refactored llm, added gemini

package.json

@@ -1,6 +1,7 @@
 {
     "type": "module",
     "dependencies": {
+        "@google/generative-ai": "^0.2.1",
         "minecraft-data": "^3.46.2",
         "mineflayer": "^4.14.0",
         "mineflayer-armor-manager": "^2.0.1",

settings.json

@@ -3,5 +3,7 @@
     "host": "localhost",
     "port": 55916,
     "auth": "offline",
+
+    "model": "gemini-1.0-pro",
     "allow_insecure_coding": false
 }

src/agent/agent.js

@@ -3,7 +3,7 @@ import { Coder } from './coder.js';
 import { initModes } from './modes.js';
 import { Examples } from '../utils/examples.js';
 import { initBot } from '../utils/mcdata.js';
-import { sendRequest } from '../utils/gpt.js';
+import { sendRequest } from '../models/model.js';
 import { containsCommand, commandExists, executeCommand } from './commands/index.js';
 
 
@@ -102,7 +102,7 @@ export class Agent {
             let command_name = containsCommand(res);
 
             if (command_name) { // contains query or command
-                console.log('Command message:', res);
+                console.log(`""${res}""`)
                 if (!commandExists(command_name)) {
                     this.history.add('system', `Command ${command_name} does not exist. Use !newAction to perform custom actions.`);
                     console.log('Agent hallucinated command:', command_name)
@@ -110,8 +110,10 @@ export class Agent {
                 }
 
                 let pre_message = res.substring(0, res.indexOf(command_name)).trim();
-
-                this.cleanChat(`${pre_message}  *used ${command_name.substring(1)}*`);
+                let message = `*used ${command_name.substring(1)}*`;
+                if (pre_message.length > 0)
+                    message = `${pre_message}  ${message}`;
+                this.cleanChat(message);
                 let execute_res = await executeCommand(this, res);
 
                 console.log('Agent executed:', command_name, 'and got:', execute_res);

src/agent/coder.js

@@ -1,5 +1,5 @@
 import { writeFile, readFile, mkdirSync } from 'fs';
-import { sendRequest } from '../utils/gpt.js';
+import { sendRequest } from '../models/model.js';
 import { getSkillDocs } from './library/index.js';
 import { Examples } from '../utils/examples.js';
 

src/agent/history.js

@@ -1,6 +1,6 @@
 import { writeFileSync, readFileSync, mkdirSync } from 'fs';
 import { stringifyTurns } from '../utils/text.js';
-import { sendRequest } from '../utils/gpt.js';
+import { sendRequest } from '../models/model.js';
 import { getCommandDocs } from './commands/index.js';
 
 

src/models/gemini.js

@@ -0,0 +1,37 @@
+import { GoogleGenerativeAI } from '@google/generative-ai';
+import settings from '../settings.js';
+
+export class Gemini {
+    constructor() {
+        if (!process.env.GEMINI_API_KEY) {
+            console.error('Gemini API key missing! Make sure you set your GEMINI_API_KEY environment variable.');
+            process.exit(1);
+        }
+        this.genAI = new GoogleGenerativeAI(process.env.GEMINI_API_KEY);
+
+        this.model = this.genAI.getGenerativeModel({ model: settings.model });
+    }
+
+    async sendRequest(turns, systemMessage) {
+        const messages = [{'role': 'system', 'content': systemMessage}].concat(turns);
+        let prompt = "";
+        let role = "";
+        messages.forEach((message) => {
+            role = message.role;
+            if (role === 'assistant') role = 'model';
+            prompt += `${role}: ${message.content}\n`;
+        });
+        if (role !== "model") // if the last message was from the user/system, add a prompt for the model. otherwise, pretend we are extending the model's own message
+            prompt += "model: ";
+        console.log(prompt)
+        const result = await this.model.generateContent(prompt);
+        const response = await result.response;
+        return response.text();
+    }
+
+    async embed(text) {
+        const model = this.genAI.getGenerativeModel({ model: "embedding-001"});
+        const result = await model.embedContent(text);
+        return result.embedding;
+    }
+}

src/models/gpt.js

@@ -0,0 +1,66 @@
+import OpenAIApi from 'openai';
+import settings from '../settings.js';
+
+export class GPT {
+    constructor() {
+        let openAiConfig = null;
+        if (process.env.OPENAI_ORG_ID) {
+            openAiConfig = {
+                organization: process.env.OPENAI_ORG_ID,
+                apiKey: process.env.OPENAI_API_KEY,
+            };
+        } 
+        else if (process.env.OPENAI_API_KEY) {
+            openAiConfig = {
+                apiKey: process.env.OPENAI_API_KEY,
+            };
+        }
+        else {
+            console.error('OpenAI API key missing! Make sure you set OPENAI_API_KEY and OPENAI_ORG_ID (optional) environment variables.');
+            process.exit(1);
+        }
+
+        this.openai = new OpenAIApi(openAiConfig);
+    }
+
+    async sendRequest(turns, systemMessage, stop_seq='***') {
+
+        let messages = [{'role': 'system', 'content': systemMessage}].concat(turns);
+
+        let res = null;
+        try {
+            console.log('Awaiting openai api response...')
+            let completion = await this.openai.chat.completions.create({
+                model: settings.model,
+                messages: messages,
+                stop: stop_seq,
+            });
+            if (completion.choices[0].finish_reason == 'length')
+                throw new Error('Context length exceeded'); 
+            console.log('Received.')
+            res = completion.choices[0].message.content;
+        }
+        catch (err) {
+            if ((err.message == 'Context length exceeded' || err.code == 'context_length_exceeded') && turns.length > 1) {
+                console.log('Context length exceeded, trying again with shorter context.');
+                return await sendRequest(turns.slice(1), systemMessage, stop_seq);
+            } else {
+                console.log(err);
+                res = 'My brain disconnected, try again.';
+            }
+        }
+        return res;
+    }
+
+    async embed(text) {
+        const embedding = await this.openai.embeddings.create({
+            model: "text-embedding-ada-002",
+            input: text,
+            encoding_format: "float",
+        });
+        return embedding.data[0].embedding;
+    }
+}
+
+
+

