pjandro/OmniRoute
1
0
Fork 0
mirror of https://github.com/diegosouzapw/OmniRoute.git synced 2026-08-10 17:22:17 +03:00
Code Issues Packages Projects Releases Wiki Activity

fix(chat): reduce file size

Browse Source 
Co-authored-by: diegosouzapw <8016841+diegosouzapw@users.noreply.github.com>
This commit is contained in:
jackjinke
2026-08-09 08:34:15 -03:00
parent 12f29bf310
commit df3b33bf1a
3 changed files with 0 additions and 103 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
open-sse/handlers/chatCore.ts
View File

@@ -393,7 +393,6 @@ import { isSmallEnoughForSemanticCache } from "../utils/estimateSize.ts";
// extractSystemRoleMessages extracted to chatCore/claudeSystemRole.ts (#3501); re-exported above so
// existing importers (e.g. tests/unit/system-role-extraction.test.ts) keep resolving it from here.
export async function handleChatCore({
body,
modelInfo,
@@ -5033,7 +5032,6 @@ export async function handleChatCore({
}),
};
}
export function isTokenExpiringSoon(expiresAt, bufferMs = 5 * 60 * 1000) {
if (!expiresAt) return false;
const expiresAtMs = new Date(expiresAt).getTime();

27
src/app/(dashboard)/dashboard/providers/[id]/components/modals/EditConnectionModal.tsx
View File

@@ -1,5 +1,4 @@
"use client";
import { useState, useEffect, useMemo } from "react";
import { useTranslations } from "next-intl";
import { Button, Badge, Input, Modal, Toggle, Select } from "@/shared/components";
@@ -58,7 +57,6 @@ import AgentrouterConsoleFields from "./AgentrouterConsoleFields";
import QuotaScrapingFields, { EMPTY_QUOTA_SCRAPING_FIELDS } from "./QuotaScrapingFields";
import GlmTeamQuotaFields, { EMPTY_GLM_TEAM_QUOTA_FIELDS } from "./GlmTeamQuotaFields";
import ProviderRegionField, { getProviderRegionConfig } from "./AlibabaProviderRegionField";
export interface EditConnectionModalConnection {
id?: string;
name?: string;
@@ -73,7 +71,6 @@ export interface EditConnectionModalConnection {
healthCheckInterval?: number;
projectId?: string | null;
}
export interface EditConnectionModalProps {
isOpen: boolean;
connection: EditConnectionModalConnection | null;
@@ -84,9 +81,7 @@ export interface EditConnectionModalProps {
onResyncModels?: (connectionId: string) => void | Promise<void>;
onClose: () => void;
}
const stringField = (value: unknown) => (typeof value === "string" ? value : "");
export default function EditConnectionModal({
isOpen,
connection,
@@ -170,7 +165,6 @@ export default function EditConnectionModal({
>({});
const [showAdvanced, setShowAdvanced] = useState(false);
const showEmail = useEmailPrivacyStore((state) => state.emailsVisible);
// #6147 — built-in providers can opt in to an advanced base-URL override.
// OAuth connections are excluded: their save path does not persist
// providerSpecificData.baseUrl.
