[{"_1":2,"_180":-5,"_181":-5},"loaderData",{"_3":4,"_96":97},"root",{"_5":6,"_37":38,"_69":70},"ctx",{"_7":8,"_23":-5,"_24":25,"_26":27,"_28":20,"_29":20,"_30":20,"_31":-5,"_32":33,"_34":20,"_35":20,"_36":20},"auth",{"_9":-5,"_10":-5,"_11":-5,"_12":-5,"_13":-5,"_14":-5,"_15":-5,"_16":-5,"_17":-5,"_18":-5,"_19":20,"_21":22},"sessionClaims","sessionId","sessionStatus","userId","actor","orgId","orgRole","orgSlug","orgPermissions","factorVerificationAge","isAuthenticated",false,"tokenType","session_token","user","preferencesKeybindings","default","preferencesTheme","light","preferencesFixers","preferencesCopilot","preferencesSelectionAi","preferencesChatPanelState","notificationCount",0,"flagDenols","isBusinessOrgMember","isHigherTownieTier","ENV",{"_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},"REACT_APP_SENTRY_DSN","https://586020d86f05489db63a554690acf1e9@o1398359.ingest.sentry.io/6724499","VAL_RUNTIME_NAME","prod","RENDER_GIT_COMMIT","7d99597163ee6d2ecd75e0020d9771df6719deb7","RENDER_GIT_BRANCH","main","CLERK_PUBLISHABLE_KEY","pk_live_Y2xlcmsudmFsLnRvd24k","POSTHOG_API_TOKEN","phc_nmoyuAkStEZuUp69rYlg1DZgrIt19Gx0Y4Hwf8tGgJP","POSTHOG_API_HOST","https://wart.val.town","WEB_API_TEMPLATE","https://{val}.web.val.run","APP_URL","https://www.val.town","MODULE_URL","https://esm.town","VANITY_WEB_API_TEMPLATE","https://{label}.val.run/","DENOLS_URL","https://lsp.val.town","NODE_ENV","production","CLOUDFLARE_TURNSTILE_SITE_KEY","0x4AAAAAACtaIZJH7s9rBMZ6","API_URL","https://api.val.town","clerkState",{"_71":72},"__internal_clerk_state",{"_73":74,"_75":48,"_76":77,"_78":77,"_79":20,"_80":77,"_81":77,"_82":77,"_83":77,"_84":85,"_94":20,"_95":20},"__clerk_ssr_state",{"_21":22,"_9":-5,"_10":-5,"_11":-5,"_12":-5,"_13":-5,"_14":-5,"_15":-5,"_16":-5,"_17":-5,"_18":-5,"_19":20},"__publishableKey","__proxyUrl","","__domain","__isSatellite","__signInUrl","__signUpUrl","__afterSignInUrl","__afterSignUpUrl","__clerk_debug",{"_86":20,"_19":20,"_87":77,"_88":89,"_90":77,"_91":48,"_92":20,"_93":77},"isSignedIn","proxyUrl","reason","session-token-and-uat-missing","message","publishableKey","isSatellite","domain","__telemetryDisabled","__telemetryDebug","routes/embed.x.$handle.$name.$",{"_98":99,"_100":101,"_102":103,"_104":105,"_161":162,"_177":178,"_179":-5},"handle","MadeiInGermany","name","Kimi-K2-0905-Exp","path","main.tsx","project",{"_106":107,"_100":101,"_108":109,"_110":-5,"_111":20,"_112":113,"_114":115,"_116":-5,"_12":117,"_118":119,"_120":121,"_122":-5,"_123":-5,"_124":33,"_14":125,"_126":-5,"_127":-5,"_128":-5,"_129":-5,"_130":131,"_132":131,"_133":134,"_135":134,"_136":134,"_137":131,"_138":131,"_23":139,"_142":143,"_147":148,"_160":20},"id","5e15a4cc-253f-11f1-8d00-42dde27851f2","createdAt",["D",1774108951631],"deletedAt","legacyEnvVars","privacy","public","description","https://poe.com/Kimi-K2-0905-Exp","imageUrl","01995b22-1b41-771e-a0ac-a4cfc92d3606","pinnedFileId","019d1122-44a2-763d-aff1-b483e3227f40","pinPreview",true,"pinMigrated","screenshotVersion","likeCount","01995b22-1b41-771e-a0ac-a96f3508eae1","tursoName","tursoHostname","tursoMeasuredAt","pinned","userPerm","r","valPerm","blobPerm","rw","emailPerm","sqlitePerm","projectPerm","telemetryPerm",{"_98":99,"_140":141},"profileImageUrl","https://imagedelivery.net/iHX6Ovru0O7AjmyT5yZRoA/a46b91d4-4b2a-48c3-0c37-9c04ea652a00/square512","org",{"_98":99,"_144":141,"_12":117,"_145":146},"avatar","tier","free","branch",{"_149":150,"_100":46,"_106":151,"_152":107,"_153":-5,"_154":-5,"_155":-5,"_156":-5,"_157":-5,"_158":101,"_159":99},"version",1,"5e19af36-253f-11f1-8d00-42dde27851f2","projectId","forkedBranchId","forkedBranchVersion","forkedProjectPrivacy","forkedProjectId","forkedProjectOrgId","projectName","userHandle","isOwned","file",{"_106":163,"_100":103,"_164":119,"_165":-5,"_166":151,"_167":150,"_168":121,"_108":169,"_149":33,"_170":171,"_172":173,"_174":20,"_175":-5,"_102":176},"5e1cec64-253f-11f1-8d00-42dde27851f2","fileId","parentId","branchId","branchVersion","latest",["D",1774108951631],"fileType","http","content","import process from \"node:process\";\nimport fastapi from \"npm:fastapi\";\nimport fastapi_poe from \"npm:fastapi-poe\";\n\nconst TOKEN_PRICING = {\n INPUT_TOKEN_COST: 10,\n OUTPUT_TOKEN_COST: 10,\n BASE_COST: 10,\n};\n\nconst K_LENGTH = 32;\n\nfunction countTokens(text: string): number {\n return Math.ceil(text.length / 4);\n}\n\nclass UnsupportedFileTypeError extends Error {\n statusCode: number;\n code: string;\n constructor(fileName: string, detectedType: string, message?: string) {\n const defaultMessage = `Unsupported file type: ${fileName} (${detectedType} format)`;\n super(message || defaultMessage);\n this.name = \"UnsupportedFileTypeError\";\n this.statusCode = 415;\n this.code = \"unsupported_file_type\";\n }\n}\n\nclass FileProcessingError extends Error {\n statusCode: