mirror of
https://github.com/diegosouzapw/OmniRoute.git
synced 2026-09-14 02:42:24 +03:00
65e81158ab
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.
71 lines
2.2 KiB
TypeScript
71 lines
2.2 KiB
TypeScript
import test from "node:test";
|
|
import assert from "node:assert/strict";
|
|
|
|
import { enforceOutputTokenBudget } from "../../open-sse/handlers/chatCore/outputTokenBudget.ts";
|
|
|
|
test("rejects a prompt that cannot leave one output token", () => {
|
|
const result = enforceOutputTokenBudget({ max_tokens: 8192 }, 527_058, 128_000);
|
|
|
|
assert.equal(result.ok, false);
|
|
if (result.ok) assert.fail("expected the rejected output-budget branch");
|
|
assert.equal(result.estimatedInputTokens, 527_058);
|
|
assert.equal(result.contextLimit, 128_000);
|
|
assert.deepEqual(result, {
|
|
ok: false,
|
|
estimatedInputTokens: 527_058,
|
|
contextLimit: 128_000,
|
|
});
|
|
});
|
|
|
|
test("caps a positive output budget to the target's remaining context", () => {
|
|
const input = { messages: [], max_tokens: 12_000 };
|
|
const result = enforceOutputTokenBudget(input, 127_000, 128_000);
|
|
|
|
assert.equal(result.ok, true);
|
|
if (!result.ok) return;
|
|
assert.equal(result.body.max_tokens, 1_000);
|
|
assert.equal(input.max_tokens, 12_000, "must not mutate the shared combo request body");
|
|
});
|
|
|
|
test("removes non-positive numeric output limits before upstream dispatch", () => {
|
|
const result = enforceOutputTokenBudget(
|
|
{ max_tokens: -398_464, max_completion_tokens: 0 },
|
|
1_000,
|
|
128_000
|
|
);
|
|
|
|
assert.equal(result.ok, true);
|
|
if (!result.ok) return;
|
|
assert.equal("max_tokens" in result.body, false);
|
|
assert.equal("max_completion_tokens" in result.body, false);
|
|
});
|
|
|
|
test("caps max_output_tokens to the target's remaining context", () => {
|
|
const result = enforceOutputTokenBudget({ max_output_tokens: 12_000 }, 127_000, 128_000);
|
|
|
|
assert.equal(result.ok, true);
|
|
if (!result.ok) return;
|
|
assert.equal(result.body.max_output_tokens, 1_000);
|
|
});
|
|
|
|
test("accepts a missing request body when output budget remains", () => {
|
|
const result = enforceOutputTokenBudget(null, 1_000, 128_000);
|
|
|
|
assert.deepEqual(result, {
|
|
ok: true,
|
|
body: {},
|
|
availableOutputTokens: 127_000,
|
|
adjustedFields: [],
|
|
});
|
|
});
|
|
|
|
test("rejects when a Claude target's default output budget does not fit", () => {
|
|
const result = enforceOutputTokenBudget({}, 70_000, 128_000, 64_000);
|
|
|
|
assert.deepEqual(result, {
|
|
ok: false,
|
|
estimatedInputTokens: 70_000,
|
|
contextLimit: 128_000,
|
|
});
|
|
});
|