@@ -247,7 +241,6 @@ export default function EditConnectionModal({
})),
[t]
);
useEffect(() => {
if (isOpen && connection) {
const effectiveProvider = connection.provider || providerId;
@@ -388,7 +381,6 @@ export default function EditConnectionModal({
defaultRegion,
setOpenRouterPreset,
]);
const handleTest = async () => {
if (!provider) return;
setTesting(true);
@@ -417,7 +409,6 @@ export default function EditConnectionModal({
setTesting(false);
}
};
const handleValidate = async () => {
if (
!provider ||
@@ -450,7 +441,6 @@ export default function EditConnectionModal({
setValidating(false);
}
};
const handleAddParsedExtraKeys = (raw: string) => {
const { added, duplicates } = parseExtraApiKeys(raw, extraApiKeys);
if (added.length > 0) {
@@ -461,7 +451,6 @@ export default function EditConnectionModal({
notify.warning(t("bulkPasteDuplicatesIgnored", { count: duplicates }));
}
};
const handleSubmit = async () => {
setSaving(true);
setSaveError(null);
@@ -477,14 +466,12 @@ export default function EditConnectionModal({
}
parsedMaxConcurrent = numericMaxConcurrent;
}
const updates: any = {
name: formData.name,
priority: formData.priority,
maxConcurrent: parsedMaxConcurrent,
healthCheckInterval: formData.healthCheckInterval,
};
const overrides: Record<string, number> = {};
if (formData.rpm.trim()) overrides.rpm = Number(formData.rpm);
if (formData.tpm.trim()) overrides.tpm = Number(formData.tpm);
@@ -493,16 +480,13 @@ export default function EditConnectionModal({
if (formData.rateLimitMaxConcurrent.trim())
overrides.maxConcurrent = Number(formData.rateLimitMaxConcurrent);
updates.rateLimitOverrides = Object.keys(overrides).length > 0 ? overrides : null;
if (isAntigravityFamily) {
updates.projectId = trimmedCloudCodeProjectId || null;
}
if (isGooglePse && !formData.cx.trim()) {
setSaveError(t("searchEngineIdRequired"));
return;
}
let validatedBaseUrl = null;
if (usesBaseUrl) {
// #6147 — an opt-in override left blank clears it (no default to fall
@@ -518,7 +502,6 @@ export default function EditConnectionModal({
validatedBaseUrl = checked.value;
}
}
if (!isOAuth && formData.apiKey) {
updates.apiKey = formData.apiKey;
let isValid = validationResult === "success";
@@ -648,15 +631,12 @@ export default function EditConnectionModal({
setSaving(false);
}
};
if (!connection) return null;
const isOAuth = connection.authType === "oauth";
const testErrorMeta =
!testResult?.valid && testResult?.diagnosis?.type
? ERROR_TYPE_LABELS[testResult.diagnosis.type] || null
: null;
const preserveEncryptedReasoningToggle = isResponsesConnection ? (
<Toggle
checked={formData.preserveEncryptedReasoning}
@@ -669,7 +649,6 @@ export default function EditConnectionModal({
)}
/>
) : null;
return (
<Modal isOpen={isOpen} title={t("editConnection")} onClose={onClose}>
<div className="flex flex-col gap-4">
@@ -1053,7 +1032,6 @@ export default function EditConnectionModal({
/>
</>
)}
{/* #6147 — opt-in "Advanced → override base URL" for eligible built-ins */}
{!usesBaseUrl && isBaseUrlOverrideEligible && (
<button
@@ -1064,7 +1042,6 @@ export default function EditConnectionModal({
{providerText(t, "overrideBaseUrlAdvanced", "Advanced: override base URL")}
</button>
)}
{usesBaseUrl && (
<Input
label={t("baseUrlLabel")}
@@ -1083,7 +1060,6 @@ export default function EditConnectionModal({
}
/>
)}
{showProtocolSelector && (
<Select
label={providerText(t, "apiProtocolLabel", "API protocol")}
@@ -1103,13 +1079,11 @@ export default function EditConnectionModal({
)}
/>
)}
<ProviderRegionField
provider={provider}
value={formData.region}
onChange={(region) => setFormData({ ...formData, region })}
/>
{isCloudflare && (
<Input
label={t("accountIdLabel")}
@@ -1119,7 +1093,6 @@ export default function EditConnectionModal({
hint={t("accountIdHint")}
/>
)}
{isGlm && (
<div className="flex flex-col gap-3">
<div>

74
tests/unit/chatcore-translation-paths.test.ts
View File

@@ -4,10 +4,8 @@ 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-chatcore-translation-"));
process.env.DATA_DIR = TEST_DATA_DIR;
const core = await import("../../src/lib/db/core.ts");
const providersDb = await import("../../src/lib/db/providers.ts");
const settingsDb = await import("../../src/lib/db/settings.ts");
@@ -49,14 +47,12 @@ const { resetPayloadRulesConfigForTests, setPayloadRulesConfig } =
await import("../../open-sse/services/payloadRules.ts");
