我正在尝试使用 chatgpt 作为我的 magento 2 网站的聊天机器人,并且我想将产品数据传递给它。为此,我收集了所有产品并将它们存储在一个 json 文件中,然后读取该文件以将数据嵌入到系统角色的 systemrolecontent 中。然而,我面临的问题是 json 文件相当大。
{ "bot_response": "error: chatbot error: unexpected api response structure: { "error": { "message": "request too large for gpt-4o on tokens per min (tpm): limit 30000, requested 501140. the input or output tokens must be reduced in order to run successfully. visit platform.openai./account/rate-limits to learn more.", "type": "tokens", "param": null, "code": "rate_limit_exceeded" }}"}
我注意到需要在 api 配置中添加一个功能,该功能允许您运行查询来根据名称或描述中与用户消息中的关键字匹配的关键字来选择产品。挑战在于用户最初可能不知道产品的名称;他们来到聊天机器人是为了发现他们。我该如何解决这个问题?
这是我现在正在使用的代码:
<?phpnamespace MetaCaresChatbotModel;use MagentoFrameworkAppObjectManager;class ChatBot{ private $authorization; private $endpoint; private $conversationHistory = []; private $productsFile; private $fetchingDateFile; private $didFetchProducts = false; public function __construct() { $this->authorization = ‘sk-proj-‘; $this->endpoint = ‘api.openai./v1/chat/pletions’; $this->productsFile = __DIR__ . ‘/products.json’; $this->fetchingDateFile = __DIR__ . ‘/fetching_date.json’; $currentTime = time(); $timeDifferenceSeconds = 24 * 3600; if (!file_exists($this->fetchingDateFile)) {file_put_contents($this->fetchingDateFile, json_encode([‘last_fetch_time’ => 0])); } $fetchingData = json_decode(file_get_contents($this->fetchingDateFile), true); $lastFetchTime = $fetchingData[‘last_fetch_time’] ?? 0; if ($currentTime – $lastFetchTime > $timeDifferenceSeconds) {$products = $this->fetchProductsUsingModel();$productsJson = json_encode($products);file_put_contents($this->productsFile, $productsJson);$fetchingData[‘last_fetch_time’] = $currentTime;file_put_contents($this->fetchingDateFile, json_encode($fetchingData));$this->didFetchProducts = true; } $jsonSampleData = file_get_contents($this->productsFile); $systemRoleContent = <<<EOTNom: Meta Cares BotDescription BOT Meta Cares répond aux questions sur les produits du site et fournit des conseils santé fiables. Tu aides les clients de Meta Cares à faire des choix éclairés tout en offrant un acpagnement personnalisé, sécurisé et adapté à leurs besoins.catalogue Meta Cares{$jsonSampleData}Liste des Sites Référencés :- PubMed : [pubmed.ncbi.nlm.nih.gov/](pubmed.ncbi.nlm.nih.gov/)- ScienceDirect : [www.sciencedirect./](www.sciencedirect./)—- Génération d’images DALL·E : DésactivéeEOT; $this->conversationHistory[] = [‘role’ => ‘system’,’content’ => $systemRoleContent ]; if (session_status() == PHP_SESSION_NONE) {session_start(); } if (isset($_SESSION[‘chat_history’])) {$this->conversationHistory = $_SESSION[‘chat_history’]; } } public function fetchProductsUsingModel(): array { return $products; } private function getCategoryNames(array $categoryIds): array { return $categoryNames; } public function sendMessage(string $message): array { try {$this->conversationHistory[] = [ ‘role’ => ‘user’, ‘content’ => $message];$data = [ ‘model’ => ‘gpt-4o’, ‘messages’ => array_map(function ($msg) { return [‘role’ => $msg[‘role’] === ‘bot’ ? ‘assistant’ : $msg[‘role’],’content’ => $msg[‘content’] ]; }, $this->conversationHistory)];$response = $this->makeApiRequest($data);$arrResult = json_decode($response, true);if (json_last_error() !== JSON_ERROR_NONE) { throw new Exception(‘Invalid API response format’);}if (!isset($arrResult[‘choices’]) || !isset($arrResult[‘choices’][0][‘message’][‘content’])) { throw new Exception(‘Unexpected API response structure: ‘ . $response);}$assistantResponse = $arrResult[‘choices’][0][‘message’][‘content’];$this->conversationHistory[] = [ ‘role’ => ‘bot’, ‘content’ => $assistantResponse];$_SESSION[‘chat_history’] = $this->conversationHistory;return [ "conversationHistory" => $_SESSION[‘chat_history’], ‘didFetchProducts’ => $this->didFetchProducts, ‘response’ => $assistantResponse,]; } catch (Exception $e) {throw new Exception(‘ChatBot Error: ‘ . $e->getMessage()); } } private function makeApiRequest(array $data): string { $ch = curl_init(); curl_setopt_array($ch, [CURLOPT_URL => $this->endpoint,CURLOPT_POST => true,CURLOPT_POSTFIELDS => json_encode($data),CURLOPT_HTTPHEADER => [ ‘Content-Type: application/json’, ‘Authorization: Bearer ‘ . $this->authorization,],CURLOPT_RETURNTRANSFER => true,CURLOPT_SSL_VERIFYPEER => false,CURLOPT_SSL_VERIFYHOST => 0 ]); $response = curl_exec($ch); if (curl_errno($ch)) {$error = curl_error($ch);curl_close($ch);throw new Exception(‘API request failed: ‘ . $error); } curl_close($ch); return $response; }}
以上就是如何优化大型 JSON 文件以与 ChatGPT API 一起使用?的详细内容,更多请关注范的资源库其它相关文章!
转载请注明:范的资源库 » 如何优化大型JSON文件以与ChatGPTAPI一起使用?