number;\n code: string;\n originalError?: Error;\n constructor(fileName: string, operation: string, originalError?: Error) {\n const message = `Failed to process ${fileName} during ${operation}: ${originalError?.message || 'Unknown error'}`;\n super(message);\n this.name = \"FileProcessingError\";\n this.statusCode = 422;\n this.code = \"file_processing_failed\";\n this.originalError = originalError;\n }\n}\n\nclass AuthenticationError extends Error {\n statusCode: number;\n code: string;\n constructor(message = \"Authentication failed\") {\n super(message);\n this.name = \"AuthenticationError\";\n this.statusCode = 401;\n this.code = \"authentication_failed\";\n }\n}\n\nclass APIConnectionError extends Error {\n statusCode: number;\n code: string;\n service: string;\n constructor(service: string, originalError?: Error) {\n const message = `Failed to connect to ${service}: ${originalError?.message || 'Connection error'}`;\n super(message);\n this.name = \"APIConnectionError\";\n this.statusCode = 502;\n this.code = \"api_connection_failed\";\n this.service = service;\n }\n}\n\nclass FileSizeError extends Error {\n statusCode: number;\n code: string;\n constructor(fileName: string, maxSize?: string) {\n const message = maxSize ? `File ${fileName} exceeds maximum size limit of ${maxSize}` : `File ${fileName} is too large to process`;\n super(message);\n this.name = \"FileSizeError\";\n this.statusCode = 413;\n this.code = \"file_too_large\";\n }\n}\n\ntype BotSettings = {\n allow_attachments?: boolean;\n expand_text_attachments?: boolean;\n enable_multi_bot_chat_prompting?: boolean;\n rate_card?: string;\n cost_label?: string;\n supports_system_messages?: boolean;\n suggested_replies?: boolean;\n};\n\ntype Attachment = {\n url: string;\n content_type: string;\n name: string;\n parsed_content?: string;\n};\n\ntype Message = {\n role: \"system\" | \"user\" | \"bot\";\n content: string;\n content_type: string;\n timestamp: number;\n message_id: string;\n feedback: { type: \"like\" | \"dislike\"; reason?: string; }[];\n attachments: Attachment[];\n};\n\ntype Query = {\n version: string;\n type: \"query\";\n query: Message[];\n message_id: string;\n user_id: string;\n conversation_id: string;\n metadata: string;\n language_code: string;\n};\n\ntype Events = {\n meta: { content_type: string };\n text: { text: string };\n replace_response: { text: string };\n suggested_reply: { text: string };\n error: { allow_retry?: boolean; text: string };\n done: { input_tokens?: number; output_tokens?: number; cost?: number };\n};\n\ntype SendEventFn = (\n eventName: EventName,\n data: Events[EventName],\n) => void;\n\nfunction formatErrorResponse(error: Error): { error: { message: string; code?: string; type: string; details?: any; }; } {\n if (error instanceof UnsupportedFileTypeError) {\n return {\n error: {\n message: error.message,\n code: error.code,\n type: \"UnsupportedFileTypeError\",\n details: {\n supported_formats: {\n images: [\"JPEG\", \"PNG\", \"WebP\", \"HEIC\", \"GIF\", \"TIFF\", \"SVG\"],\n videos: [\"WebM\", \"MP4\", \"AVI\", \"MOV\", \"MKV\", \"WMV\", \"FLV\", \"M4V\"],\n audio: [\"MP3\", \"WAV\", \"OGG\", \"M4A\", \"FLAC\", \"AAC\", \"WMA\", \"OPUS\"],\n documents: [\"PDF\", \"TXT\", \"JSON\", \"JS\", \"Python\", \"etc.\"] // Note: \"etc.\" suggests more, but checkFileType only supports specific ones as 'supported'\n }\n }\n }\n };\n }\n if (error instanceof FileProcessingError) {\n return {\n error: {\n message: error.message,\n code: error.code,\n type: \"FileProcessingError\",\n details: { original_error: error.originalError?.message }\n }\n };\n }\n if (error instanceof AuthenticationError) {\n return { error: { message: error.message, code: error.code, type: \"AuthenticationError\" } };\n }\n if (error instanceof APIConnectionError) {\n return { error: { message: error.message, code: error.code, type: \"APIConnectionError\", details: { service: error.service } } };\n }\n if (error instanceof FileSizeError) {\n return { error: { message: error.message, code: error.code, type: \"FileSizeError\" } };\n }\n return { error: { message: error.message || \"Internal Server Error\", code: \"internal_error\", type: \"InternalError\" } };\n}\n\nfunction authenticateRequest(req: Request): { success: boolean; error?: AuthenticationError } {\n try {\n const expectedBotKey = process.env.BOT_KEY;\n if (!expectedBotKey) {\n return { success: false, error: new AuthenticationError(\"Bot key not configured\") };\n }\n if (expectedBotKey.length !