const { FORMATS } = await import("../../open-sse/translator/formats.ts");
const { register, getRequestTranslator } = await import("../../open-sse/translator/registry.ts");
const originalFetch = globalThis.fetch;
const originalResponsesToOpenAI = getRequestTranslator(FORMATS.OPENAI_RESPONSES, FORMATS.OPENAI);
const originalSetTimeout = globalThis.setTimeout;
const originalBackgroundConfig = getBackgroundDegradationConfig();
const originalCallLogPipelineCaptureStreamChunks =
process.env.CALL_LOG_PIPELINE_CAPTURE_STREAM_CHUNKS;
function noopLog() {
return {
debug() {},
@@ -65,7 +61,6 @@ function noopLog() {
error() {},
};
}
function restorePipelineCaptureEnv() {
if (originalCallLogPipelineCaptureStreamChunks === undefined) {
delete process.env.CALL_LOG_PIPELINE_CAPTURE_STREAM_CHUNKS;
@@ -74,7 +69,6 @@ function restorePipelineCaptureEnv() {
originalCallLogPipelineCaptureStreamChunks;
}
}
function toPlainHeaders(headers) {
if (!headers) return {};
if (headers instanceof Headers) return Object.fromEntries(headers.entries());
@@ -82,7 +76,6 @@ function toPlainHeaders(headers) {
Object.entries(headers).map(([key, value]) => [key, value == null ? "" : String(value)])
);
}
function buildOpenAIResponse(stream, text = "ok") {
if (stream) {
return new Response(
@@ -97,7 +90,6 @@ function buildOpenAIResponse(stream, text = "ok") {
}
);
}
return new Response(
JSON.stringify({
id: "chatcmpl-json",
@@ -122,7 +114,6 @@ function buildOpenAIResponse(stream, text = "ok") {
}
);
}
function buildClaudeResponse(stream, text = "ok") {
if (stream) {
return new Response(
@@ -170,7 +161,6 @@ function buildClaudeResponse(stream, text = "ok") {
}
);
}
return new Response(
JSON.stringify({
id: "msg_json",
@@ -389,7 +379,6 @@ test.after(async () => {
await resetStorage();
fs.rmSync(TEST_DATA_DIR, { recursive: true, force: true });
});
test("chatCore times out upstream execution before provider response headers", async () => {
// This test asserts pendingDetail.providerRequest — only attached when the
// call-log pipeline capture is enabled. Declare the dependency explicitly
@@ -456,7 +445,6 @@ test("chatCore times out upstream execution before provider response headers", a
globalThis.fetch = originalFetch;
}
});
test("chatCore can disable pipeline stream chunk capture through environment", async () => {
process.env.CALL_LOG_PIPELINE_CAPTURE_STREAM_CHUNKS = "false";
await settingsDb.updateSettings({ call_log_pipeline_enabled: true });
@@ -479,7 +467,6 @@ test("chatCore can disable pipeline stream chunk capture through environment", a
assert.ok(detail.pipelinePayloads, "expected pipeline payloads when capture is enabled");
assert.equal((detail.pipelinePayloads as any).streamChunks, undefined);
});
test("chatCore keeps Responses-native Codex payloads in native passthrough mode", async () => {
const { call, result } = await invokeChatCore({
provider: "codex",
@@ -507,7 +494,6 @@ test("chatCore keeps Responses-native Codex payloads in native passthrough mode"
assert.deepEqual(call.body.metadata, { source: "codex-client" });
assert.equal("messages" in call.body, false);
});
test("chatCore honors providerSpecificData.apiType for legacy openai-compatible providers", async () => {
const { call, result } = await invokeChatCore({
provider: "openai-compatible-sp-openai",
@@ -537,7 +523,6 @@ test("chatCore honors providerSpecificData.apiType for legacy openai-compatible
assert.equal("messages" in call.body, false);
assert.equal(payload.choices[0].message.content, "ok");
});
test("chatCore applies Responses input policy to openai-compatible targets", async () => {
const reasoningItems = [
{ id: "rs_valid", type: "reasoning", encrypted_content: "encrypted-blob" },
@@ -578,7 +563,6 @@ test("chatCore applies Responses input policy to openai-compatible targets", asy
assert.equal(input.find((item) => item.type === "function_call")?.id, undefined);
}
});
test("chatCore preserves opted-in encrypted reasoning for Codex", async () => {
const { call, result } = await invokeChatCore({
provider: "codex",
@@ -611,7 +595,6 @@ test("chatCore preserves opted-in encrypted reasoning for Codex", async () => {
false
);
});
test("chatCore helper exports detect responses passthrough paths and token expiry windows", () => {
assert.equal(
shouldUseNativeCodexPassthrough({
