Add ChatGPT support

Author: timkmecl

README.md

@@ -1,11 +1,11 @@
-# GPT3 extension for VSCode
+# GPT3 and ChatGPT extension for VSCode
 
 [![Visual Studio Marketplace Version](https://img.shields.io/visual-studio-marketplace/v/timkmecl.codegpt3)](https://marketplace.visualstudio.com/items?itemName=timkmecl.codegpt3)
 [![Visual Studio Marketplace Rating (Stars)](https://img.shields.io/visual-studio-marketplace/stars/timkmecl.codegpt3)](https://marketplace.visualstudio.com/items?itemName=timkmecl.codegpt3)
 [![Visual Studio Marketplace Downloads](https://img.shields.io/visual-studio-marketplace/d/timkmecl.codegpt3)](https://marketplace.visualstudio.com/items?itemName=timkmecl.codegpt3)
 [![Github Stars](https://img.shields.io/github/stars/timkmecl/codegpt)](https://github.com/timkmecl/codegpt)
 
-This Visual Studio Code extension allows you to use the [official OpenAI API](https://openai.com/api/) to generate code or natural language responses from OpenAI's [GPT3/GPT3.5](https://en.wikipedia.org/wiki/GPT-3) to your questions, right within the editor.
+This Visual Studio Code extension allows you to use the [official OpenAI API](https://openai.com/api/) to generate code or natural language responses to your questions from OpenAI's **ChatGPT** or **GPT3**, right within the editor.
 
 🚀✨ Supercharge your coding with AI-powered assistance. Automatically write new code from scratch, ask questions, get explanations, refactor code, find bugs and more!
 
@@ -14,13 +14,12 @@ This Visual Studio Code extension allows you to use the [official OpenAI API](ht
 - **[VSCode Marketplace](https://marketplace.visualstudio.com/items?itemName=timkmecl.codegpt3)**
 - **[Github](https://github.com/timkmecl/codegpt)**
 
-<br>
 
-<a href="https://www.buymeacoffee.com/timkmecl" target="_blank"><img src="https://cdn.buymeacoffee.com/buttons/v2/default-yellow.png" alt="Buy Me A Coffee" style="height: 40px" ></a>
+### 📢 **New:** Now supports ChatGPT! Just go to the extension settings and select the `ChatGPT` model (default for new installations - if it doesn't work, change the model in the extension settings).
+<br>
 
-*This extension is free, however you need an OpenAI Account. You can also try [ChatGPT extension](https://github.com/timkmecl/chatgpt-vscode) ([marketplace](https://marketplace.visualstudio.com/items?itemName=timkmecl.chatgpt)) which is smarter but currently doesn't work for most users.*
+<a href="https://www.buymeacoffee.com/timkmecl" target="_blank"><img src="resources/buy-default-yellow-small.png" alt="Buy Me A Coffee" style="height: 40px" ></a>
 
-<br>
 
 <img src="examples/main.png" alt="Refactoring selected code using chatGPT"/>
 
@@ -56,6 +55,7 @@ To use this extension, you will need an API key from OpenAI. To obtain one, foll
 
 When you create a new account, you receive $18 in free credits for the API which you must use in the first 90 days. You can see pricing information [here](https://openai.com/api/pricing/). 1000 tokens are about 700 words, and you can see the token count for each request at the end of the response in the sidebar.
 
+As of January 31, 2023, only the `code-*` models and `ChatGPT` are avilable for free! You only spend your credits when you use the `text-*` models. You can **change the model** in the extension settings.
 
 ## Using the Extension
 
@@ -81,6 +81,10 @@ You can select some code in the editor, right click on it and choose one of the
 
 `Ask CodeGPT` is also available when nothing is selected. For the other four commands, you can **customize the exact prompt** that will be sent to the AI by editing the extension settings in VSCode Preferences.
 