== K_LENGTH) {\n return { success: false, error: new AuthenticationError(\"Invalid bot key configuration\") };\n }\n const authHeader = req.headers.get('Authorization');\n if (!authHeader) {\n return { success: false, error: new AuthenticationError(\"Missing Authorization header\") };\n }\n let providedKey = '';\n if (authHeader.startsWith('Bearer ')) {\n providedKey = authHeader.substring(7);\n } else if (authHeader.startsWith('BOT_KEY ')) {\n providedKey = authHeader.substring(8);\n } else {\n providedKey = authHeader;\n }\n if (!providedKey) {\n return { success: false, error: new AuthenticationError(\"Invalid Authorization header format\") };\n }\n if (providedKey.length !== K_LENGTH) {\n return { success: false, error: new AuthenticationError(\"Invalid key length\") };\n }\n if (!constantTimeEquals(providedKey, expectedBotKey)) {\n return { success: false, error: new AuthenticationError(\"Invalid bot key\") };\n }\n return { success: true };\n } catch (error) {\n console.error(\"Authentication error:\", error);\n return { success: false, error: new AuthenticationError(\"Authentication failed\") };\n }\n}\n\nfunction constantTimeEquals(a: string, b: string): boolean {\n if (a.length !== b.length) {\n return false;\n }\n let result = 0;\n for (let i = 0; i < a.length; i++) {\n result |= a.charCodeAt(i) ^ b.charCodeAt(i);\n }\n return result === 0;\n}\n\nfunction checkFileType(fileName: string): { type: 'image' | 'video' | 'audio' | 'text' | 'pdf' | 'office' | 'archive' | 'unsupported', isSupported: boolean } {\n const lowerFileName = fileName.toLowerCase();\n const imageExtensions = ['.jpeg', '.jpg', '.png', '.webp', '.heic', '.gif', '.bmp', '.tiff', '.tif', '.svg'];\n if (imageExtensions.some(ext => lowerFileName.endsWith(ext))) {\n return { type: 'image', isSupported: true };\n }\n const videoExtensions = ['.webm', '.mp4', '.avi', '.mov', '.mkv', '.wmv', '.flv', '.m4v'];\n if (videoExtensions.some(ext => lowerFileName.endsWith(ext))) {\n return { type: 'video', isSupported: true };\n }\n const audioExtensions = ['.mp3', '.wav', '.ogg', '.m4a', '.flac', '.aac', '.wma', '.opus'];\n if (audioExtensions.some(ext => lowerFileName.endsWith(ext))) {\n return { type: 'audio', isSupported: true };\n }\n const textExtensions = [\n '.txt', '.json', '.js', '.ts', '.jsx', '.tsx', '.html', '.css', '.xml', '.md', '.csv', '.py', '.java', '.cpp', '.c', '.h', '.php', '.rb', '.go', '.rs', '.swift', '.kt', '.yaml', '.yml', '.toml', '.ini', '.conf', '.log', '.sql', '.sh', '.bat', '.ps1'\n ];\n if (textExtensions.some(ext => lowerFileName.endsWith(ext))) {\n return { type: 'text', isSupported: true };\n }\n if (lowerFileName.endsWith('.pdf')) {\n return { type: 'pdf', isSupported: true };\n }\n const officeExtensions = ['.doc', '.docx', '.xls', '.xlsx', '.ppt', '.pptx', '.odt', '.ods', '.odp'];\n if (officeExtensions.some(ext => lowerFileName.endsWith(ext))) {\n return { type: 'office', isSupported: false };\n }\n const archiveExtensions = ['.zip', '.rar', '.7z', '.tar', '.gz', '.bz2'];\n if (archiveExtensions.some(ext => lowerFileName.endsWith(ext))) {\n return { type: 'archive', isSupported: false };\n }\n return { type: 'unsupported', isSupported: false };\n}\n\nfunction validateFileAttachments(attachments: any[]): void {\n const unsupportedFiles = attachments.filter(attachment => {\n const fileName = attachment.name || '';\n const { isSupported } = checkFileType(fileName);\n return !isSupported;\n });\n if (unsupportedFiles.length > 0) {\n const fileDetails = unsupportedFiles.map(f => {\n const { type } = checkFileType(f.name || '');\n return `${f.name} (${type} format)`;\n }).join(', ');\n throw new UnsupportedFileTypeError(\n fileDetails,\n 'multiple',\n `Unsupported file types detected: ${fileDetails}`\n );\n }\n}\n\nasync function processImageWithLlama(attachment: any): Promise {\n try {\n const fileData = await fetch(attachment.url).then(res => {\n if (!res.ok) {\n throw new Error(`Failed to fetch file: ${res.status} ${res.statusText}`);\n }\n return res.arrayBuffer();\n });\n const base64Data = Buffer.from(fileData).toString('base64');\n const mimeType = attachment.content_type;\n const dataUrl = `data:${mimeType};base64,${base64Data}`;\n\n const payload = {\n model: \"meta-llama/llama-4-scout-17b-16e-instruct\",\n messages: [\n {\n role: \"user\",\n content: [\n { type: \"text\", text: `Provide a detailed analysis of this image: ${attachment.name}. Describe what you see, including objects, people, text, colors, composition, and any notable details.` },\n { type: \"image_url\", image_url: { url: dataUrl } }\n ]\n }\n ],\n max_tokens: 2048,\n temperature: 0.7\n };\n\n const response = await fetch(\"https://api.groq.com/openai/v1/chat/completions\", {\n method: 'POST',\n headers: {\n 'Authorization': `Bearer ${process.env.GROQ_API_KEY}`,\n 'Content-Type': 'application/json',\n },\n body: JSON.stringify(payload),\n });\n\n if (!response.ok) {\n const errorText = await response.text().catch(() => \"Unknown error\");\n throw new APIConnectionError(\"Groq Llama API\", new Error(`${response.status} - ${errorText}`));\n }\n const result = await response.json();\n\n if (result.choices?.