@@ -639,7 +622,6 @@ test("chatCore helper exports detect responses passthrough paths and token expir
);
assert.equal(isTokenExpiringSoon(null), false);
});
test("chatCore helper detects Claude Code semantic passthrough only for direct Claude-Code routes", () => {
assert.equal(
isClaudeCodeSemanticPassthroughRequest({
@@ -679,7 +661,6 @@ test("chatCore helper detects Claude Code semantic passthrough only for direct C
false
);
});
test("chatCore applies payload rules after translating Responses input into Chat payloads", async () => {
setPayloadRulesConfig({
default: [
@@ -725,7 +706,6 @@ test("chatCore applies payload rules after translating Responses input into Chat
assert.equal(call.body.messages[0].metadata.routeTag, "feature-110");
assert.equal(call.body.messages[0].role, "user");
});
test("chatCore builds Claude Code-compatible upstream requests for CC providers", async () => {
const { call, result } = await invokeChatCore({
provider: "anthropic-compatible-cc-test",
@@ -835,7 +815,6 @@ test("chatCore normalizes native Claude Code messages for native Claude OAuth pa
// user msg[2] (was clientMessages[3]): tool_result preserved (preserveToolResultBlocks:true)
assert.equal(call.body.messages[2].content[0].type, "tool_result");
});
test("chatCore preserves Opus 5 mid-conversation system cache breakpoints", async () => {
await settingsDb.updateSettings({ alwaysPreserveClientCache: "auto" });
invalidateCacheControlSettingsCache();
@@ -889,7 +868,6 @@ test("chatCore preserves Opus 5 mid-conversation system cache breakpoints", asyn
);
assert.equal(call.body.messages[3].content[0].cache_control, undefined);
});
test("chatCore keeps Claude normalization for non-Claude-Code Claude passthrough", async () => {
const { call, result } = await invokeChatCore({
provider: "claude",
@@ -1017,7 +995,6 @@ test("chatCore normalizes native Claude Code messages before CC-compatible relay
// user msg[2] (was clientMessages[3]): tool_result preserved (preserveToolResultBlocks:true)
assert.equal(call.body.messages[2].content[0].type, "tool_result");
});
test("chatCore preserves cache_control automatically for Claude Code single-model requests", async () => {
await settingsDb.updateSettings({ alwaysPreserveClientCache: "auto" });
invalidateCacheControlSettingsCache();
@@ -1064,7 +1041,6 @@ test("chatCore preserves cache_control automatically for Claude Code single-mode
// base.ts executor explicitly strips cache_control from tools for Claude Code clients
assert.equal(call.body.tools[0].cache_control, undefined);
});
test("chatCore supplements a missing message cache breakpoint for native Claude Code requests", async () => {
await settingsDb.updateSettings({ alwaysPreserveClientCache: "auto" });
invalidateCacheControlSettingsCache();
@@ -1110,7 +1086,6 @@ test("chatCore supplements a missing message cache breakpoint for native Claude
assert.deepEqual(call.body.messages[2].content[0].cache_control, { type: "ephemeral" });
assert.equal(call.body.tools[0].cache_control, undefined);
});
test("chatCore auto cache policy becomes false for nondeterministic combos", async () => {
await settingsDb.updateSettings({ alwaysPreserveClientCache: "auto" });
invalidateCacheControlSettingsCache();
@@ -1144,7 +1119,6 @@ test("chatCore auto cache policy becomes false for nondeterministic combos", asy
true
);
});
test("chatCore always-preserve mode keeps cache_control even without Claude Code user-agent", async () => {
await settingsDb.updateSettings({ alwaysPreserveClientCache: "always" });
invalidateCacheControlSettingsCache();
@@ -1166,7 +1140,6 @@ test("chatCore always-preserve mode keeps cache_control even without Claude Code
assert.equal(hasCacheControl(call.body), true);
assert.deepEqual(call.body.system[0].cache_control, { type: "ephemeral", ttl: "5m" });
});
test("chatCore disables raw Claude passthrough when cache preservation is off and normalizes through OpenAI", async () => {
await settingsDb.updateSettings({ alwaysPreserveClientCache: "never" });
invalidateCacheControlSettingsCache();
@@ -1202,7 +1175,6 @@ test("chatCore disables raw Claude passthrough when cache preservation is off an
// Tools disable flag is applied
assert.equal("_disableToolPrefix" in call.body, false);
});
test("chatCore default translation converts Claude requests to OpenAI and strips cache markers for non-Claude providers", async () => {
const { call } = await invokeChatCore({
