import { createChatOptions } from '@tanstack/ai' import { createOpenaiChat } from '@tanstack/ai-openai' import Anthropic from '@anthropic-ai/sdk' import { createAnthropicChatWithClient } from '@tanstack/ai-anthropic' import { createGeminiChat } from '@tanstack/ai-gemini' import { createGeminiTextInteractions } from '@tanstack/ai-gemini/experimental' import { vertexText } from '@tanstack/ai-vertex' import { grokVertexText } from '@tanstack/ai-grok/vertex' import { mistralVertexText } from '@tanstack/ai-mistral/vertex' import { vertexE2eAuthClient, vertexE2eConfig } from '@/lib/vertex-e2e' import { createOllamaChat } from '@tanstack/ai-ollama' import { createGroqText } from '@tanstack/ai-groq' import { createGrokText } from '@tanstack/ai-grok' import { createBedrockText } from '@tanstack/ai-bedrock' import { openaiCompatibleText } from '@tanstack/ai-openai/compatible' import { createOpenRouterResponsesText, createOpenRouterText, } from '@tanstack/ai-openrouter' import { createVercelGatewayText } from '@tanstack/ai-vercel-gateway' import { createLovableText } from '@tanstack/ai-lovable' import { createMistralText } from '@tanstack/ai-mistral' import { createBytePlusText } from '@tanstack/ai-byteplus' import { createLLMGatewayText } from '@tanstack/ai-llmgateway' import { createCloudflareText } from '@tanstack/ai-cloudflare' import { createNeonText } from '@tanstack/ai-neon' import { createMuseText } from '@tanstack/ai-muse' import { HTTPClient } from '@openrouter/sdk' import type { AnyTextAdapter } from '@tanstack/ai' import type { BytePlusChatModel } from '@tanstack/ai-byteplus' import type { Feature, Provider } from '@/lib/types' export const LLMOCK_DEFAULT_BASE = process.env.LLMOCK_URL || 'http://127.0.0.1:4010' const DUMMY_KEY = 'sk-e2e-test-dummy-key' /** * Re-keys aimock's reasoning frames to the pre-July-2025 * `response.reasoning.delta` event name that Mantle-served Bedrock models * still emit, so openai-base's legacy-event mapping is exercised over a real * wire. aimock authors everything else about the stream. */ const legacyReasoningFetch: typeof fetch = async (input, init) => { const response = await fetch(input, init) if ( !response.body || !response.headers.get('content-type')?.includes('text/event-stream') ) { return response } const decoder = new TextDecoder() const encoder = new TextEncoder() const rewriteFrame = (frame: string) => frame .replace( 'event: response.reasoning_summary_text.delta', 'event: response.reasoning.delta', ) .replace( '"type":"response.reasoning_summary_text.delta"', '"type":"response.reasoning.delta"', ) let buffer = '' const rewritten = response.body.pipeThrough( new TransformStream({ transform(chunk, controller) { buffer += decoder.decode(chunk, { stream: true }) const frames = buffer.split(/\r?\n\r?\n/) buffer = frames.pop() ?? '' for (const frame of frames) { controller.enqueue(encoder.encode(rewriteFrame(frame) + '\n\n')) } }, flush(controller) { if (buffer) controller.enqueue(encoder.encode(rewriteFrame(buffer))) }, }), ) return new Response(rewritten, { status: response.status, statusText: response.statusText, headers: response.headers, }) } const defaultModels: Record = { openai: 'gpt-4o', anthropic: 'claude-sonnet-4-5', gemini: 'gemini-2.5-flash', vertex: 'gemini-2.5-flash', 'vertex-grok': 'grok-4.3', 'vertex-mistral': 'mistral-medium-3', ollama: 'mistral', groq: 'openai/gpt-oss-120b', grok: 'grok-build-0.1', bedrock: 'openai.gpt-oss-120b-1:0', 'bedrock-responses': 'openai.gpt-oss-120b-1:0', openrouter: 'openai/gpt-4o', 'openrouter-responses': 'openai/gpt-4o', 'vercel-gateway': 'openai/gpt-5.5', 'vercel-gateway-responses': 'openai/gpt-5.5', lovable: 'openai/gpt-5.5', 'lovable-responses': 'openai/gpt-5.5', 'openai-compatible': 'gpt-4o', // Arbitrary id on purpose: the provider models "any model behind an // OpenAI-compatible Responses endpoint frozen on the pre-July-2025 spec", // not a specific vendor model. 