mirror of
https://github.com/mauriceboe/TREK.git
synced 2026-10-21 14:36:31 +00:00
b1c05bca83
Whether the AI Parsing model reads images is an admin setting now, and it lands with the two features that send a photo, so it has something to control. The setting is `vision: 'auto' | 'on' | 'off'` on the addon config. On Automatic a local Ollama is asked what the model can do (`/api/show`), and a cloud provider means no; Yes and No are applied as set. The old `multimodal` flag is not read: the admin screens wrote it false on every save, so a stored false meant nothing. The per-user toggle keeps its on/off meaning and is read by the extraction now. The lookup is server-side only: capped read, validated body, never throws, cached per server and model. GET /api/llm/capabilities answers it for the signed-in caller, and both photo pickers below ask it. The booking import takes JPG, PNG and WEBP when the model reads images. The photo goes as an image to every provider (image_url, an Anthropic image block, Ollama's native chat with `images`), capped at 1600px, and skips KDE Itinerary. A scanned PDF, which used to reach the model as empty text, has its first two pages drawn by pdf-parse and sent the same way. The rest is the import as it was: background job, review, the file attached to the booking, the linked cost. On a real qwen3.5:4b a train ticket came back as bare clock times until the router gave the model every type's field meanings, and its price as an object inside a string, which read as 4900; both are handled. Costs gets Scan receipt: a dialog laid out like the booking import's takes one photo, read on the import job queue, then reviewed in the existing expense editor pre-filled with merchant, total, currency, day and lines, the photo waiting in Receipts & Invoices to be uploaded on save. No table, no migration, no category guess, no document-type detection.
99 lines
4.3 KiB
TypeScript
99 lines
4.3 KiB
TypeScript
import { describe, it, expect, vi, beforeEach } from 'vitest';
|
|
|
|
// The client goes through safeFetchLlm (SSRF guard: blocks the cloud-metadata
|
|
// range, allows a local/LAN Ollama). Mock it so the tests never do a real DNS
|
|
// lookup; its (url, init) signature matches the raw fetch it wraps, so the
|
|
// recorded-call assertions are unchanged.
|
|
const { safeFetchLlmMock } = vi.hoisted(() => ({ safeFetchLlmMock: vi.fn() }));
|
|
vi.mock('../../../../src/utils/ssrfGuard', () => ({ safeFetchLlm: safeFetchLlmMock }));
|
|
|
|
import { toNativeBase, extractEnforced } from '../../../../src/nest/llm-parse/router/ollama-format.client';
|
|
|
|
function mockFetch(impl: (url: string, init: RequestInit) => Promise<Response> | Response) {
|
|
safeFetchLlmMock.mockImplementation(impl as unknown as typeof fetch);
|
|
return safeFetchLlmMock;
|
|
}
|
|
|
|
function jsonResponse(body: unknown, ok = true, status = 200): Response {
|
|
return { ok, status, json: async () => body, text: async () => JSON.stringify(body) } as unknown as Response;
|
|
}
|
|
|
|
const INPUT = {
|
|
baseUrl: 'http://ollama:11434/v1',
|
|
model: 'qwen3:8b',
|
|
system: 'sys',
|
|
user: 'doc',
|
|
schema: { type: 'object' as const },
|
|
};
|
|
|
|
beforeEach(() => safeFetchLlmMock.mockReset());
|
|
|
|
describe('toNativeBase', () => {
|
|
it('strips a /v1 suffix and trailing slashes', () => {
|
|
expect(toNativeBase('http://ollama:11434/v1')).toBe('http://ollama:11434');
|
|
expect(toNativeBase('http://ollama:11434/v1/')).toBe('http://ollama:11434');
|
|
expect(toNativeBase('http://ollama:11434/')).toBe('http://ollama:11434');
|
|
expect(toNativeBase('http://ollama:11434')).toBe('http://ollama:11434');
|
|
});
|
|
});
|
|
|
|
describe('extractEnforced', () => {
|
|
it('posts to the native /api/chat with the grammar format and thinking disabled', async () => {
|
|
const fetchFn = mockFetch(() => jsonResponse({ message: { content: '{"name":"Hotel"}' } }));
|
|
const out = await extractEnforced(INPUT);
|
|
expect(out).toEqual({ name: 'Hotel' });
|
|
const [url, init] = fetchFn.mock.calls[0];
|
|
expect(url).toBe('http://ollama:11434/api/chat');
|
|
const body = JSON.parse((init as RequestInit).body as string);
|
|
expect(body.format).toEqual({ type: 'object' });
|
|
expect(body.think).toBe(false);
|
|
expect(body.stream).toBe(false);
|
|
expect(body.options.temperature).toBe(0);
|
|
expect((init as RequestInit).headers).not.toHaveProperty('authorization');
|
|
});
|
|
|
|
it('attaches images to the user turn only when given', async () => {
|
|
const fetchFn = mockFetch(() => jsonResponse({ message: { content: '{}' } }));
|
|
await extractEnforced({ ...INPUT, images: ['aW1n'] });
|
|
await extractEnforced(INPUT);
|
|
const userTurn = (i: number) => JSON.parse((fetchFn.mock.calls[i][1] as RequestInit).body as string).messages[1];
|
|
expect(userTurn(0).images).toEqual(['aW1n']);
|
|
expect(userTurn(1)).not.toHaveProperty('images');
|
|
});
|
|
|
|
it('sends a bearer header only when an apiKey is given', async () => {
|
|
const fetchFn = mockFetch(() => jsonResponse({ message: { content: '{}' } }));
|
|
await extractEnforced({ ...INPUT, apiKey: 'sk-123', numPredict: 900, numCtx: 16000 });
|
|
const init = fetchFn.mock.calls[0][1] as RequestInit;
|
|
expect((init.headers as Record<string, string>).authorization).toBe('Bearer sk-123');
|
|
const body = JSON.parse(init.body as string);
|
|
expect(body.options.num_predict).toBe(900);
|
|
expect(body.options.num_ctx).toBe(16000);
|
|
});
|
|
|
|
it('strips a ```json code fence before parsing', async () => {
|
|
mockFetch(() => jsonResponse({ message: { content: '```json\n{"a":1}\n```' } }));
|
|
expect(await extractEnforced(INPUT)).toEqual({ a: 1 });
|
|
});
|
|
|
|
it('returns null when the content parses to a non-object', async () => {
|
|
mockFetch(() => jsonResponse({ message: { content: '"just a string"' } }));
|
|
expect(await extractEnforced(INPUT)).toBeNull();
|
|
});
|
|
|
|
it('returns null for unparseable content', async () => {
|
|
mockFetch(() => jsonResponse({ message: { content: 'not json at all' } }));
|
|
expect(await extractEnforced(INPUT)).toBeNull();
|
|
});
|
|
|
|
it('returns null when the response has no content', async () => {
|
|
mockFetch(() => jsonResponse({ message: {} }));
|
|
expect(await extractEnforced(INPUT)).toBeNull();
|
|
});
|
|
|
|
it('throws with the status when Ollama responds non-ok', async () => {
|
|
mockFetch(() => jsonResponse({ error: 'model not found' }, false, 404));
|
|
await expect(extractEnforced(INPUT)).rejects.toThrow(/Ollama \/api\/chat failed \(404\)/);
|
|
});
|
|
});
|