provider: "openai",
@@ -1228,7 +1200,6 @@ test("chatCore default translation converts Claude requests to OpenAI and strips
assert.equal(call.body.messages[0].role, "system");
assert.equal(JSON.stringify(call.body).includes("cache_control"), false);
});
test("chatCore sets Claude tool prefix disabling, strips empty Anthropic text blocks, and cleans helper flags", async () => {
const { call } = await invokeChatCore({
provider: "claude",
@@ -1271,7 +1242,6 @@ test("chatCore sets Claude tool prefix disabling, strips empty Anthropic text bl
["hello"]
);
});
test("chatCore restores prefixed Claude passthrough tool names in upstream responses", async () => {
const { result } = await invokeChatCore({
provider: "claude",
@@ -1323,7 +1293,6 @@ test("chatCore restores prefixed Claude passthrough tool names in upstream respo
assert.equal(result.success, true);
assert.equal(payload.content[0].name, "Bash");
});
test("chatCore strips unsupported reasoning params and caps provider token fields", async () => {
const { call } = await invokeChatCore({
provider: "openai",
@@ -1345,7 +1314,6 @@ test("chatCore strips unsupported reasoning params and caps provider token field
assert.equal(call.body.max_tokens, undefined);
assert.equal(call.body.max_completion_tokens, 16384);
});
test("chatCore preserves reasoning_effort for assistant-prefill OpenAI-compatible requests", async () => {
const { call, result } = await invokeChatCore({
provider: "openai-compatible-aio",
@@ -1367,7 +1335,6 @@ test("chatCore preserves reasoning_effort for assistant-prefill OpenAI-compatibl
assert.equal(call.body.model, "glm-5.1");
assert.equal(call.body.reasoning_effort, "xhigh");
});
test("chatCore logs chat completions endpoint as OpenAI protocol", async () => {
const { call, result } = await invokeChatCore({
provider: "openrouter",
@@ -1394,7 +1361,6 @@ test("chatCore logs chat completions endpoint as OpenAI protocol", async () => {
assert.equal(logEntry.path, "/v1/chat/completions");
assert.equal(logEntry.sourceFormat, FORMATS.OPENAI);
});
test("chatCore surfaces translation errors with explicit status codes", async () => {
register(
FORMATS.OPENAI_RESPONSES,
@@ -1421,7 +1387,6 @@ test("chatCore surfaces translation errors with explicit status codes", async ()
assert.equal(result.status, 409);
assert.equal(result.error, "responses translator rejected the payload");
});
test("chatCore surfaces typed translation errors with the declared error type", async () => {
register(
FORMATS.OPENAI_RESPONSES,
@@ -1452,7 +1417,6 @@ test("chatCore surfaces typed translation errors with the declared error type",
assert.equal(payload.error.type, "unsupported_feature");
assert.equal(payload.error.code, "unsupported_feature");
});
test("chatCore returns 500 when translation throws a generic error", async () => {
register(
FORMATS.OPENAI_RESPONSES,
@@ -1477,7 +1441,6 @@ test("chatCore returns 500 when translation throws a generic error", async () =>
assert.equal(result.status, 500);
assert.equal(result.error, "unexpected translator crash");
});
test("chatCore refreshes GitHub credentials after 401 and retries with the refreshed Copilot token", async () => {
let refreshedCredentials = null;
const { calls, result } = await invokeChatCore({
@@ -1543,7 +1506,6 @@ test("chatCore refreshes GitHub credentials after 401 and retries with the refre
assert.equal(refreshedCredentials?.providerSpecificData?.copilotToken, "copilot-refreshed-token");
assert.equal(payload.choices[0].message.content, "retry succeeded after refresh");
});
test("chatCore uses the native executor when no upstream proxy mode is enabled", async () => {
const { call } = await invokeChatCore({
provider: "openai",
@@ -1557,7 +1519,6 @@ test("chatCore uses the native executor when no upstream proxy mode is enabled",
assert.match(call.url, /^https:\/\/api\.openai\.com\/v1\/chat\/completions$/);
});
test("chatCore routes providers through CLIProxyAPI in passthrough mode", async () => {
await upstreamProxyDb.upsertUpstreamProxyConfig({
providerId: "qoder",
@@ -1579,7 +1540,6 @@ test("chatCore routes providers through CLIProxyAPI in passthrough mode", async
assert.match(call.url, /^http:\/\/127\.0\.0\.1:8317\/v1\/chat\/completions$/);
assert.equal(call.headers.Authorization ?? call.headers.authorization, "Bearer qoder-token");
});