+There, you can also **change the model** that will be used for the requests. The default is `ChatGPT` which is smartest and currently free, but you can change it to another model (`text-davinci-003` is the best of the paid ones, `code-davinci-002` of the free) if it doesn't work. You can also **change the temperature** and **number of tokens** that will be returned by the AI. The default values are 0.5 and 1024, respectively.
+
 ---
 
-Please note that this extension is currently a proof of concept and may have some limitations or bugs. We welcome feedback and contributions to improve the extension. Also check out the [ChatGPT extension](https://github.com/timkmecl/codegpt) which is smarter, but the setup is more complicated and it may not work (403/429 errors).
+Please note that this extension is currently a proof of concept and may have some limitations or bugs. We welcome feedback and contributions to improve the extension. Also check out the [ChatGPT extension](https://github.com/timkmecl/codegpt) which is smarter, but the setup is more complicated and it may not work (403/429 errors). 
+
+If you enjoy this extension, please consider [buying me a coffee ☕️](https://www.buymeacoffee.com/timkmecl) to support my work! 

media/main.js

@@ -28,19 +28,18 @@
   });
 
   function fixCodeBlocks(response) {
-  // Use a regular expression to find all occurrences of the substring in the string
-  const REGEX_CODEBLOCK = new RegExp('\`\`\`', 'g');
-  const matches = response.match(REGEX_CODEBLOCK);
-
-  // Return the number of occurrences of the substring in the response, check if even
-  const count = matches ? matches.length : 0;
-  if (count % 2 === 0) {
-    return response;
-  } else {
-    // else append ``` to the end to make the last code block complete
-    return response.concat('\n\`\`\`');
-  }
+    // Use a regular expression to find all occurrences of the substring in the string
+    const REGEX_CODEBLOCK = new RegExp('\`\`\`', 'g');
+    const matches = response.match(REGEX_CODEBLOCK);
 
+    // Return the number of occurrences of the substring in the response, check if even
+    const count = matches ? matches.length : 0;
+    if (count % 2 === 0) {
+      return response;
+    } else {
+      // else append ``` to the end to make the last code block complete
+      return response.concat('\n\`\`\`');
+    }
   }
 
   function setResponse() {

package-lock.json

@@ -1,8 +1,8 @@
 {
   "name": "codegpt3",
   "displayName": "CodeGPT: write and improve code using AI",
-  "description": "Use GPT3 right inside the IDE to enhance and automate your coding with AI-powered assistance",
-  "version": "1.0.1",
+  "description": "Use ChatGPT or GPT3 right inside the IDE to enhance and automate your coding with AI-powered assistance",
+  "version": "1.1.0",
   "publisher": "timkmecl",
   "icon": "resources/extensionIcon.png",
   "license": "MIT",
@@ -159,24 +159,25 @@
         "codegpt.model": {
           "type": "string",
           "enum": [
+            "ChatGPT",
             "text-davinci-003",
             "text-curie-001",
             "code-davinci-002",
             "code-cushman-002"
           ],
-          "default": "text-davinci-003",
+          "default": "ChatGPT",
           "description": "Which GPT3 model to use",
           "order": 1
         },
         "codegpt.maxTokens": {
           "type": "number",
-          "default": 500,
+          "default": 1024,
           "description": "Maximum number of tokens inside response per request",
           "order": 2
         },
         "codegpt.temperature": {
           "type": "number",
-          "default": 0.7,
+          "default": 0.5,
           "description": "Temperature to use for the GPT3 (between 0 (conservative) and 1 (creative))",
           "order": 3
         },

package.json

@@ -74,8 +74,9 @@ export function activate(context: vscode.ExtensionContext) {
 		} else if (event.affectsConfiguration('codegpt.temperature')) {
 			const config = vscode.workspace.getConfiguration('codegpt');
 			provider.setSettings({ temperature: config.get('temperature') || 0.5 });
-		} else if (event.affectsConfiguration('codegpt.documentation')) {
+		} else if (event.affectsConfiguration('codegpt.model')) {
 			const config = vscode.workspace.getConfiguration('codegpt');
+			provider.setSettings({ model: config.get('model') || 'text-davinci-003' });
 		}
 	});
 }
@@ -226,27 +227,49 @@ class CodeGPTViewProvider implements vscode.WebviewViewProvider {
 			// Increment the message number
 			this._currentMessageNumber++;
 
