Press n or j to go to the next uncovered block, b, p or k for the previous block.
| 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 | /** * API route for streaming text chat with a geometry teacher — GPT-5.2 Responses API. * * POST /api/realtime/euclid/chat * Body: { messages, propositionId, currentStep, isComplete, playgroundMode, * constructionGraph, toolState, proofFacts, stepList, screenshot?, * characterId?, attitudeId?, tools? } * Returns: SSE stream of { text } deltas and { toolCall } events, ending with [DONE] */ import { withAuth } from '@/lib/auth/withAuth' import { recordOpenAiResponsesStreamUsage } from '@/lib/ai-usage/helpers' import { AiFeature } from '@/lib/ai-usage/features' import { getTeacherConfig } from '@/components/toys/euclid/characters/registry' import type { AttitudeId } from '@/components/toys/euclid/agent/attitudes/types' import { getAttitude } from '@/components/toys/euclid/agent/attitudes' import type { RealtimeTool } from '@/lib/voice/types' interface ChatMessage { role: 'user' | 'assistant' | 'assistant_tool_call' | 'tool' content: string toolCallId?: string toolCallName?: string toolCallArgs?: string } /** Convert our RealtimeTool format to OpenAI Responses API function tool format. */ function toOpenAITool(tool: RealtimeTool) { return { type: 'function' as const, name: tool.name, description: tool.description, parameters: tool.parameters, } } export const POST = withAuth(async (request, { userId }) => { const body = await request.json() const { messages, propositionId, currentStep, isComplete, playgroundMode, constructionGraph, toolState, proofFacts, stepList, screenshot, isMobile, characterId, attitudeId, tools: clientTools, } = body as { messages: ChatMessage[] propositionId: number currentStep: number isComplete: boolean playgroundMode: boolean constructionGraph: string toolState: string proofFacts: string stepList: string screenshot?: string isMobile?: boolean characterId?: string attitudeId?: AttitudeId tools?: RealtimeTool[] } if (!Array.isArray(messages) || messages.length === 0) { return new Response(JSON.stringify({ error: 'messages is required' }), { status: 400, headers: { 'Content-Type': 'application/json' }, }) } const apiKey = process.env.LLM_OPENAI_API_KEY || process.env.OPENAI_API_KEY if (!apiKey) { return new Response(JSON.stringify({ error: 'OpenAI API key not configured' }), { status: 503, headers: { 'Content-Type': 'application/json' }, }) } const config = getTeacherConfig(characterId, attitudeId) // Build system context — append chatDirective for author mode const attitude = attitudeId ? getAttitude(attitudeId) : undefined let systemText = config.buildChatSystemPrompt({ propositionId: typeof propositionId === 'number' ? propositionId : 1, currentStep, isComplete, playgroundMode, constructionGraph, toolState, proofFacts, stepList, isMobile, attitudeId, }) if (attitude?.chatDirective) { systemText += `\n\n=== AUTHOR COLLABORATION DIRECTIVE ===\n${attitude.chatDirective}` } // Build input for the Responses API // System context goes as the first user message, then conversation history const input: Array<Record<string, unknown>> = [ { role: 'user', content: [ { type: 'input_text', text: `[SYSTEM CONTEXT — not from the student]\n\n${systemText}`, }, ], }, { role: 'assistant', content: [ { type: 'output_text', text: config.chatAssistantPriming, }, ], }, ] // Add conversation history — handle tool call/result messages for (const msg of messages) { if (msg.role === 'user') { const contentParts: Array<Record<string, unknown>> = [ { type: 'input_text', text: msg.content }, ] input.push({ role: 'user', content: contentParts }) } else if (msg.role === 'assistant_tool_call' && msg.toolCallId) { // Tool call the model made — must precede the function_call_output input.push({ type: 'function_call', call_id: msg.toolCallId, name: msg.toolCallName, arguments: msg.toolCallArgs ?? '{}', }) } else if (msg.role === 'tool' && msg.toolCallId) { // Tool result — follows the function_call_output format for Responses API input.push({ type: 'function_call_output', call_id: msg.toolCallId, output: msg.content, }) } else { input.push({ role: 'assistant', content: [{ type: 'output_text', text: msg.content }], }) } } // Attach screenshot to the last user message if provided if (screenshot && typeof screenshot === 'string') { let lastUserMsg: Record<string, unknown> | undefined for (let i = input.length - 1; i >= 0; i--) { if (input[i].role === 'user') { lastUserMsg = input[i] break } } if (lastUserMsg && Array.isArray(lastUserMsg.content)) { const base64 = screenshot.includes(',') ? screenshot.split(',')[1] : screenshot ;(lastUserMsg.content as Array<Record<string, unknown>>).push({ type: 'input_image', image_url: `data:image/png;base64,${base64}`, }) } } // Build OpenAI request body const openaiBody: Record<string, unknown> = { model: 'gpt-5.2', input, stream: true, reasoning: { effort: 'none', }, } // Include tools if provided (author mode) if (clientTools && clientTools.length > 0) { openaiBody.tools = clientTools.map(toOpenAITool) } // Call the Responses API with streaming console.log( '[euclid-chat-api] calling OpenAI Responses API, messageCount=%d, hasScreenshot=%s, character=%s, attitude=%s, tools=%d', messages.length, !!screenshot, config.definition.id, attitudeId ?? 