test("chatCore fallback proxy mode retries through CLIProxyAPI after retryable native failures", async () => {
await upstreamProxyDb.upsertUpstreamProxyConfig({
providerId: "github",
@@ -1620,7 +1580,6 @@ test("chatCore fallback proxy mode retries through CLIProxyAPI after retryable n
assert.match(calls[0].url, /^https:\/\/api\.githubcopilot\.com\/chat\/completions$/);
assert.match(calls[1].url, /^http:\/\/127\.0\.0\.1:8317\/v1\/chat\/completions$/);
});
test("chatCore fallback proxy mode surfaces CLIProxyAPI errors after a retryable native status", async () => {
await upstreamProxyDb.upsertUpstreamProxyConfig({
providerId: "github",
@@ -1661,7 +1620,6 @@ test("chatCore fallback proxy mode surfaces CLIProxyAPI errors after a retryable
assert.equal(result.status, 502);
assert.equal(result.error, "[502]: cliproxy retry failed");
});
test("chatCore fallback proxy mode surfaces CLIProxyAPI errors after native executor throws", async () => {
await upstreamProxyDb.upsertUpstreamProxyConfig({
providerId: "github",
@@ -1699,7 +1657,6 @@ test("chatCore fallback proxy mode surfaces CLIProxyAPI errors after native exec
assert.equal(result.status, 502);
assert.equal(result.error, "[502]: cliproxy transport exploded");
});
test("chatCore serves a cached idempotent response without hitting the provider twice", async () => {
const sharedHeaders = { "idempotency-key": "unit-idempotent-key" };
@@ -1735,7 +1692,6 @@ test("chatCore serves a cached idempotent response without hitting the provider
const payload = (await second.result.response.json()) as any;
assert.equal(payload.choices[0].message.content, "ok");
});
test("chatCore returns a semantic cache HIT for repeated deterministic requests", async () => {
let upstreamHits = 0;
const sharedBody = {
@@ -1788,7 +1744,6 @@ test("chatCore returns a semantic cache HIT for repeated deterministic requests"
assert.equal(semanticLog.path, "/v1/chat/completions");
assert.equal(semanticLog.status, 200);
});
test("chatCore skips semantic cache when disabled in settings", async () => {
await settingsDb.updateSettings({ semanticCacheEnabled: false });
@@ -1831,7 +1786,6 @@ test("chatCore skips semantic cache when disabled in settings", async () => {
const payload = (await second.result.response.json()) as any;
assert.equal(payload.choices[0].message.content, "fresh-2");
});
test("chatCore attaches OmniRoute response metadata headers to non-stream responses", async () => {
const { result } = await invokeChatCore({
provider: "claude",
@@ -1853,7 +1807,6 @@ test("chatCore attaches OmniRoute response metadata headers to non-stream respon
assert.ok(Number(result.response.headers.get("X-OmniRoute-Latency-Ms")) >= 0);
assert.match(String(result.response.headers.get("X-OmniRoute-Response-Cost")), /^\d+\.\d{10}$/);
});
test("chatCore does not expose provider request credentials in non-stream response headers", async () => {
const { result } = await invokeChatCore({
provider: "openai",
@@ -1872,7 +1825,6 @@ test("chatCore does not expose provider request credentials in non-stream respon
assert.equal(result.response.headers.get("Content-Type"), "application/json");
assert.equal(result.response.headers.get("X-OmniRoute-Cache"), "MISS");
});
test("chatCore normalizes tool finish reasons and estimates usage when upstream omits it", async () => {
const { result } = await invokeChatCore({
provider: "openai",
@@ -1924,7 +1876,6 @@ test("chatCore normalizes tool finish reasons and estimates usage when upstream
assert.ok(payload.usage.total_tokens > 0);
assert.ok(payload.usage.prompt_tokens > 0);
});
test("chatCore bypasses Claude CLI warmup probes before touching the provider", async () => {
const { calls, result } = await invokeChatCore({
model: "gpt-5",
@@ -1941,7 +1892,6 @@ test("chatCore bypasses Claude CLI warmup probes before touching the provider",
assert.equal(calls.length, 0);
assert.match(payload.choices[0].message.content, /CLI Command Execution/);
});
test("chatCore redirects background utility tasks to a cheaper mapped model", async () => {
setBackgroundDegradationConfig({
enabled: true,
@@ -1968,7 +1918,6 @@ test("chatCore redirects background utility tasks to a cheaper mapped model", as
assert.equal(result.success, true);
assert.equal(call.body.model, "gpt-5-mini");
});
test("chatCore preserves Codex dual-window scope cooldowns on 429 responses", async () => {
