HOTEL_KI], warnings: [] }),
+    });
+    const res = await svc.preview([file()], 'fallback-on-empty', 1);
+    expect(extractor.extract).toHaveBeenCalled();
+    expect(llmParse.parse).toHaveBeenCalled();
+    expect(res.items).toHaveLength(1);
+    expect(res.items[0].needs_review).toBe(true);
+    expect(res.files![0]).toEqual({ fileName: 'a.pdf', aiAvailable: true, aiUsed: true });
+  });
+
+  it('fallback-on-empty: skips the LLM when kitinerary already found items', async () => {
+    const { svc, llmParse } = make({ kit: true, ai: true, extract: async () => [HOTEL_KI] });
+    const res = await svc.preview([file()], 'fallback-on-empty', 1);
+    expect(llmParse.parse).not.toHaveBeenCalled();
+    expect(res.files![0].aiUsed).toBe(false);
+  });
+
+  it('force-ai: skips kitinerary entirely and uses the LLM', async () => {
+    const { svc, extractor, llmParse } = make({
+      kit: true, ai: true,
+      parse: async () => ({ kiItems: [HOTEL_KI], warnings: [] }),
+    });
+    const res = await svc.preview([file()], 'force-ai', 1);
+    expect(extractor.extract).not.toHaveBeenCalled();
+    expect(llmParse.parse).toHaveBeenCalled();
+    expect(res.items[0].needs_review).toBe(true);
+  });
+});

server/tests/unit/nest/llm-parse/clients.test.ts

@@ -0,0 +1,96 @@
+import { describe, it, expect, vi, beforeEach } from 'vitest';
+import { OpenAiCompatibleClient } from '../../../../src/nest/llm-parse/clients/openai-compatible.client';
+import { AnthropicClient } from '../../../../src/nest/llm-parse/clients/anthropic.client';
+import type { LlmExtractionInput } from '../../../../src/nest/llm-parse/llm-provider.interface';
+
+const baseInput: LlmExtractionInput = {
+  prompt: 'system',
+  jsonSchema: { type: 'object' },
+  model: 'm',
+  text: 'Flight AB123',
+};
+
+function mockFetch(impl: (url: string, init: RequestInit) => Promise<Response> | Response) {
+  const fn = vi.fn(impl as any);
+  vi.stubGlobal('fetch', fn);
+  return fn;
+}
+
+function jsonResponse(body: unknown, ok = true, status = 200): Response {
+  return { ok, status, json: async () => body, text: async () => JSON.stringify(body) } as unknown as Response;
+}
+
+beforeEach(() => vi.unstubAllGlobals());
+
+describe('OpenAiCompatibleClient', () => {
+  it('posts to {baseUrl}/chat/completions and returns the reservations array', async () => {
+    const fetchFn = mockFetch(() =>
+      jsonResponse({ choices: [{ message: { content: JSON.stringify({ reservations: [{ '@type': 'FlightReservation' }] }) } }] }),
+    );
+    const out = await new OpenAiCompatibleClient().extract({ ...baseInput, baseUrl: 'http://localhost:11434/v1/' });
+    expect(out).toEqual([{ '@type': 'FlightReservation' }]);
+    expect(fetchFn.mock.calls[0][0]).toBe('http://localhost:11434/v1/chat/completions');
+  });
+
+  it('tolerates code-fenced JSON', async () => {
+    mockFetch(() =>
+      jsonResponse({ choices: [{ message: { content: '```json\n{"reservations":[{"@type":"TrainReservation"}]}\n```' } }] }),
+    );
+    const out = await new OpenAiCompatibleClient().extract(baseInput);
+    expect(out).toEqual([{ '@type': 'TrainReservation' }]);
+  });
+
+  it('returns [] on malformed content', async () => {
+    mockFetch(() => jsonResponse({ choices: [{ message: { content: 'not json' } }] }));
+    expect(await new OpenAiCompatibleClient().extract(baseInput)).toEqual([]);
+  });
+
+  it('throws on non-2xx', async () => {
+    mockFetch(() => jsonResponse({ error: 'bad' }, false, 401));
+    await expect(new OpenAiCompatibleClient().extract(baseInput)).rejects.toThrow(/401/);
+  });
+
+  it('sends an image natively as image_url but never a file/pdf part', async () => {
+    const fetchFn = mockFetch(() => jsonResponse({ choices: [{ message: { content: '{"reservations":[]}' } }] }));
+    await new OpenAiCompatibleClient().extract({ ...baseInput, file: { mimeType: 'image/png', data: Buffer.from('IMG') } });
+    let parts = JSON.parse((fetchFn.mock.calls[0][1] as RequestInit).body as string).messages[1].content;
+    expect(parts.some((p: any) => p.type === 'image_url')).toBe(true);
+    expect(parts.some((p: any) => p.type === 'file')).toBe(false);
+
+    // A PDF must NOT be sent as a content part (Ollama rejects it).
+    await new OpenAiCompatibleClient().extract({ ...baseInput, file: { mimeType: 'application/pdf', data: Buffer.from('PDF') } });
+    parts = JSON.parse((fetchFn.mock.calls[1][1] as RequestInit).body as string).messages[1].content;
+    expect(parts.every((p: any) => p.type !== 'file' && p.type !== 'image_url')).toBe(true);
+  });
+});
+
+describe('AnthropicClient', () => {
+  it('forces the emit_reservations tool and reads its input', async () => {
+    const fetchFn = mockFetch(() =>
+      jsonResponse({ stop_reason: 'tool_use', content: [{ type: 'tool_use', name: 'emit_reservations', input: { reservations: [{ '@type': 'LodgingReservation' }] } }] }),
+    );
+    const out = await new AnthropicClient().extract(baseInput);
+    expect(out).toEqual([{ '@type': 'LodgingReservation' }]);
+    const body = JSON.parse((fetchFn.mock.calls[0][1] as RequestInit).body as string);
+    expect(body.tool_choice).toEqual({ type: 'tool', name: 'emit_reservations' });
+    expect(body.tools[0].name).toBe('emit_reservations');
+  });
+
+  it('throws on a refusal stop_reason', async () => {
+    mockFetch(() => jsonResponse({ stop_reason: 'refusal', content: [] }));
+    await expect(new AnthropicClient().extract(baseInput)).rejects.toThrow(/declined/i);
+  });
+
+  it('throws on non-2xx', async () => {
+    mockFetch(() => jsonResponse({ error: 'bad' }, false, 500));
+    await expect(new AnthropicClient().extract(baseInput)).rejects.toThrow(/500/);
+  });
+
+  it('sends a native pdf as a base64 document block', async () => {
+    const fetchFn = mockFetch(() => jsonResponse({ content: [{ type: 'tool_use', name: 'emit_reservations', input: { reservations: [] } }] }));
+    await new AnthropicClient().extract({ ...baseInput, file: { mimeType: 'application/pdf', data: Buffer.from('PDF') } });
+    const body = JSON.parse((fetchFn.mock.calls[0][1] as RequestInit).body as string);
+    const blocks = body.messages[0].content;
+    expect(blocks.some((b: any) => b.type === 'document' && b.source.type === 'base64')).toBe(true);
+  });
+});