'default', clientTools?.length ?? 0 ) const response = await fetch('https://api.openai.com/v1/responses', { method: 'POST', headers: { Authorization: `Bearer ${apiKey}`, 'Content-Type': 'application/json', }, body: JSON.stringify(openaiBody), }) console.log('[euclid-chat-api] OpenAI response status: %d', response.status) if (!response.ok) { const errText = await response.text() console.error('[euclid-chat-api] API error:', response.status, errText) return new Response( JSON.stringify({ error: `Could not reach ${config.definition.displayName} right now.` }), { status: 502, headers: { 'Content-Type': 'application/json' }, } ) } // Stream the response as SSE const encoder = new TextEncoder() const reader = response.body?.getReader() if (!reader) { return new Response(JSON.stringify({ error: 'No response stream' }), { status: 502, headers: { 'Content-Type': 'application/json' }, }) } const stream = new ReadableStream({ async start(controller) { const decoder = new TextDecoder() let buffer = '' // Track function calls being accumulated const pendingCalls = new Map<string, { callId: string; name: string; arguments: string }>() try { while (true) { const { done, value } = await reader.read() if (done) break buffer += decoder.decode(value, { stream: true }) // Process SSE lines from the OpenAI stream const lines = buffer.split('\n') buffer = lines.pop() ?? '' for (const line of lines) { if (!line.startsWith('data: ')) continue const data = line.slice(6).trim() if (data === '[DONE]') continue try { const event = JSON.parse(data) // Text deltas if (event.type === 'response.output_text.delta' && event.delta) { controller.enqueue( encoder.encode(`data: ${JSON.stringify({ text: event.delta })}\n\n`) ) } // Function call: output item added else if ( event.type === 'response.output_item.added' && event.item?.type === 'function_call' ) { pendingCalls.set(event.item.id || event.output_index?.toString(), { callId: event.item.call_id || event.item.id || '', name: event.item.name || '', arguments: '', }) } // Function call: arguments delta else if (event.type === 'response.function_call_arguments.delta') { const key = event.item_id || event.output_index?.toString() const pending = pendingCalls.get(key) if (pending) { pending.arguments += event.delta || '' } } // Function call: arguments complete else if (event.type === 'response.function_call_arguments.done') { const key = event.item_id || event.output_index?.toString() const pending = pendingCalls.get(key) if (pending) { // Use the final arguments from the done event if available const finalArgs = event.arguments || pending.arguments try { const parsedArgs = JSON.parse(finalArgs) controller.enqueue( encoder.encode( `data: ${JSON.stringify({ toolCall: { callId: pending.callId, name: pending.name, arguments: parsedArgs, }, })}\n\n` ) ) } catch { console.error('[euclid-chat-api] failed to parse function args:', finalArgs) } pendingCalls.delete(key) } } // Response completed — record usage else if (event.type === 'response.completed' && event.response?.usage) { recordOpenAiResponsesStreamUsage( { input_tokens: event.response.usage.input_tokens, output_tokens: event.response.usage.output_tokens, }, event.response.model ?? 'gpt-5.2', { userId, feature: AiFeature.EUCLID_CHAT } ) } // Error events else if (event.type === 'error') { const errMsg = event.error?.message || 'An error occurred' const errCode = event.error?.code || 'unknown' console.error('[euclid-chat-api] stream error: %s — %s', errCode, errMsg) const isQuota = /quota/i.test(errCode) const userMessage = isQuota ? `${config.definition.displayName} is unavailable right now. Try again later.` : `Something went wrong. Try again later.` controller.enqueue( encoder.encode(`data: ${JSON.stringify({ error: userMessage })}\n\n`) ) } } catch { // Skip malformed JSON lines } } } } catch (err) { console.error('[euclid-chat] Stream error:', err) } finally { controller.enqueue(encoder.encode('data: [DONE]\n\n')) controller.close() } }, }) return new Response(stream, { headers: { 'Content-Type': 'text/event-stream', 'Cache-Control': 'no-cache', Connection: 'keep-alive', }, }) }) |