const connection = await providersDb.createProviderConnection({
provider: "codex",
@@ -2022,7 +1971,6 @@ test("chatCore preserves Codex dual-window scope cooldowns on 429 responses", as
);
assert.equal((updated as any).providerSpecificData.codexExhaustedWindow, "5h");
});
test("chatCore 429 lets account fallback apply the configured resilience cooldown", async () => {
await settingsDb.updateSettings({
resilienceSettings: {
@@ -2083,7 +2031,6 @@ test("chatCore 429 lets account fallback apply the configured resilience cooldow
assert.equal((afterFallback as any).testStatus, "unavailable");
assert.ok(cooldownRemaining > 0 && cooldownRemaining <= 2_000);
});
test("chatCore falls back to the next family model when the requested model is unavailable", async () => {
const { calls, result } = await invokeChatCore({
provider: "openai",
@@ -2110,7 +2057,6 @@ test("chatCore falls back to the next family model when the requested model is u
assert.equal(calls[1].body.model, "gpt-5.1-mini");
assert.equal(payload.choices[0].message.content, "family fallback ok");
});
test("chatCore falls back to a larger-context sibling when the request overflows context", async () => {
saveModelsDevCapabilities({
unknown: {
@@ -2145,7 +2091,6 @@ test("chatCore falls back to a larger-context sibling when the request overflows
assert.equal(calls[1].body.model, "gpt-4o");
assert.equal(payload.choices[0].message.content, "larger context fallback");
});
test("chatCore parses upstream SSE payloads for non-streaming requests", async () => {
const { result } = await invokeChatCore({
provider: "openai",
@@ -2164,7 +2109,6 @@ test("chatCore parses upstream SSE payloads for non-streaming requests", async (
assert.equal(result.success, true);
assert.equal(payload.choices[0].message.content, "sse json");
});
test("chatCore rejects malformed non-streaming SSE payloads", async () => {
const { result } = await invokeChatCore({
provider: "openai",
@@ -2186,7 +2130,6 @@ test("chatCore rejects malformed non-streaming SSE payloads", async () => {
assert.equal(result.status, 502);
assert.match(result.error, /Invalid SSE response/);
});
test("chatCore rejects malformed non-streaming JSON payloads", async () => {
const { result } = await invokeChatCore({
provider: "openai",
@@ -2208,7 +2151,6 @@ test("chatCore rejects malformed non-streaming JSON payloads", async () => {
assert.equal(result.status, 502);
assert.equal(result.error, "Invalid JSON response from provider");
});
test("chatCore falls back after an empty-content success response", async () => {
const { calls, result } = await invokeChatCore({
provider: "openai",
@@ -2249,7 +2191,6 @@ test("chatCore falls back after an empty-content success response", async () =>
assert.equal(calls[1].body.model, "gpt-5.1-mini");
assert.equal(payload.choices[0].message.content, "empty-content fallback ok");
});
test("chatCore returns a gateway error when the empty-content fallback responds with invalid JSON", async () => {
const { result, calls } = await invokeChatCore({
provider: "openai",
@@ -2294,7 +2235,6 @@ test("chatCore returns a gateway error when the empty-content fallback responds
assert.equal(calls.length, 2);
assert.equal(calls[1].body.model, "gpt-5.1-mini");
});
test("chatCore records Claude prompt cache and cache usage metadata in call logs", async () => {
await settingsDb.updateSettings({ alwaysPreserveClientCache: "always" });
invalidateCacheControlSettingsCache();
@@ -2372,7 +2312,6 @@ test("chatCore records Claude prompt cache and cache usage metadata in call logs
cacheCreationTokens: 2,
});
});
test("chatCore propagates budget errors without an executor-level emergency hop", async () => {
// The emergency budget fallback is orchestrated by the routing layer
// (src/sse/handlers/chat.ts), which resolves credentials FOR the emergency
@@ -2412,7 +2351,6 @@ test("chatCore propagates budget errors without an executor-level emergency hop"
"emergency fallback model must not be called at executor level"
);
});
test("chatCore injects progress events into streaming responses when requested", async () => {
const { result } = await invokeChatCore({
provider: "openai",
@@ -2434,7 +2372,6 @@ test("chatCore injects progress events into streaming responses when requested",
assert.equal(result.response.headers.get("X-OmniRoute-Progress"), "enabled");