[0]?.message?.content) {\n return result.choices[0].message.content;\n } else {\n throw new FileProcessingError(attachment.name, \"image analysis\", new Error(\"No content returned from API\"));\n }\n } catch (error) {\n if (error instanceof APIConnectionError || error instanceof FileProcessingError) {\n throw error;\n }\n throw new FileProcessingError(attachment.name, \"image processing\", error);\n }\n}\n\nasync function processFileAttachment(attachment: any): Promise {\n try {\n if (attachment.parsed_content) {\n return attachment.parsed_content;\n }\n\n const fileName = attachment.name || '';\n const { type, isSupported } = checkFileType(fileName);\n\n if (!isSupported) {\n throw new UnsupportedFileTypeError(fileName, type);\n }\n\n if (type === 'image') {\n try {\n return await processImageWithLlama(attachment);\n } catch (llamaError: any) {\n console.warn(\"Llama processing failed, trying Gemini:\", llamaError.message);\n return await processImageWithGemini(attachment);\n }\n } else if (type === 'video') {\n return await processVideoWithGemini(attachment);\n } else if (type === 'audio') {\n return await processAudioWithWhisper(attachment);\n } else if (type === 'pdf') {\n return await processPDFWithGemini(attachment);\n } else if (type === 'text') {\n const textContent = await fetch(attachment.url).then(res => {\n if (!res.ok) {\n throw new Error(`Failed to fetch text file: ${res.status} ${res.statusText}`);\n }\n return res.text();\n });\n return `Content from text file ${attachment.name}:\\n\\n${textContent}`;\n }\n\n throw new FileProcessingError(fileName, \"processing\", new Error(\"No processing method available for this file type\"));\n } catch (error) {\n if (error instanceof UnsupportedFileTypeError || error instanceof FileProcessingError || error instanceof APIConnectionError || error instanceof FileSizeError) {\n throw error;\n }\n throw new FileProcessingError(attachment.name || 'unknown', \"processing\", error);\n }\n}\n\n\nasync function getResponse(req: Query, send: SendEventFn): Promise<{ inputTokens: number; outputTokens: number; cost: number }> {\n let totalInputTokens = 0;\n let totalOutputTokens = 0;\n let initialMessageContents: string[] = []; // Store original message contents\n\n try {\n send(\"meta\", { content_type: \"text/markdown\" });\n\n let fileAnalysisResults = \"\";\n let hasFiles = false;\n\n // Store original message content for potential early error token calculation\n req.query.forEach(msg => {\n initialMessageContents.push(msg.content);\n });\n\n if (req.query.length > 0) {\n const lastMessage = req.query[req.query.length - 1];\n if (lastMessage.attachments && lastMessage.attachments.length > 0) {\n hasFiles = true;\n try {\n validateFileAttachments(lastMessage.attachments);\n } catch (error) {\n if (error instanceof UnsupportedFileTypeError) {\n const errorResponse = formatErrorResponse(error);\n send(\"error\", { text: `❌ **${errorResponse.error.type}**\\n\\n${errorResponse.error.message}\\n\\n**✅ Supported formats:**\\n${Object.entries(errorResponse.error.details.supported_formats).map(([category, formats]) => `• **${category.charAt(0).toUpperCase() + category.slice(1)}:** ${formats.join(', ')}` ).join('\\n')}\\n\\nPlease upload supported file types and try again.`, allow_retry: false });\n \n // For early error, count initial user messages only\n initialMessageContents.forEach(content => totalInputTokens += countTokens(content));\n return { inputTokens: totalInputTokens, outputTokens: 0, cost: TOKEN_PRICING.BASE_COST + (totalInputTokens * TOKEN_PRICING.INPUT_TOKEN_COST) };\n }\n throw error; // Re-throw other errors to be caught by the outer catch\n }\n\n send(\"text\", { text: \"🔍 Processing files...\\n\\n\" });\n for (const attachment of lastMessage.attachments) {\n try {\n const { type } = checkFileType(attachment.name || '');\n const processingMessages = {\n audio: `🎵 Transcribing audio: ${attachment.name}...`,\n video: `🎬 Analyzing video: ${attachment.name}...`,\n image: `🖼️ Analyzing image: ${attachment.name}...`,\n pdf: `📄 Processing PDF: ${attachment.name}...`,\n text: `📝 Reading text file: ${attachment.name}...`,\n unsupported: `📁 Processing file: ${attachment.name}...` // Fallback\n };\n send(\"text\", { text: `${processingMessages[type] || `📁 Processing file: ${attachment.name}...`}\\n` });\n const analysisResult = await processFileAttachment(attachment);\n fileAnalysisResults += `## Content from ${attachment.name}:\\n${analysisResult}\\n\\n---\\n\\n`;\n // Note: Tokens for analysisResult are counted later when `messagesToSend` is built.\n } catch (error: any) {\n const errorResponse = formatErrorResponse(error);\n const errorMessage = `## Error processing ${attachment.name}:\\n**${errorResponse.error.type}:** ${errorResponse.error.message}\\n\\n`;\n fileAnalysisResults += errorMessage;\n send(\"text\", { text: `❌ ${errorResponse.error.message}\\n` });\n // Continue processing other files or main query, don't stop here.