'openai-compatible-legacy': 'legacy-spec-model', mistral: 'mistral-large-latest', // Structured-output features override this in `features.ts`: // seed-2-0-lite-260428 rejects both `response_format: json_schema` and // `json_object` (live-probed), so it can't drive any structured feature. byteplus: 'seed-2-0-lite-260428', // ElevenLabs has no chat/text model — the support matrix already filters // it out of text features, but we still need an entry to satisfy the // Record constraint. elevenlabs: '', llmgateway: 'gpt-5.6-terra', cloudflare: '@cf/zai-org/glm-5.3-flash', neon: 'gemini-3-8-flash', 'neon-anthropic': 'claude-sonnet-4-5', 'neon-responses': 'gpt-5-3-codex', muse: 'muse-ember-1.0', } export function createTextAdapter( provider: Provider, modelOverride?: string, _aimockPort?: number, testId?: string, feature?: Feature, ): { adapter: AnyTextAdapter } { const model = modelOverride ?? defaultModels[provider] // OpenAI, Grok SDKs need /v1 in baseURL. Groq SDK appends /openai/v1/ internally. // Anthropic, Gemini, Ollama SDKs include their path prefixes internally const base = LLMOCK_DEFAULT_BASE const openaiUrl = `${base}/v1` // BytePlus Ark's data plane lives under /api/v3. aimock's compat-path // normalizer rewrites any non-/v1//v2 path ending in /chat/completions to // /v1/chat/completions, so the Ark prefix reaches the native OpenAI handler // untouched — see testing/e2e/README.md § "BytePlus (Ark) path handling". const arkUrl = `${base}/api/v3` // X-Test-Id header for per-test sequenceIndex isolation in aimock const testHeaders = testId ? { 'X-Test-Id': testId } : undefined // The Gemini Interactions API lives at a different endpoint // (POST /v1beta/interactions) and uses a different adapter than the // standard Gemini chat path. if (provider === 'gemini' && feature === 'stateful-interactions') { return createChatOptions({ adapter: createGeminiTextInteractions( model as 'gemini-2.5-flash', DUMMY_KEY, { baseURL: base, defaultHeaders: testHeaders }, ), }) } // Agentic video understanding uses the standard Gemini chat adapter, but a // video part with metadata.processing:'agentic' makes it route through the // Interactions API (POST /v1beta/interactions). aimock defaults that endpoint // to streaming SSE, which the non-streaming interactions.create() can't read, // so this feature gets a dedicated mount prefix that returns a completed JSON // interaction. Keeps stateful-interactions on the native handler untouched. if (provider === 'gemini' && feature === 'video-understanding') { return createChatOptions({ adapter: createGeminiChat(model as 'gemini-2.5-flash', DUMMY_KEY, { baseURL: `${base}/vu-interactions`, defaultHeaders: testHeaders, }), }) } const factories: Record { adapter: AnyTextAdapter }> = { openai: () => createChatOptions({ adapter: createOpenaiChat(model as 'gpt-4o', DUMMY_KEY, { baseURL: openaiUrl, defaultHeaders: testHeaders, }), }), anthropic: () => createChatOptions({ adapter: createAnthropicChatWithClient( model as 'claude-sonnet-4-5', new Anthropic({ apiKey: DUMMY_KEY, baseURL: base, defaultHeaders: testHeaders, }), ), }), gemini: () => createChatOptions({ adapter: createGeminiChat(model as 