server/tests/unit/nest/llm-parse/llm-config.resolver.test.ts

@@ -0,0 +1,67 @@
+import { describe, it, expect, vi, beforeEach } from 'vitest';
+
+const { dbMock } = vi.hoisted(() => {
+  const stmt = { get: vi.fn() };
+  return { dbMock: { prepare: vi.fn(() => stmt), _stmt: stmt } };
+});
+vi.mock('../../../../src/db/database', () => ({ db: dbMock, closeDb: () => {}, reinitialize: () => {} }));
+
+const { isAddonEnabled } = vi.hoisted(() => ({ isAddonEnabled: vi.fn() }));
+vi.mock('../../../../src/services/adminService', () => ({ isAddonEnabled }));
+
+const { getUserSettings, getDecryptedUserSetting } = vi.hoisted(() => ({
+  getUserSettings: vi.fn(() => ({}) as Record<string, unknown>),
+  getDecryptedUserSetting: vi.fn(() => null as string | null),
+}));
+vi.mock('../../../../src/services/settingsService', () => ({ getUserSettings, getDecryptedUserSetting }));
+
+import { resolveLlmConfig } from '../../../../src/nest/llm-parse/llm-config.resolver';
+
+function setInstanceConfig(config: unknown) {
+  dbMock._stmt.get.mockReturnValue(config === undefined ? undefined : { config: JSON.stringify(config) });
+}
+
+beforeEach(() => {
+  vi.clearAllMocks();
+  isAddonEnabled.mockReturnValue(true);
+  setInstanceConfig(undefined);
+  getUserSettings.mockReturnValue({});
+  getDecryptedUserSetting.mockReturnValue(null);
+});
+
+describe('resolveLlmConfig', () => {
+  it('returns null when the addon is disabled', () => {
+    isAddonEnabled.mockReturnValue(false);
+    expect(resolveLlmConfig(1)).toBeNull();
+  });
+
+  it('uses instance config when present (and decrypts the key)', () => {
+    setInstanceConfig({ provider: 'anthropic', model: 'claude-opus-4-8', apiKey: 'sk-plain', multimodal: true });
+    expect(resolveLlmConfig(1)).toEqual({
+      provider: 'anthropic',
+      model: 'claude-opus-4-8',
+      baseUrl: undefined,
+      apiKey: 'sk-plain',
+      multimodal: true,
+    });
+  });
+
+  it('falls back to per-user config when instance config is incomplete', () => {
+    setInstanceConfig({ provider: 'anthropic' }); // no model → not usable
+    getUserSettings.mockReturnValue({ llm_provider: 'local', llm_model: 'nuextract', llm_base_url: 'http://x/v1', llm_multimodal: true });
+    getDecryptedUserSetting.mockReturnValue('user-key');
+    expect(resolveLlmConfig(7)).toEqual({
+      provider: 'local',
+      model: 'nuextract',
+      baseUrl: 'http://x/v1',
+      apiKey: 'user-key',
+      multimodal: true,
+    });
+    expect(getDecryptedUserSetting).toHaveBeenCalledWith(7, 'llm_api_key');
+  });
+
+  it('returns null when neither instance nor user config is usable', () => {
+    getUserSettings.mockReturnValue({ llm_provider: 'openai' }); // no model
+    expect(resolveLlmConfig(1)).toBeNull();
+  });
+});