src/models/model.js

@@ -0,0 +1,19 @@
+import { GPT } from './gpt.js';
+import { Gemini } from './gemini.js';
+import settings from '../settings.js';
+
+console.log('Initializing model...');
+let model = null;
+if (settings.model.includes('gemini')) {
+    model = new Gemini();
+} else {
+    model = new GPT();
+}
+
+export async function sendRequest(turns, systemMessage) {
+    return await model.sendRequest(turns, systemMessage);
+}
+
+export async function embed(text) {
+    return await model.embed(text);
+}

src/utils/examples.js

@@ -1,6 +1,7 @@
 import { readFileSync } from 'fs';
-import { embed, cosineSimilarity } from './gpt.js';
+import { cosineSimilarity } from './math.js';
 import { stringifyTurns } from './text.js';
+import { embed } from '../models/model.js';
 
 
 export class Examples {

src/utils/gpt.js

@@ -1,75 +0,0 @@
-import OpenAIApi from 'openai';
-
-
-let openAiConfig = null;
-if (process.env.OPENAI_ORG_ID) {
-    openAiConfig = {
-        organization: process.env.OPENAI_ORG_ID,
-        apiKey: process.env.OPENAI_API_KEY,
-    };
-} 
-else if (process.env.OPENAI_API_KEY) {
-    openAiConfig = {
-        apiKey: process.env.OPENAI_API_KEY,
-    };
-}
-else {
-    console.error('OpenAI API key missing! Make sure you set OPENAI_API_KEY and OPENAI_ORG_ID (optional) environment variables.');
-    process.exit(1);
-}
-
-const openai = new OpenAIApi(openAiConfig);
-
-
-export async function sendRequest(turns, systemMessage, stop_seq='***') {
-
-    let messages = [{'role': 'system', 'content': systemMessage}].concat(turns);
-
-    let res = null;
-    try {
-        console.log('Awaiting openai api response...')
-        let completion = await openai.chat.completions.create({
-            model: 'gpt-3.5-turbo',
-            messages: messages,
-            stop: stop_seq,
-        });
-        if (completion.choices[0].finish_reason == 'length')
-            throw new Error('Context length exceeded'); 
-        console.log('Received.')
-        res = completion.choices[0].message.content;
-    }
-    catch (err) {
-        if ((err.message == 'Context length exceeded' || err.code == 'context_length_exceeded') && turns.length > 1) {
-            console.log('Context length exceeded, trying again with shorter context.');
-            return await sendRequest(turns.slice(1), systemMessage, stop_seq);
-        } else {
-            console.log(err);
-            res = 'My brain disconnected, try again.';
-        }
-    }
-    return res;
-}
-
-
-export async function embed(text) {
-    const embedding = await openai.embeddings.create({
-        model: "text-embedding-ada-002",
-        input: text,
-        encoding_format: "float",
-    });
-    return embedding.data[0].embedding;
-}
-
-export function cosineSimilarity(a, b) {
-    let dotProduct = 0;
-    let magnitudeA = 0;
-    let magnitudeB = 0;
-    for (let i = 0; i < a.length; i++) {
-        dotProduct += a[i] * b[i];  // calculate dot product
-        magnitudeA += Math.pow(a[i], 2);  // calculate magnitude of a
-        magnitudeB += Math.pow(b[i], 2);  // calculate magnitude of b
-    }
-    magnitudeA = Math.sqrt(magnitudeA);
-    magnitudeB = Math.sqrt(magnitudeB);
-    return dotProduct / (magnitudeA * magnitudeB);  // calculate cosine similarity
-}

src/utils/math.js

@@ -0,0 +1,13 @@
+export function cosineSimilarity(a, b) {
+    let dotProduct = 0;
+    let magnitudeA = 0;
+    let magnitudeB = 0;
+    for (let i = 0; i < a.length; i++) {
+        dotProduct += a[i] * b[i];  // calculate dot product
+        magnitudeA += Math.pow(a[i], 2);  // calculate magnitude of a
+        magnitudeB += Math.pow(b[i], 2);  // calculate magnitude of b
+    }
+    magnitudeA = Math.sqrt(magnitudeA);
+    magnitudeB = Math.sqrt(magnitudeB);
+    return dotProduct / (magnitudeA * magnitudeB);  // calculate cosine similarity
+}