mirror of
https://github.com/diegosouzapw/OmniRoute.git
synced 2026-09-16 03:42:21 +03:00
3d4f3e4960
* test(infra): retry recursive temp-dir removal instead of failing a shard on ENOTEMPTY (#11966) Two shards on release/v3.8.51 went red in one day with the same signature — "ENOTEMPTY, Directory not empty: /tmp/omniroute-<test>-XXXXXX" — from combo-same-provider-cascade (Unit Tests fast-path 4/4, on a PR that touches only .github/) and auth-policy-embeddings-webfetch-7785 (the 20k-test TIA step). Both pass alone and on re-run: the cleanup races something still writing into the directory (SQLite WAL/-shm checkpoint, a worker, the backup) and under a loaded hosted runner the window opens. 1154 test files do their own cleanup with fs.rmSync(dir, { recursive: true, force: true }); 57 already asked for retries. One-shot codemod (scripts/ad-hoc/codemod-rm-maxretries.mjs, kept for the record): every rm / rmSync / rmdirSync option object with `recursive: true` and no `maxRetries` gains `maxRetries: 5, retryDelay: 100` — Node itself then retries ENOTEMPTY/EBUSY/EPERM for up to ~0.5 s before giving up. 2243 call sites in 1292 files under tests/, the shared tests/_setup/isolateDataDir.ts exit hook included. Only the option object changes: no call site, assertion or import is touched. Validation: prettier and ESLint (with the frozen suppressions) clean on all 1292 files; a random 20-file sample runs green (quota-redis-store hangs identically on the untouched tree — it needs a Redis on localhost, an environment matter). The four unit shards on this PR are the full run. * fix(quality): let check-forgotten-sibling-tests read a 1,000-file diff The gate shells out to `git diff` through execFileSync with Node's default 1 MB maxBuffer; the 1,292-file codemod in this PR is the first diff large enough to overflow it, and the gate died with `spawnSync git ENOBUFS` before comparing anything. 64 MB is far above any real PR and costs nothing when unused.
87 lines
3.4 KiB
TypeScript
87 lines
3.4 KiB
TypeScript
import test from "node:test";
|
|
import assert from "node:assert/strict";
|
|
import fs from "node:fs";
|
|
import os from "node:os";
|
|
import path from "node:path";
|
|
|
|
const TEST_DATA_DIR = fs.mkdtempSync(path.join(os.tmpdir(), "omniroute-reasoning-guards-"));
|
|
process.env.DATA_DIR = TEST_DATA_DIR;
|
|
process.env.API_KEY_SECRET = process.env.API_KEY_SECRET || "test-reasoning-guards-secret";
|
|
|
|
const core = await import("../../src/lib/db/core.ts");
|
|
const rulesDb = await import("../../src/lib/db/reasoningRoutingRules.ts");
|
|
const handler = await import("../../src/sse/handlers/reasoningRouting.ts");
|
|
|
|
/**
|
|
* `applyReasoningRouting` short-circuits on two separate conditions before
|
|
* applying a decision: no rule matched, and the resolver having failed
|
|
* (`resolveDecision` catches and returns `{ error }`).
|
|
*
|
|
* They used to be written as one compound guard —
|
|
* `if (decision && "error" in decision)` — which only narrows its *true*
|
|
* branch, so the `{ error }` arm survived in the union all the way down. The
|
|
* guards are now separate. These tests pin the observable behaviour of the
|
|
* paths that reorder touched; they pass identically before and after it.
|
|
*/
|
|
|
|
test.after(() => {
|
|
core.resetDbInstance();
|
|
fs.rmSync(TEST_DATA_DIR, { recursive: true, force: true, maxRetries: 5, retryDelay: 100 });
|
|
});
|
|
|
|
function req() {
|
|
return new Request("https://example.test/v1/chat/completions", { method: "POST" });
|
|
}
|
|
|
|
test("no matching rule passes the request through untouched", async () => {
|
|
rulesDb.invalidateReasoningRoutingRuleCache();
|
|
|
|
const body = { model: "gpt-4o", messages: [{ role: "user", content: "hi" }] };
|
|
const result = await handler.applyReasoningRouting({
|
|
request: req(),
|
|
body,
|
|
modelStr: "gpt-4o",
|
|
policy: { apiKey: null, apiKeyInfo: null },
|
|
apiKeyInfo: null,
|
|
});
|
|
|
|
assert.equal(result.response, null, "no rule must not short-circuit the request");
|
|
assert.equal(result.reasoningDecision, null, "and must not report a decision");
|
|
assert.equal(result.modelStr, "gpt-4o", "the model is forwarded unchanged");
|
|
assert.equal(result.body, body, "the body is forwarded by reference, unmodified");
|
|
});
|
|
|
|
test("the pass-through still reports the extracted reasoning intent", async () => {
|
|
rulesDb.invalidateReasoningRoutingRuleCache();
|
|
|
|
const result = await handler.applyReasoningRouting({
|
|
request: req(),
|
|
body: { model: "gpt-4o", reasoning_effort: "high", messages: [] },
|
|
modelStr: "gpt-4o",
|
|
policy: { apiKey: null, apiKeyInfo: null },
|
|
apiKeyInfo: null,
|
|
});
|
|
|
|
// Short-circuiting early must not skip intent extraction — downstream routing
|
|
// reads this even when no reasoning rule applied. The model comes back
|
|
// provider-qualified, which is what the rule matcher would have been given.
|
|
assert.equal(result.reasoningIntent.model, "openai/gpt-4o");
|
|
assert.equal(result.reasoningIntent.sourceEffort, "high");
|
|
assert.ok(Array.isArray(result.requestRoutingTags.tags));
|
|
});
|
|
|
|
test("an unknown model with no rules is also a clean pass-through", async () => {
|
|
rulesDb.invalidateReasoningRoutingRuleCache();
|
|
|
|
const result = await handler.applyReasoningRouting({
|
|
request: req(),
|
|
body: { model: "not-a-real-provider/not-a-real-model", messages: [] },
|
|
modelStr: "not-a-real-provider/not-a-real-model",
|
|
policy: { apiKey: null, apiKeyInfo: null },
|
|
apiKeyInfo: null,
|
|
});
|
|
|
|
assert.equal(result.response, null);
|
|
assert.equal(result.reasoningDecision, null);
|
|
});
|