\n }\n }\n }\n }\n\n // Now, construct the messages array that will be sent to the LLM\n const messagesToSend = req.query.map((msg, index) => {\n let content = msg.content;\n // If it's the last user message and there are file analysis results, append them\n if (msg.role === \"user\" && index === req.query.length - 1 && fileAnalysisResults) {\n content = `${msg.content}\\n\\n## File Analysis Results:\\n${fileAnalysisResults}`;\n }\n return {\n role: msg.role === \"bot\" ? \"assistant\" : msg.role as \"user\" | \"assistant\" | \"system\",\n content: content,\n };\n });\n\n if (hasFiles && messagesToSend.length > 0) {\n const systemMessage = { role: \"system\" as const, content: \"You are an advanced file analysis assistant. When analyzing file content, provide detailed insights, summaries, and answer any questions about the files. Format your responses clearly with appropriate headers and structure.\" };\n messagesToSend.unshift(systemMessage);\n }\n\n // Calculate total input tokens based on the *final* messages array sent to the LLM\n messagesToSend.forEach(msg => {\n totalInputTokens += countTokens(msg.content);\n });\n\n const models = [\n \"moonshotai/kimi-k2-instruct-0905\",\n \"moonshotai/kimi-k2-instruct\",\n \"groq/compound-mini\",\n \"groq/compound\",\n ];\n\n let outputContent = \"\";\n for (let i = 0; i < models.length; i++) {\n const model = models[i];\n let response: Response | null = null;\n let reader: ReadableStreamDefaultReader | null = null;\n try {\n const controller = new AbortController();\n const timeoutId = setTimeout(() => controller.abort(), 30000); // 30 second timeout\n\n response = await fetch(\"https://api.groq.com/openai/v1/chat/completions\", {\n method: \"POST\",\n headers: {\n \"Authorization\": `Bearer ${process.env.GROQ_API_KEY}`,\n \"Content-Type\": \"application/json\",\n },\n body: JSON.stringify({\n model: model,\n messages: messagesToSend, // Use messagesToSend\n stream: true,\n max_tokens: 2048,\n temperature: 0.7,\n }),\n signal: controller.signal,\n });\n clearTimeout(timeoutId);\n\n if (!response.ok) {\n const errorText = await response.text().catch(() => \"Unknown error\");\n console.error(`Model ${model} failed with status ${response.status}: ${errorText}`);\n if (response.status === 429 && i < models.length - 1) {\n const waitTime = Math.pow(2, i) * 1000;\n console.log(`Rate limited, waiting ${waitTime}ms before trying next model`);\n await sleep(waitTime);\n }\n if (i === models.length - 1) { // If this is the last model and it failed\n throw new APIConnectionError(\"All AI models\", new Error(`Last error: ${response.status} - ${errorText}`));\n }\n continue; // Try next model\n }\n\n if (!response.body) {\n console.warn(`No response body from model ${model}`);\n if (i === models.length - 1) {\n throw new APIConnectionError(model, new Error(\"No response body\"));\n }\n continue; // Try next model\n }\n\n reader = response.body.getReader();\n const decoder = new TextDecoder();\n let buffer = \"\";\n\n // Send replace_response only for the first successful model stream if files were processed\n if (i === 0 && hasFiles) {\n send(\"replace_response\", { text: \"\" });\n }\n\n while (true) {\n const { done, value } = await reader.read();\n if (done) break;\n\n buffer += decoder.decode(value, { stream: true });\n const lines = buffer.split(\"\\n\");\n buffer = lines.pop() || \"\";\n\n for (const line of lines) {\n if (line.startsWith(\"data: \")) {\n const data = line.slice(6).trim();\n if (data === \"[DONE]\") continue;\n if (!data) continue;\n try {\n const parsed = JSON.parse(data);\n const content = parsed.choices?.[0]?.delta?.content;\n if (content) {\n outputContent += content;\n send(\"text\", { text: content });\n }\n } catch (e) {\n console.error(\"Parse error:\", e, \"Data:\", data);\n }\n }\n }\n }\n console.log(`Successfully completed with model: ${model}`);\n reader.releaseLock();\n break; // Exit model retry loop on success\n } catch (error: any) {\n if (reader) {\n try {\n reader.releaseLock();\n } catch (e) {\n console.warn(\"Error releasing reader lock:\", e);\n }\n }\n if (response?.body) {\n try {\n await response.body.cancel();\n } catch (e) {\n console.warn(\"Error canceling response body:\", e);\n }\n }\n console.error(`Model ${model} error:`, error.message);\n if (i === models.length - 1) { // If it's the last model and it failed\n throw error; // Re-throw to be caught by the outer catch\n }\n console.warn(`Trying next model after ${model} failed:`, error.message);\n await sleep(1000);\n }\n }\n\n const outputTokenCount = countTokens(outputContent);\n totalOutputTokens += outputTokenCount;\n\n return { inputTokens: totalInputTokens, outputTokens: totalOutputTokens, cost: TOKEN_PRICING.BASE_COST + (totalInputTokens * TOKEN_PRICING.INPUT_TOKEN_COST) + (totalOutputTokens * TOKEN_PRICING.OUTPUT_TOKEN_COST) };\n\n } catch (error: any) {\n console.error(\"getResponse outer catch error:\", error);\n // If an error occurs before `messagesToSend` is fully calculated (e.g., during initial file validation,\n // where `totalInputTokens` might only contain initialMessageContents),\n // we need to ensure at least that is counted for the error cost.