server/tests/unit/nest/llm-parse/llm-local.service.test.ts

@@ -0,0 +1,60 @@
+import { describe, it, expect, vi, beforeEach } from 'vitest';
+import { HttpException } from '@nestjs/common';
+import { LlmLocalService } from '../../../../src/nest/llm-parse/llm-local.service';
+
+const svc = () => new LlmLocalService();
+
+function mockFetch(impl: any) {
+  const fn = vi.fn(impl);
+  vi.stubGlobal('fetch', fn);
+  return fn;
+}
+
+beforeEach(() => vi.unstubAllGlobals());
+
+describe('LlmLocalService.ollamaRoot', () => {
+  it('strips a trailing /v1 and slashes', () => {
+    expect(svc().ollamaRoot('http://localhost:11434/v1')).toBe('http://localhost:11434');
+    expect(svc().ollamaRoot('http://localhost:11434/v1/')).toBe('http://localhost:11434');
+    expect(svc().ollamaRoot('http://host:1/')).toBe('http://host:1');
+  });
+
+  it('defaults when no base URL is given', () => {
+    expect(svc().ollamaRoot(undefined)).toBe('http://localhost:11434');
+  });
+
+  it('rejects non-http(s) and invalid URLs', () => {
+    expect(() => svc().ollamaRoot('ftp://x')).toThrow(HttpException);
+    expect(() => svc().ollamaRoot('not a url')).toThrow(HttpException);
+  });
+});
+
+describe('LlmLocalService.listModels', () => {
+  it('returns named models from /api/tags', async () => {
+    const fetchFn = mockFetch(async () => ({ ok: true, json: async () => ({ models: [{ name: 'nuextract', size: 100 }, { name: '' }] }) }));
+    const out = await svc().listModels('http://localhost:11434/v1');
+    expect(out.models).toEqual([{ name: 'nuextract', size: 100 }]);
+    expect(fetchFn.mock.calls[0][0]).toBe('http://localhost:11434/api/tags');
+  });
+
+  it('502s when the server is unreachable', async () => {
+    mockFetch(async () => { throw new Error('ECONNREFUSED'); });
+    await expect(svc().listModels('http://localhost:11434')).rejects.toThrow(HttpException);
+  });
+});
+
+describe('LlmLocalService.pull', () => {
+  it('requires a model', async () => {
+    await expect(svc().pull('http://localhost:11434', '')).rejects.toThrow(HttpException);
+  });
+
+  it('posts to /api/pull and returns the stream body', async () => {
+    const body = {} as ReadableStream<Uint8Array>;
+    const fetchFn = mockFetch(async () => ({ ok: true, body }));
+    const out = await svc().pull('http://localhost:11434/v1', 'nuextract');
+    expect(out).toBe(body);
+    expect(fetchFn.mock.calls[0][0]).toBe('http://localhost:11434/api/pull');
+    const init = fetchFn.mock.calls[0][1];
+    expect(JSON.parse(init.body)).toEqual({ model: 'nuextract', stream: true });
+  });
+});

