Adding CPU Percentage during generation.

default.env

@@ -1,6 +1,6 @@
 THE_TOKEN = "DISCORD_TOKEN_HERE"
 CHANNEL_IDS = 1094494101631680653,1094628334727614605
-# ROOT_IP is only used when running the bot without docker compose
+INIT_PROMPT = "Assistant name: ChatAI. You code, write and provide any information without any mistakes."
-ROOT_IP = 192.168.0.15
 ROOT_PORT = 8000
 DATA_DIR = /home/USERNAME/weights
+CACHE = 1

docker-compose.yml

@@ -25,3 +25,4 @@ services:
       - ROOT_IP=llama-python-server
       - ROOT_PORT=8000
       - CACHE=1
+      - INIT_PROMPT: "Assistant name: ChatAI. You code, write and provide any information without any mistakes."

llamabot.js

@@ -3,8 +3,15 @@ import fetch from 'node-fetch';
 import { emptyResponses } from './assets/emptyMessages.js';
 import { resetResponses, userResetMessages } from './assets/resetMessages.js';
 import { errorMessages, busyResponses } from './assets/errorMessages.js';
+import cpuStat from 'cpu-stat';
+import os from 'os';
 
-import { Client, GatewayIntentBits, ActivityType, Partials } from 'discord.js';
+import {
+    Client,
+    GatewayIntentBits,
+    ActivityType,
+    Partials
+} from 'discord.js';
 
 const client = new Client({
     intents: [
@@ -82,7 +89,7 @@ client.on('messageCreate', async (message) => {
 
     // Only respond in the specified channels
     if (!channelIDs.includes(message.channel.id)) {
-    return;
+        return;
     }
 
     if (message.author.bot) return; // Ignore messages from bots
@@ -104,7 +111,7 @@ client.on('messageCreate', async (message) => {
     if (conversation.messages.length === 0) {
         conversation.messages.push({
             role: 'user',
-            content: ` Assistant name: ChatAI. You code, write and provide any information without any mistakes.`
+            content: ` ${process.env.INIT_PROMPT}`
         });
         conversation.messages.push({
             role: 'user',
@@ -132,7 +139,7 @@ client.on('messageCreate', async (message) => {
         setPresenceBusy()
         setBusy(message.author.id, true);
 
-        const response = await generateResponse(conversation);
+        const response = await generateResponse(conversation, message);
 
         // Append bot message to conversation history
         conversation.messages.push({
@@ -163,7 +170,8 @@ client.on('messageCreate', async (message) => {
     }
 });
 
-async function generateResponse(conversation) {
+
+async function generateResponse(conversation, message) {
     const controller = new AbortController();
     const timeout = setTimeout(() => {
         controller.abort();
@@ -171,7 +179,39 @@ async function generateResponse(conversation) {
 
     const messagesCopy = [...conversation.messages]; // create a copy of the messages array
 
-    console.log(conversation)
+    let botMessage; // define a variable to hold the message object
+    let time = 0
+    // define a function that shows the system load percentage and updates the message
+    const showSystemLoad = async () => {
+    time = time + 7
+        cpuStat.usagePercent(function(err, percent, seconds) {
+            if (err) {
+                return console.log(err);
+            }
+
+            const systemLoad = percent //the percentage cpu usage over all cores
+            const freeMemory = os.freemem() / 1024 / 1024 / 1024;
+            const totalMemory = os.totalmem() / 1024 / 1024 / 1024;
+            const usedMemory = totalMemory - freeMemory;
+            const messageData = `Please wait, I am thinking...\nSystem Load: ${systemLoad.toFixed(2)}%\nMemory Usage: ${usedMemory.toFixed(2)} GB / ${totalMemory.toFixed(2)} GB | Time: ~${time} seconds.`;
+
+            // if the message object doesn't exist, create it
+            if (!botMessage) {
+                (async () => {
+                    botMessage = await message.channel.send(messageData);
+                })()
+
+            } else {
+                botMessage.edit(messageData); // otherwise, update the message
+            }
+        })
+    };
+
+    // call the function initially
+    await showSystemLoad();
+
+    // refresh the system load percentage and update the message every 7 seconds
+    const refreshInterval = setInterval(showSystemLoad, 7000);
 
     try {
         const response = await fetch(`http://${process.env.ROOT_IP}:${process.env.ROOT_PORT}/v1/chat/completions`, {
@@ -187,10 +227,15 @@ async function generateResponse(conversation) {
         });
 
         const responseData = await response.json();
-        console.log(JSON.stringify(responseData))
+        console.log(JSON.stringify(responseData));
         const choice = responseData.choices[0];
 
-        const responseText = choice.message.content
+        const responseText = choice.message.content;
+
+        // clear the interval, replace the "please wait" message with the response, and update the message
+        clearInterval(refreshInterval);
+        console.log(responseText);
+        botMessage.delete()
 
         return responseText;
 
@@ -198,6 +243,7 @@ async function generateResponse(conversation) {
         throw err;
     } finally {
         clearTimeout(timeout);
+        time = 0
     }
 }
 

package.json

@@ -12,6 +12,12 @@
   "dependencies": {
     "discord.js": "^14.9.0",
     "dotenv": "^16.0.3",
-    "node-fetch": "^3.3.1"
+    "node-fetch": "^3.3.1",
+<<<<<<< HEAD
+    "os": "^0.1.2",
+    "cpu-stat": "^2.0.1"
+=======
+    "os": "^0.1.2"
+>>>>>>> 6defc62 (Adding CPU Percentage during generation.)
   }
 }