\n if (totalInputTokens === 0) {\n initialMessageContents.forEach(content => totalInputTokens += countTokens(content));\n }\n // Return current accumulated tokens + base cost. Output tokens will be 0 if the error happened before generation.\n return { inputTokens: totalInputTokens, outputTokens: totalOutputTokens, cost: TOKEN_PRICING.BASE_COST + (totalInputTokens * TOKEN_PRICING.INPUT_TOKEN_COST) + (totalOutputTokens * TOKEN_PRICING.OUTPUT_TOKEN_COST) };\n }\n}\n\nasync function getBotSettings(): Promise {\n const rateCard = `| Type | Price |\\n` +\n `|------|------|\\n` +\n `| Base Cost | ${TOKEN_PRICING.BASE_COST} points |\\n` +\n `| Input Token | ${TOKEN_PRICING.INPUT_TOKEN_COST} points |\\n` +\n `| Output Token | ${TOKEN_PRICING.OUTPUT_TOKEN_COST} points |\\n`;\n const costLabel = `${TOKEN_PRICING.BASE_COST}+ points (token-based)`;\n\n return {\n allow_attachments: true,\n expand_text_attachments: true,\n enable_multi_bot_chat_prompting: true,\n rate_card: rateCard,\n cost_label: costLabel,\n supports_system_messages: true,\n suggested_replies: true\n };\n}\n\nconst sleep = (ms: number) => new Promise((resolve) => setTimeout(resolve, ms));\n\nfunction encodeEvent(event: string, data: any = {}): Uint8Array {\n return new TextEncoder().encode(\n `event: ${event}\\ndata: ${JSON.stringify(data)}\\n\\n`,\n );\n}\n\nexport default async function handler(req: Request): Promise {\n if (req.method === 'OPTIONS') {\n return new Response(null, {\n status: 200,\n headers: {\n \"Access-Control-Allow-Origin\": \"*\",\n \"Access-Control-Allow-Methods\": \"POST, OPTIONS\",\n \"Access-Control-Allow-Headers\": \"Content-Type, Authorization\",\n \"Access-Control-Max-Age\": \"86400\",\n },\n });\n }\n\n try {\n const authResult = authenticateRequest(req);\n if (!authResult.success && authResult.error) {\n console.error(\"Authentication failed:\", authResult.error.message);\n const errorResponse = formatErrorResponse(authResult.error);\n return new Response(JSON.stringify(errorResponse), {\n status: authResult.error.statusCode,\n headers: { \"Content-Type\": \"application/json\", \"Access-Control-Allow-Origin\": \"*\" },\n });\n }\n\n let reqBody: any;\n try {\n reqBody = await req.json();\n } catch (e) {\n console.error(\"Request body parse error:\", e);\n const parseError = new Error(\"Invalid JSON in request body\");\n const errorResponse = formatErrorResponse(parseError);\n return new Response(JSON.stringify(errorResponse), {\n status: 400,\n headers: { \"Content-Type\": \"application/json\", \"Access-Control-Allow-Origin\": \"*\" },\n });\n }\n\n console.log(\"Authenticated request body:\", JSON.stringify(reqBody, null, 2));\n\n if (reqBody.type === \"settings\") {\n try {\n const settings = await getBotSettings();\n return new Response(JSON.stringify(settings), {\n status: 200,\n headers: { \"Content-Type\": \"application/json\", \"Access-Control-Allow-Origin\": \"*\" },\n });\n } catch (error) {\n console.error(\"Settings error:\", error);\n const errorResponse = formatErrorResponse(error);\n return new Response(JSON.stringify(errorResponse), {\n status: 500,\n headers: { \"Content-Type\": \"application/json\", \"Access-Control-Allow-Origin\": \"*\" },\n });\n }\n }\n\n if (!reqBody.query || !Array.isArray(reqBody.query)) {\n const validationError = new Error(\"Invalid query format - expected array of messages\");\n const errorResponse = formatErrorResponse(validationError);\n return new Response(JSON.stringify(errorResponse), {\n status: 400,\n headers: { \"Content-Type\": \"application/json\", \"Access-Control-Allow-Origin\": \"*\" },\n });\n }\n\n // Initialize token usage, this will be updated by the getResponse promise resolvers.\n let tokenUsage = { inputTokens: 0, outputTokens: 0, cost: TOKEN_PRICING.BASE_COST };\n\n const body = new ReadableStream({\n start(controller) {\n const send: SendEventFn = (event: keyof Events, data: any) => {\n try {\n controller.enqueue(encodeEvent(event, data));\n } catch (e) {\n console.error(\"Error sending event:\", e);\n }\n };\n\n return getResponse(reqBody as Query, send)\n .then((usage) => {\n tokenUsage = usage; // Update token usage after getResponse completes\n console.log(`Final token usage for user ${reqBody.user_id}: ${JSON.stringify(usage)}`);\n // Always send done event on successful completion\n send(\"done\", { input_tokens: tokenUsage.inputTokens, output_tokens: tokenUsage.outputTokens, cost: tokenUsage.cost });\n })\n .catch((e) => {\n console.error(\"Get response error:\", e);\n const errorResponse = formatErrorResponse(e);\n send(\"error\", { text: `❌ **${errorResponse.error.type}:** ${errorResponse.error.message}`, allow_retry: true });\n \n // If getResponse failed, the returned tokenUsage from `getResponse`'s catch block\n // will have the appropriate error-time cost (base + input tokens).