Files
OmniRoute/tests/unit/pricing-sync-memoization.test.ts
Praveen K Palaniswamy 65e81158ab fix(ollama): route models by advertised capability (#11088)
Landed with the design call resolved per the owner's pick — **option 1**: the synced store is now endpoint-agnostic (persistDiscoveredModels and managedModelImport no longer drop non-chat models at write time), and chat selectability moved to read time (auto-pool expansion in autoStrategy applies filterChatSelectableModels; the models-route projection already had its chatOnly filter). Your discovery test now passes end-to-end (3/3): /api/show capabilities persist per connection and image/embedding requests route through the advertising host.

Reconciliation notes: conflicted areas merged onto the current tip (adobe discovery import, requestedModel preflight signature, resolvedProvider fast-path coexists with the synced-route override — explicit resolution wins); carried base-red drains (#10055 memoization, #11071 test variants) dropped as already-landed; the managed-model-import exclusion test was propagated to the new contract (image/video models persist; the read filter still hides them from chat pickers — pinned by a new assertion). Full battery: 205/206 focused (the one red is a confirmed periodic-timer timing flake on the loaded devbox — 20/20 isolated), autoCombo vitest 30/30, combo suites 46/46, gates + typecheck clean.

Thank you @yourspraveen — the capability probe + routing design was right; it just needed the store contract opened up. Fixes #11087.
2026-08-23 11:45:01 -03:00

69 lines
2.3 KiB
TypeScript

import assert from "node:assert/strict";
import { describe, it, before, after, mock } from "node:test";
import { getDbInstance } from "../../src/lib/db/core.ts";
import {
getSyncedPricing,
saveSyncedPricing,
clearSyncedPricing,
} from "../../src/lib/pricingSync.ts";
describe("getSyncedPricing memoization", () => {
before(() => {
saveSyncedPricing({
openai: {
"gpt-4o": { input: 2.5, output: 10 },
},
});
});
after(() => {
try {
clearSyncedPricing();
} catch {
// ignore
}
});
it("returns the same object reference for repeated reads within the same cache version", () => {
const first = getSyncedPricing();
const second = getSyncedPricing();
const third = getSyncedPricing();
// The saturation bug rebuilt a fresh object on every call; resolveCatalogPricing()
// calls this per model, so a fresh object per call re-ran the SELECT + JSON.parse
// and rebuilt the findInsensitive() lowercase index per lookup (~400 warnings/s).
assert.equal(second, first);
assert.equal(third, first);
});
it("hits the DB once for repeated reads within the same cache version", () => {
const db = getDbInstance();
const prepareSpy = mock.method(db, "prepare");
const callsBefore = prepareSpy.mock.calls.length;
getSyncedPricing();
getSyncedPricing();
getSyncedPricing();
const callsAfter = prepareSpy.mock.calls.length;
prepareSpy.mock.restore();
// Memoized, 3 calls should cost at most 1 real DB round-trip (0 if a prior
// test already warmed the cache at the same version).
assert.ok(
callsAfter - callsBefore <= 1,
`expected at most 1 db.prepare() call across 3 reads, got ${callsAfter - callsBefore}`
);
});
it("returns a new reference with fresh data after a pricing write invalidates the cache", () => {
const warm = getSyncedPricing(); // warm the memo at the current cache version
saveSyncedPricing({
anthropic: { "claude-x": { input: 1, output: 2 } },
});
const pricing = getSyncedPricing();
assert.notEqual(pricing, warm, "invalidation must rebuild, not reuse the stale object");
assert.ok(pricing.anthropic, "cache should reflect the write, not a stale snapshot");
assert.equal(pricing.anthropic["claude-x"].input, 1);
});
});