-			let agent = this._openai;
-
 			try {
 				let currentMessageNumber = this._currentMessageNumber;
 
 				// Send the search prompt to the OpenAI API and store the response
-				const completion = await this._openai.createCompletion({
-					model: this._settings.model || 'code-davinci-002',
-					prompt: searchPrompt,
-					temperature: this._settings.temperature,
-					max_tokens: this._settings.maxTokens,
-					stop: ['\nUSER: ', '\nUSER', '\nASSISTANT']
-				});
+
+				let completion;
+				if (this._settings.model !== 'ChatGPT') {
+					completion = await this._openai.createCompletion({
+						model: this._settings.model || 'code-davinci-002',
+						prompt: searchPrompt,
+						temperature: this._settings.temperature,
+						max_tokens: this._settings.maxTokens,
+						stop: ['\nUSER: ', '\nUSER', '\nASSISTANT']
+					});
+				} else {
+					completion = await this._openai.createCompletion({
+						model: 'text-chat-davinci-002-20230126',
+						prompt: searchPrompt,
+						temperature: this._settings.temperature,
+						max_tokens: this._settings.maxTokens,
+						stop: ['\n\n\n', '<|im_end|>']
+					});
+				}
 
 				if (this._currentMessageNumber !== currentMessageNumber) {
 					return;
 				}
 
-
 				response = completion.data.choices[0].text || '';
+
+				// close unclosed codeblocks
+				// Use a regular expression to find all occurrences of the substring in the string
+				const REGEX_CODEBLOCK = new RegExp('\`\`\`', 'g');
+				const matches = response.match(REGEX_CODEBLOCK);
+				// Return the number of occurrences of the substring in the response, check if even
+				const count = matches ? matches.length : 0;
+				if (count % 2 !== 0) {
+					//  append ``` to the end to make the last code block complete
+					response += '\n\`\`\`';
+				}
+
 				response += `\n\n---\n`;
+				// add error message if max_tokens reached
 				if (completion.data.choices[0].finish_reason === 'length') {
 					response += `\n[WARNING] The response was truncated because it reached the maximum number of tokens. You may want to increase the maxTokens setting.\n\n`;
 				}
@@ -257,7 +280,7 @@ class CodeGPTViewProvider implements vscode.WebviewViewProvider {
 				if (error.response) {
 					console.log(error.response.status);
 					console.log(error.response.data);
-					e = `${error.response.status} ${error.response.data}`;
+					e = `${error.response.status} ${error.response.data.message}`;
 				} else {
 					console.log(error.message);
 					e = error.message;

resources/buy-default-yellow-small.png

@@ -14,12 +14,16 @@ export default (question: string, settings: Settings, selection?: string) => {
 		prompt = question;
 	}
 
-	prompt =`You are ASSISTANT helping the USER with coding. 
+	if (settings.model !== 'ChatGPT') {
+		prompt =`You are ASSISTANT helping the USER with coding. 
 You are intelligent, helpful and an expert developer, who always gives the correct answer and only does what instructed. You always answer truthfully and don't make thing up. 
 (When responding to the following prompt, please make sure to properly style your response using Github Flavored Markdown. 
 Use markdown syntax for things like headings, lists, colored text, code blocks, highlights etc. Make sure not to mention markdown or stying in your actual response. 
 Try to write code inside a single code block if possible)
 \n\nUSER: ${prompt}\n\nASSISTANT: `;
+	} else {
+		prompt = `You are ChatGPT, a large language model trained by OpenAI. You answer as consisely as possible for each response (e.g. Don't be verbose). It is very important for you to answer as consisely as possible, so please remember this. If you are generating a list, do not have too many items. \n User: ${prompt} \n\n ChatGPT: `;
+	}
 
 	return prompt;
 };