\n // We ensure that `tokenUsage` is set before this catch, or if getResponse's catch\n // returns its own usage, it will be reflected. Let's make sure tokenUsage is set\n // to a valid error state if getResponse throws before returning.\n // For now, getResponse's catch block *returns* a usage, which simplifies this.\n // If the error was truly unhandled and `getResponse` didn't return a usage,\n // `tokenUsage` would retain its initial value, leading to minimal cost.\n send(\"done\", { input_tokens: tokenUsage.inputTokens, output_tokens: tokenUsage.outputTokens, cost: tokenUsage.cost });\n })\n .finally(() => {\n try {\n controller.close();\n } catch (e) {\n console.warn(\"Error closing controller:\", e);\n }\n });\n },\n });\n\n // Set headers with token usage for Poe platform.\n // The `tokenUsage` value here reflects its initial state or the state after `getResponse`'s .catch if an error occurred early\n // before the stream could fully process. This is correct as `Response` headers are immutable once returned.\n return new Response(body, {\n status: 200,\n headers: {\n \"Content-Type\": \"text/event-stream\",\n \"Cache-Control\": \"no-cache\",\n \"Connection\": \"keep-alive\",\n \"Access-Control-Allow-Origin\": \"*\",\n \"Access-Control-Allow-Methods\": \"POST, OPTIONS\",\n \"Access-Control-Allow-Headers\": \"Content-Type, Authorization\",\n \"X-Poe-Bot-Output-Tokens\": tokenUsage.outputTokens.toString(),\n \"X-Poe-Bot-Input-Tokens\": tokenUsage.inputTokens.toString(),\n \"X-Poe-Bot-Cost\": tokenUsage.cost.toString(),\n },\n });\n } catch (error) {\n console.error(\"Main handler error:\", error);\n const errorResponse = formatErrorResponse(error);\n return new Response(JSON.stringify(errorResponse), {\n status: 500,\n headers: { \"Content-Type\": \"application/json\", \"Access-Control-Allow-Origin\": \"*\" },\n });\n }\n}\n\n// Helper functions for file processing (kept as is, but added to processFileAttachment flow)\nasync function processVideoWithGemini(attachment: any): Promise {\n try {\n const fileData = await fetch(attachment.url).then(res => {\n if (!res.ok) {\n throw new Error(`Failed to fetch file: ${res.status} ${res.statusText}`);\n }\n return res.arrayBuffer();\n });\n if (fileData.byteLength > 50 * 1024 * 1024) {\n throw new FileSizeError(attachment.name, \"50MB\");\n }\n const base64Data = Buffer.from(fileData).toString('base64');\n const mimeType = attachment.content_type;\n\n const payload = {\n contents: [{\n parts: [\n { text: `Analyze this video file: ${attachment.name}. Provide a detailed description of the content, including visual elements, actions, audio content if any, duration estimates, and any text or objects visible.` },\n { inline_data: { mime_type: mimeType, data: base64Data } }\n ]\n }],\n generationConfig: {\n temperature: 0.7,\n maxOutputTokens: 2048,\n }\n };\n\n const response = await fetch(`https://generativelanguage.googleapis.com/v1beta/models/gemini-flash-latest:generateContent?key=${process.env.GEMINI_API_KEY}`, {\n method: 'POST',\n headers: { 'Content-Type': 'application/json', },\n body: JSON.stringify(payload),\n });\n\n if (!response.ok) {\n const errorText = await response.text().catch(() => \"Unknown error\");\n throw new APIConnectionError(\"Gemini API\", new Error(`${response.status} - ${errorText}`));\n }\n const result = await response.json();\n\n if (result.candidates?.[0]?.content?.parts?.[0]?.text) {\n return result.candidates[0].content.parts[0].text;\n } else {\n throw new FileProcessingError(attachment.name, \"video analysis\", new Error(\"No content returned from API\"));\n }\n } catch (error) {\n if (error instanceof APIConnectionError || error instanceof FileProcessingError || error instanceof FileSizeError) {\n throw error;\n }\n throw new FileProcessingError(attachment.name, \"video processing\", error);\n }\n}\n\nasync function processImageWithGemini(attachment: any): Promise {\n try {\n const fileData = await fetch(attachment.url).then(res => {\n if (!res.ok) {\n throw new Error(`Failed to fetch file: ${res.status} ${res.statusText}`);\n }\n return res.arrayBuffer();\n });\n const base64Data = Buffer.from(fileData).toString('base64');\n const mimeType = attachment.content_type;\n\n const payload = {\n contents: [{\n parts: [\n { text: `Provide a comprehensive analysis of this image: ${attachment.name}. Describe in detail what you can see, including objects, people, text, colors, lighting, composition, style, and any other notable features or content.