'gemini-2.5-flash', DUMMY_KEY, { baseURL: base, defaultHeaders: testHeaders, }), }), // Gemini on Vertex. Dummy ADC + project/location so the SDK posts // `/v1/projects/{p}/locations/{l}/publishers/google/models/{m}:…`, // which aimock already serves. See vertex-e2e.ts. vertex: () => createChatOptions({ adapter: vertexText( model as 'gemini-2.5-flash', vertexE2eConfig(base, testHeaders), ), }), // Grok on Vertex. Dummy ADC + aimock `/v1` so the OpenAI Responses // client hits the same path as the xAI Grok row. The factory still // prefixes the wire model with `xai/`. 'vertex-grok': () => createChatOptions({ adapter: grokVertexText(model as 'grok-4.3', { project: 'e2e-project', location: 'global', baseURL: openaiUrl, authClient: vertexE2eAuthClient(), defaultHeaders: testHeaders, }), }), // Mistral on Vertex. `resolveRequestUrl` skips the publisher // `:rawPredict` rewrite and posts to aimock `/v1/chat/completions`. 'vertex-mistral': () => createChatOptions({ adapter: mistralVertexText(model as 'mistral-medium-3', { project: 'e2e-project', location: 'us-central1', authClient: vertexE2eAuthClient(), defaultHeaders: testHeaders, resolveRequestUrl: () => `${base}/v1/chat/completions`, }), }), ollama: () => createChatOptions({ adapter: createOllamaChat(model as 'mistral', { baseURL: base, defaultHeaders: testHeaders, }), }), groq: () => createChatOptions({ adapter: createGroqText(model as 'openai/gpt-oss-120b', DUMMY_KEY, { baseURL: base, defaultHeaders: testHeaders, }), }), grok: () => createChatOptions({ adapter: createGrokText(model as 'grok-build-0.1', DUMMY_KEY, { baseURL: openaiUrl, defaultHeaders: testHeaders, }), }), // NOTE: Only the OpenAI-compatible Bedrock paths are E2E-covered here. // The default `bedrock-converse` adapter uses the AWS binary event-stream // (vnd.amazon.eventstream) Converse protocol, which aimock cannot replay — // that path is covered by unit tests in packages/ai-bedrock/tests/converse/ // instead. See testing/e2e/README.md § "Bedrock Converse coverage gap". bedrock: () => createChatOptions({ adapter: createBedrockText( model as 'openai.gpt-oss-120b-1:0', DUMMY_KEY, { baseURL: openaiUrl, defaultHeaders: testHeaders, // Converse is now the default; this matrix entry exercises the // OpenAI-compatible Chat Completions path, so pin api: 'chat'. api: 'chat', }, ), }), 'bedrock-responses': () => createChatOptions({ adapter: createBedrockText( model as 'openai.gpt-oss-120b-1:0', DUMMY_KEY, { baseURL: openaiUrl, defaultHeaders: testHeaders, api: 'responses', }, ), }), openrouter: () => { // OpenRouter SDK exposes an HTTPClient with beforeRequest hooks. Use // that to inject X-Test-Id, since `defaultHeaders` isn't supported and // the SDK strips query params off `serverURL` when building per-request // URLs (it does `new URL(path, baseURL)` which drops the search), so // the previous `?testId=...` trick never actually reached aimock and // multiple openrouter tests collided on the `__default__` test bucket. const httpClient = new HTTPClient() if (testId) { httpClient.addHook('beforeRequest', (req) => { const next = new Request(req) next.headers.set('X-Test-Id', testId) return next }) } return createChatOptions({ adapter: createOpenRouterText(model as 'openai/gpt-4o', DUMMY_KEY, { serverURL: openaiUrl, httpClient, }), }) }, 'openrouter-responses': () => { // Same X-Test-Id injection rationale as the chat-completions factory // above. The beta Responses endpoint uses the same SDK base URL + // HTTPClient surface. const httpClient = new HTTPClient() if (testId) { httpClient.addHook('beforeRequest', (req) => { const