server/tests/unit/nest/llm-parse/llm-parse.service.test.ts

@@ -0,0 +1,116 @@
+import { describe, it, expect, vi, beforeEach } from 'vitest';
+
+const { resolveLlmConfig } = vi.hoisted(() => ({ resolveLlmConfig: vi.fn() }));
+vi.mock('../../../../src/nest/llm-parse/llm-config.resolver', () => ({ resolveLlmConfig }));
+
+const { createLlmClient, extract } = vi.hoisted(() => {
+  const extract = vi.fn();
+  return { createLlmClient: vi.fn(() => ({ extract })), extract };
+});
+vi.mock('../../../../src/nest/llm-parse/llm-client.factory', () => ({ createLlmClient }));
+
+const { extractText } = vi.hoisted(() => ({ extractText: vi.fn(async () => 'Flight AB123') }));
+vi.mock('../../../../src/nest/llm-parse/text-extract', async (orig) => {
+  const actual = await orig() as Record<string, unknown>;
+  return { ...actual, extractText };
+});
+
+import { LlmParseService } from '../../../../src/nest/llm-parse/llm-parse.service';
+
+const cfg = (over: Record<string, unknown> = {}) => ({ provider: 'openai', model: 'm', multimodal: false, ...over });
+const svc = () => new LlmParseService();
+const file = (name: string, body = 'Flight AB123') => ({ buffer: Buffer.from(body), originalName: name });
+
+beforeEach(() => {
+  vi.clearAllMocks();
+  resolveLlmConfig.mockReturnValue(cfg());
+  extract.mockResolvedValue([{ '@type': 'FlightReservation' }]);
+  extractText.mockResolvedValue('Flight AB123');
+});
+
+describe('LlmParseService', () => {
+  it('isAvailable reflects whether a config resolves', () => {
+    resolveLlmConfig.mockReturnValueOnce(null);
+    expect(svc().isAvailable(1)).toBe(false);
+    expect(svc().isAvailable(1)).toBe(true);
+  });
+
+  it('returns a not-configured warning when no config resolves', async () => {
+    resolveLlmConfig.mockReturnValue(null);
+    const res = await svc().parse(file('a.txt'), 1);
+    expect(res.kiItems).toEqual([]);
+    expect(res.warnings[0]).toMatch(/not configured/i);
+    expect(extract).not.toHaveBeenCalled();
+  });
+
+  it('sends extracted text for a text-like file', async () => {
+    const res = await svc().parse(file('a.txt'), 1);
+    expect(res.kiItems).toEqual([{ '@type': 'FlightReservation' }]);
+    const input = extract.mock.calls[0][0];
+    expect(input.text).toBe('Flight AB123');
+    expect(input.file).toBeUndefined();
+  });
+
+  it('extracts text for a pdf on the OpenAI-compatible/local path (no native bytes)', async () => {
+    extractText.mockResolvedValue('Hotel X');
+    await svc().parse(file('a.pdf', '%PDF'), 1);
+    const input = extract.mock.calls[0][0];
+    expect(input.text).toBe('Hotel X');
+    expect(input.file).toBeUndefined();
+  });
+
+  it('sends a pdf as native bytes only for Anthropic', async () => {
+    resolveLlmConfig.mockReturnValue(cfg({ provider: 'anthropic' }));
+    await svc().parse(file('a.pdf', '%PDF'), 1);
+    const input = extract.mock.calls[0][0];
+    expect(input.file).toEqual({ mimeType: 'application/pdf', data: expect.any(Buffer) });
+    expect(input.text).toBeUndefined();
+    expect(extractText).not.toHaveBeenCalled();
+  });
+
+  it('warns when a pdf yields no readable text (e.g. a scan)', async () => {
+    extractText.mockResolvedValue('   ');
+    const res = await svc().parse(file('a.pdf', '%PDF'), 1);
+    expect(res.kiItems).toEqual([]);
+    expect(res.warnings[0]).toMatch(/no readable text/i);
+    expect(extract).not.toHaveBeenCalled();
+  });
+
+  it('folds flattened type fields into reservationFor (small-model output)', async () => {
+    extract.mockResolvedValue([{
+      '@type': 'FlightReservation',
+      reservationNumber: 'ABC',
+      flightNumber: 'EZY1357',
+      airline: { iataCode: 'EG' },
+      departureAirport: { iataCode: 'GEG' },
+      arrivalAirport: { iataCode: 'AMS' },
+      departureTime: '2026-06-11T10:00:00',
+    }]);
+    const res = await svc().parse(file('a.txt'), 1);
+    const item = res.kiItems[0] as any;
+    expect(item.reservationNumber).toBe('ABC');
+    expect(item.reservationFor).toMatchObject({ flightNumber: 'EZY1357', departureAirport: { iataCode: 'GEG' } });
+    // root-level keys are not duplicated into reservationFor
+    expect(item.reservationFor.reservationNumber).toBeUndefined();
+  });
+
+  it('leaves already-nested reservationFor untouched', async () => {
+    extract.mockResolvedValue([{ '@type': 'FlightReservation', reservationFor: { flightNumber: 'X1' } }]);
+    const res = await svc().parse(file('a.txt'), 1);
+    expect((res.kiItems[0] as any).reservationFor).toEqual({ flightNumber: 'X1' });
+  });
+
+  it('drops nodes without a string @type and warns', async () => {
+    extract.mockResolvedValue([{ '@type': 'FlightReservation' }, { foo: 'bar' }]);
+    const res = await svc().parse(file('a.txt'), 1);
+    expect(res.kiItems).toEqual([{ '@type': 'FlightReservation' }]);
+    expect(res.warnings.some(w => /unrecognized/i.test(w))).toBe(true);
+  });
+
+  it('degrades to a warning when the client throws', async () => {
+    extract.mockRejectedValue(new Error('boom'));
+    const res = await svc().parse(file('a.txt'), 1);
+    expect(res.kiItems).toEqual([]);
+    expect(res.warnings[0]).toMatch(/AI parsing failed/i);
+  });
+});