assert.match(streamText, /event: progress/);
});
test("chatCore emits final SSE metadata comments before [DONE] on streaming responses", async () => {
const { result } = await invokeChatCore({
provider: "openai",
@@ -2462,7 +2399,6 @@ test("chatCore emits final SSE metadata comments before [DONE] on streaming resp
streamText.indexOf(": x-omniroute-response-cost=") < streamText.indexOf("data: [DONE]")
);
});
test("buildStreamingResponseHeaders drops upstream compression and framing headers", () => {
const headers = new Headers(
buildStreamingResponseHeaders(
@@ -2491,7 +2427,6 @@ test("buildStreamingResponseHeaders drops upstream compression and framing heade
assert.equal(headers.get("X-Upstream-Trace"), "trace-1");
assert.equal(headers.get("X-OmniRoute-Cache"), "MISS");
});
test("chatCore strips upstream compression and length headers from streaming responses", async () => {
const upstreamPayload = `data: ${JSON.stringify({
id: "chatcmpl-stream-headers",
@@ -2526,7 +2461,6 @@ test("chatCore strips upstream compression and length headers from streaming res
assert.equal(result.response.headers.get("X-OmniRoute-Cache"), "MISS");
await result.response.text();
});
test("chatCore maps upstream aborts to request-aborted errors", async () => {
const { result } = await invokeChatCore({
provider: "openai",
@@ -2547,7 +2481,6 @@ test("chatCore maps upstream aborts to request-aborted errors", async () => {
assert.equal(result.status, 499);
assert.equal(result.error, "Request aborted");
});
test("chatCore maps raw string abort reasons to 499, not 502 (#7907)", async () => {
// abort(reason) rejects the upstream fetch with the raw reason — often a
// bare string with no `name`/`status`. It must map to 499 like a named
@@ -2610,7 +2543,6 @@ test("chatCore does not log a synthetic clientResponse body for a client abort",
"an aborted request never delivered anything to the client — clientResponse must stay unset"
);
});
test("chatCore returns streaming responses without waiting for upstream completion", async () => {
const encoder = new TextEncoder();
let closeUpstream: (() => void) | null = null;
@@ -2679,7 +2611,6 @@ test("chatCore returns streaming responses without waiting for upstream completi
assert.equal(result.success, true);
assert.match(streamText, /streamed-without-buffering/);
});
test("chatCore releases account semaphore slots when upstream execution throws", async () => {
const connectionId = "sem-exception";
const semaphoreKey = buildAccountSemaphoreKey({
@@ -2712,7 +2643,6 @@ test("chatCore releases account semaphore slots when upstream execution throws",
assert.equal(result.status, 502);
assert.equal(getAccountSemaphoreStats()[semaphoreKey], undefined);
});
test("chatCore locks per-model quota failures without dropping quota helper references", async () => {
const model = "gemini-1.5-pro";
const connection = await providersDb.createProviderConnection({
@@ -2758,7 +2688,6 @@ test("chatCore locks per-model quota failures without dropping quota helper refe
});
// ── Streaming semantic cache tests ──────────────────────────────────────────
test("chatCore caches streaming response and serves cache HIT on repeat", async () => {
let upstreamHits = 0;
const sharedBody = {
@@ -2813,7 +2742,6 @@ test("chatCore caches streaming response and serves cache HIT on repeat", async
assert.match(sse, /^data:/m, "cache HIT should be SSE-framed");
assert.match(sse, /streamed-once/, "SSE cache HIT should carry the cached content");
});
test("chatCore does not cache streaming response when temperature > 0", async () => {
let upstreamHits = 0;
const sharedBody = {
@@ -2854,7 +2782,6 @@ test("chatCore does not cache streaming response when temperature > 0", async ()
assert.equal(upstreamHits, 2, "both requests should hit upstream");
assert.equal(second.calls.length, 1, "second request should reach upstream");
});
test("chatCore skips streaming cache when X-OmniRoute-No-Cache header is set", async () => {
let upstreamHits = 0;
const sharedBody = {
@@ -2901,7 +2828,6 @@ test("chatCore skips streaming cache when X-OmniRoute-No-Cache header is set", a
await second.result.response.text();
assert.equal(upstreamHits, 2, "both requests should hit upstream with no-cache");
});
test("chatCore returns cache HIT as SSE when the client requests streaming", async () => {
const sharedBody = {
model: "gpt-4o-mini",