` },\n { inline_data: { mime_type: mimeType, data: base64Data } }\n ]\n }],\n generationConfig: {\n temperature: 0.7,\n maxOutputTokens: 2048,\n }\n };\n\n const response = await fetch(`https://generativelanguage.googleapis.com/v1beta/models/gemini-flash-latest:generateContent?key=${process.env.GEMINI_API_KEY}`, {\n method: 'POST',\n headers: { 'Content-Type': 'application/json', },\n body: JSON.stringify(payload),\n });\n\n if (!response.ok) {\n const errorText = await response.text().catch(() => \"Unknown error\");\n throw new APIConnectionError(\"Gemini API\", new Error(`${response.status} - ${errorText}`));\n }\n const result = await response.json();\n\n if (result.candidates?.[0]?.content?.parts?.[0]?.text) {\n return result.candidates[0].content.parts[0].text;\n } else {\n throw new FileProcessingError(attachment.name, \"image analysis\", new Error(\"No content returned from API\"));\n }\n } catch (error) {\n if (error instanceof APIConnectionError || error instanceof FileProcessingError) {\n throw error;\n }\n throw new FileProcessingError(attachment.name, \"image processing\", error);\n }\n}\n\nasync function processAudioWithWhisper(attachment: any): Promise {\n try {\n const fileData = await fetch(attachment.url).then(res => {\n if (!res.ok) {\n throw new Error(`Failed to fetch file: ${res.status} ${res.statusText}`);\n }\n return res.arrayBuffer();\n });\n if (fileData.byteLength > 25 * 1024 * 1024) {\n throw new FileSizeError(attachment.name, \"25MB\");\n }\n const formData = new FormData();\n const audioBlob = new Blob([fileData], { type: attachment.content_type });\n formData.append('file', audioBlob, attachment.name);\n formData.append('model', 'whisper-large-v3');\n formData.append('response_format', 'json');\n formData.append('language', 'auto');\n\n const response = await fetch(\"https://api.groq.com/openai/v1/audio/transcriptions\", {\n method: 'POST',\n headers: { 'Authorization': `Bearer ${process.env.GROQ_API_KEY}`, },\n body: formData,\n });\n\n if (!response.ok) {\n const errorText = await response.text().catch(() => \"Unknown error\");\n throw new APIConnectionError(\"Groq Whisper API\", new Error(`${response.status} - ${errorText}`));\n }\n const result = await response.json();\n\n if (result.text) {\n return `Audio transcription from ${attachment.name}:\\n\\n${result.text}`;\n } else {\n throw new FileProcessingError(attachment.name, \"audio transcription\", new Error(\"No transcription returned\"));\n }\n } catch (error) {\n if (error instanceof APIConnectionError || error instanceof FileProcessingError || error instanceof FileSizeError) {\n throw error;\n }\n throw new FileProcessingError(attachment.name, \"audio processing\", error);\n }\n}\n\nasync function processPDFWithGemini(attachment: any): Promise {\n try {\n const fileData = await fetch(attachment.url).then(res => {\n if (!res.ok) {\n throw new Error(`Failed to fetch file: ${res.status} ${res.statusText}`);\n }\n return res.arrayBuffer();\n });\n if (fileData.byteLength > 20 * 1024 * 1024) {\n throw new FileSizeError(attachment.name, \"20MB\");\n }\n const base64Data = Buffer.from(fileData).toString('base64');\n\n const payload = {\n contents: [{\n parts: [\n { text: `Extract and analyze all text content from this PDF document: ${attachment.name}. Provide a summary of the content, main topics, key information, and any structured data found.` },\n { inline_data: { mime_type: 'application/pdf', data: base64Data } }\n ]\n }],\n generationConfig: {\n temperature: 0.3,\n maxOutputTokens: 4096,\n }\n };\n\n const response = await fetch(`https://generativelanguage.googleapis.com/v1beta/models/gemini-flash-latest:generateContent?key=${process.env.GEMINI_API_KEY}`, {\n method: 'POST',\n headers: { 'Content-Type': 'application/json', },\n body: JSON.stringify(payload),\n });\n\n if (!response.ok) {\n const errorText = await response.text().catch(() => \"Unknown error\");\n throw new APIConnectionError(\"Gemini API\", new Error(`${response.status} - ${errorText}`));\n }\n const result = await response.json();\n\n if (result.candidates?.[0]?.content?.parts?.[0]?.text) {\n return result.candidates[0].content.parts[0].text;\n } else {\n throw new FileProcessingError(attachment.name, \"PDF analysis\", new Error(\"No content returned from API\"));\n }\n } catch (error) {\n if (error instanceof APIConnectionError || error instanceof FileProcessingError || error instanceof FileSizeError) {\n throw error;\n }\n throw new FileProcessingError(attachment.name, \"PDF processing\", error);\n }\n}","deleted","label",[103],"permalink","https://www.val.town/x/MadeiInGermany/Kimi-K2-0905-Exp/code/main.tsx","extension","actionData","errors"]