server/tests/unit/nest/llm-parse/llm-prompt.test.ts

@@ -0,0 +1,26 @@
+import { describe, it, expect } from 'vitest';
+import { buildSystemPrompt, KI_RESERVATION_JSON_SCHEMA } from '../../../../src/nest/llm-parse/llm-prompt';
+import { KI_RESERVATION_TYPES } from '@trek/shared';
+
+describe('llm-prompt', () => {
+  it('names every recognized @type the mapper supports', () => {
+    const prompt = buildSystemPrompt();
+    for (const t of KI_RESERVATION_TYPES) expect(prompt).toContain(t);
+  });
+
+  it('instructs JSON-only output wrapped in reservations', () => {
+    const prompt = buildSystemPrompt();
+    expect(prompt).toMatch(/"reservations"/);
+    expect(prompt.toLowerCase()).toContain('iso 8601');
+  });
+
+  it('exposes a strict-safe object-root JSON schema enumerating the types', () => {
+    const schema = KI_RESERVATION_JSON_SCHEMA as any;
+    expect(schema.type).toBe('object');
+    expect(schema.additionalProperties).toBe(false);
+    expect(schema.required).toContain('reservations');
+    const item = schema.properties.reservations.items;
+    expect(item.properties['@type'].enum).toEqual([...KI_RESERVATION_TYPES]);
+    expect(item.required).toContain('@type');
+  });
+});

server/tests/unit/nest/llm-parse/text-extract.test.ts

@@ -0,0 +1,40 @@
+import { describe, it, expect, vi } from 'vitest';
+
+const { getText } = vi.hoisted(() => ({ getText: vi.fn(async () => ({ text: 'Hotel X — confirmation ABC' })) }));
+vi.mock('pdf-parse', () => ({
+  PDFParse: class {
+    getText = getText;
+    destroy = vi.fn(async () => {});
+  },
+}));
+
+import { isTextLike, isPdf, extractText } from '../../../../src/nest/llm-parse/text-extract';
+
+describe('text-extract', () => {
+  it('classifies text-like and pdf extensions', () => {
+    expect(isTextLike('a.txt')).toBe(true);
+    expect(isTextLike('a.html')).toBe(true);
+    expect(isTextLike('a.eml')).toBe(true);
+    expect(isTextLike('a.pdf')).toBe(false);
+    expect(isPdf('a.PDF')).toBe(true);
+    expect(isPdf('a.txt')).toBe(false);
+  });
+
+  it('decodes plain text', async () => {
+    expect(await extractText(Buffer.from('hello world'), 'a.txt')).toBe('hello world');
+  });
+
+  it('strips markup from html/eml', async () => {
+    const html = '<html><style>x{}</style><body><p>Flight AB123</p><script>1</script></body></html>';
+    const out = await extractText(Buffer.from(html), 'a.html');
+    expect(out).toContain('Flight AB123');
+    expect(out).not.toContain('<p>');
+    expect(out).not.toContain('x{}');
+  });
+
+  it('extracts the embedded text layer from a pdf', async () => {
+    const out = await extractText(Buffer.from('%PDF-1.4'), 'a.pdf');
+    expect(out).toBe('Hotel X — confirmation ABC');
+    expect(getText).toHaveBeenCalled();
+  });
+});