next = new Request(req) next.headers.set('X-Test-Id', testId) return next }) } return createChatOptions({ adapter: createOpenRouterResponsesText( model as 'openai/gpt-4o', DUMMY_KEY, { serverURL: openaiUrl, httpClient }, ), }) }, 'vercel-gateway': () => createChatOptions({ adapter: createVercelGatewayText(model as 'openai/gpt-5.5', DUMMY_KEY, { api: 'chat', baseURL: openaiUrl, defaultHeaders: testHeaders, }), }), 'vercel-gateway-responses': () => createChatOptions({ adapter: createVercelGatewayText(model as 'openai/gpt-5.5', DUMMY_KEY, { api: 'responses', baseURL: openaiUrl, defaultHeaders: testHeaders, }), }), lovable: () => createChatOptions({ adapter: createLovableText(model as 'openai/gpt-5.5', DUMMY_KEY, { api: 'chat', baseURL: openaiUrl, defaultHeaders: testHeaders, }), }), 'lovable-responses': () => createChatOptions({ adapter: createLovableText(model as 'openai/gpt-5.5', DUMMY_KEY, { api: 'responses', baseURL: openaiUrl, defaultHeaders: testHeaders, }), }), 'openai-compatible': () => createChatOptions({ adapter: openaiCompatibleText(model, { baseURL: openaiUrl, apiKey: DUMMY_KEY, defaultHeaders: testHeaders, }), }), // Same wire on the Responses API, but reasoning frames arrive under the // legacy `response.reasoning.delta` name via `legacyReasoningFetch`. 'openai-compatible-legacy': () => createChatOptions({ adapter: openaiCompatibleText(model, { baseURL: openaiUrl, apiKey: DUMMY_KEY, defaultHeaders: testHeaders, api: 'responses', fetch: legacyReasoningFetch, }), }), mistral: () => createChatOptions({ adapter: createMistralText(model as 'mistral-large-latest', DUMMY_KEY, { baseURL: base, defaultHeaders: testHeaders, }), }), byteplus: () => createChatOptions({ adapter: createBytePlusText(model as BytePlusChatModel, DUMMY_KEY, { baseURL: arkUrl, defaultHeaders: testHeaders, }), }), elevenlabs: () => { throw new Error( 'ElevenLabs has no text/chat adapter — use createTTSAdapter or createTranscriptionAdapter.', ) }, llmgateway: () => createChatOptions({ adapter: createLLMGatewayText(model as 'gpt-5.6-terra', DUMMY_KEY, { baseURL: openaiUrl, defaultHeaders: testHeaders, }), }), cloudflare: () => createChatOptions({ adapter: createCloudflareText(model, { accountId: 'e2e-account', apiKey: DUMMY_KEY, baseURL: openaiUrl, defaultHeaders: testHeaders, }), }), // Neon's baseURL is the bare branch host; the adapter appends each // route. aimock serves `/v1/chat/completions` and strips the `/openai/` // prefix of `/openai/v1/responses` itself. neon: () => createChatOptions({ adapter: createNeonText(model as 'gemini-3-8-flash', DUMMY_KEY, { baseURL: base, defaultHeaders: testHeaders, }), }), // Claude goes to Neon's `/anthropic/v1/messages`, which aimock does not // map, so this fetch moves it to aimock's `/v1/messages`. The unit tests // pin the real path. 'neon-anthropic': () => createChatOptions({ adapter: createNeonText(model as 'claude-sonnet-4-5', DUMMY_KEY, { baseURL: base, defaultHeaders: testHeaders, fetch: (input, init) => fetch(String(input).replace('/anthropic/v1/', '/v1/'), init), }), }), 'neon-responses': () => createChatOptions({ adapter: createNeonText(model as 'gpt-5-3-codex', DUMMY_KEY, { baseURL: base, defaultHeaders: testHeaders, }), }), // Muse speaks OpenAI-compatible Chat Completions (openai-base), same wire // as the openai-compatible row — point it at aimock's /v1. muse: () => createChatOptions({ adapter: createMuseText(model as 'muse-ember-1.0', DUMMY_KEY, { baseURL: openaiUrl, defaultHeaders: testHeaders, }), }), } return factories[provider]() }