Skip to content
TheChatArchiveThe Chat Archive
gpt-4o-mini-2024-07-18· OpenAIAi2 WildChat-4.8M (AI2)

const axios = require('axios'); const { sendMessage } =…

  1. User

    User: const axios = require('axios'); const { sendMessage } = require('../handles/sendMessage');

    const conversationHistory = new Map(); const timeouts = new Map();

    const persona = "you're rick , you're an ai assistant, you're the best and the best, use emojies in u're answers, you speak just arabic";

    module.exports = { name: 'gpt4', description: 'Interact with Gemini API with images support', usage: 'gpt4 [your message]', author: 'coffee',

    async execute(senderId, args, pageAccessToken) { const prompt = args.join(' '); if (!prompt) return sendMessage(senderId, { text: "Usage: gpt4 " }, pageAccessToken);

    if (!conversationHistory.has(senderId)) {
      conversationHistory.set(senderId, []);
    }
    
    let chatHistory = conversationHistory.get(senderId);
    if (chatHistory.length === 0) {
      chatHistory.push(`Instruction: ${persona}`);
    }
    
    chatHistory.push(`User: ${prompt}`);
    
    if (chatHistory.length > 6) {
      chatHistory.shift();
    }
    
    const fullConversation = chatHistory.join("\");
    
    try {
      const { data } = await axios.get(`http://sgp1.hmvhostings.com:25721/gemini?question=${encodeURIComponent(fullConversation)}`);
    
      let responseText = data.answer ? data.answer : "لم أتمكن من فهم الإجابة.";
    
      chatHistory.push(`Bot: ${responseText}`);
    
      if (chatHistory.length > 6) {
        chatHistory.shift();
      }
    
      
      await sendMessage(senderId, { text: responseText }, pageAccessToken);
    
      
      if (data.web_images && data.web_images.length > 0) {
        for (let imageUrl of data.web_images) {
          await sendMessage(senderId, {
            attachment: {
              type: "image",
              payload: { url: imageUrl, is_reusable: true }
            }
          }, pageAccessToken);
        }
      }
    
    
      if (timeouts.has(senderId)) {
        clearTimeout(timeouts.get(senderId));
      }
    
      const timeout = setTimeout(() => {
        conversationHistory.delete(senderId);
        timeouts.delete(senderId);
      }, 10 * 60 * 1000);
    
      timeouts.set(senderId, timeout);
    
    } catch (error) {
      console.error("Error fetching data:", error);
      sendMessage(senderId, { text: 'حدث خطأ أثناء معالجة الطلب. حاول مرة أخرى لاحقًا.' }, pageAccessToken);
    }
    

    } }; User: const axios = require('axios'); const { sendMessage } = require('../handles/sendMessage');

    const conversationHistory = new Map(); const timeouts = new Map();

    const persona = "you're rick , you're an ai assistant, you're the best and the best, use emojies in u're answers, you speak just arabic";

    module.exports = { name: 'gpt4', description: 'Interact with Gemini API with images support', usage: 'gpt4 [your message]', author: 'coffee',

    async execute(senderId, args, pageAccessToken) { const prompt = args.join(' '); if (!prompt) return sendMessage(senderId, { text: "Usage: gpt4 " }, pageAccessToken);

    if (!conversationHistory.has(senderId)) {
      conversationHistory.set(senderId, []);
    }
    
    let chatHistory = conversationHistory.get(senderId);
    if (chatHistory.length === 0) {
      chatHistory.push(`Instruction: ${persona}`);
    }
    
    chatHistory.push(`User: ${prompt}`);
    
    if (chatHistory.length > 6) {
      chatHistory.shift();
    }
    
    const fullConversation = chatHistory.join("\");
    
    try {
      const { data } = await axios.get(`http://sgp1.hmvhostings.com:25721/gemini?question=${encodeURIComponent(fullConversation)}`);
    
      let responseText = data.answer ? data.answer : "لم أتمكن من فهم الإجابة.";
    
      chatHistory.push(`Bot: ${responseText}`);
    
      if (chatHistory.length > 6) {
        chatHistory.shift();
      }
    
      
      await sendMessage(senderId, { text: responseText }, pageAccessToken);
    
      
      if (data.web_images && data.web_images.length > 0) {
        for (let imageUrl of data.web_images) {
          await sendMessage(senderId, {
            attachment: {
              type: "image",
              payload: { url: imageUrl, is_reusable: true }
            }
          }, pageAccessToken);
        }
      }
    
    
      if (timeouts.has(senderId)) {
        clearTimeout(timeouts.get(senderId));
      }
    
      const timeout = setTimeout(() => {
        conversationHistory.delete(senderId);
        timeouts.delete(senderId);
      }, 10 * 60 * 1000);
    
      timeouts.set(senderId, timeout);
    
    } catch (error) {
      console.error("Error fetching data:", error);
      sendMessage(senderId, { text: 'حدث خطأ أثناء معالجة الطلب. حاول مرة أخرى لاحقًا.' }, pageAccessToken);
    }
    

    } }; Assistant:

  2. Assistant

    It looks like you've posted a JavaScript code snippet for a module that interacts with a Gemini API using Axios to send messages and handle conversations for a chatbot. Below are some comments on the code along with potential improvements:

    Overview of the Code

    1. Imports: It imports Axios for making HTTP requests and a sendMessage function for sending responses back to the user.

    2. Conversation Management:

      • The conversationHistory Map keeps track of the conversation history for each sender.
      • The timeouts Map manages session timeouts, allowing the chat context to persist for a duration of 10 minutes.
    3. Persona: A persona string defines the bot's character and response language. In this case, it's a playful character called Rick who responds in Arabic.

    4. Execution Logic:

      • The bot listens for a message and constructs a prompt based on user input.
      • It checks if there’s existing conversation history; if not, it initializes one.
      • The conversation history is updated, and when it surpasses 6 messages, the oldest message is discarded.
    5. API Interaction: The bot constructs a full prompt and sends it to the specified Gemini API, handling the response appropriately.

    6. Error Handling: There are catch blocks for logging errors and sending error messages back to the user.

    Improvements & Suggestions

    1. String Interpolation: There seems to be a syntax error in the line where the full conversation string is constructed:

      const fullConversation = chatHistory.join("\");
      

      It should be fixed as:

      const fullConversation = chatHistory.join("\n");
      
    2. Encapsulation of Persona Logic: Consider encapsulating the persona definition and chat history management into a separate module to make the code cleaner and more maintainable.

    3. Error Handling: Enhance error messages based on different error types (e.g., network errors vs. invalid responses) to provide better feedback.

    4. Timeout Configurability: Instead of hardcoding the timeout duration, consider making it a configurable parameter to allow for flexibility (e.g., through environment variables).

    5. Image Handling: When sending images, check if the sendMessage function can process messages in bulk to reduce the number of sending calls.

    6. Logging: Enhance your logging functionality for better debugging and monitoring of user interactions.

    By implementing these improvements, you'll make the bot more robust, maintainable, and user-friendly. If you have specific questions or requests about expanding on certain aspects or adding features, feel free to ask!