replace dall-e w/ gpt-image

Author: Decryptu

README.md

@@ -1,12 +1,15 @@
 # DecryptGPT
 
-DecryptGPT is an advanced Discord bot developed by Decrypt, utilizing the `discord.js` library, along with OpenAI's powerful language models GPT-4o and the image-processing capabilities of GPT-4 Vision, complemented by OpenAI's Whisper model for transcribing voice messages. This bot offers a comprehensive AI experience within Discord channels, capable of engaging in text conversations, responding to voice inputs, analyzing visual content, processing documents (pdf, csv...), and seamlessly switching between different GPT models using slash commands. Whether for advanced medical revisions, casual interactions, or exploring the capabilities of AI, DecryptGPT is designed to enrich the Discord chat experience.
+DecryptGPT is an advanced Discord bot developed by Decrypt, utilizing the `discord.js` library, along with OpenAI's powerful language models GPT-4o and the image-processing capabilities of GPT-4 Vision and GPT-Image-1, complemented by OpenAI's Whisper model for transcribing voice messages. This bot offers a comprehensive AI experience within Discord channels, capable of engaging in text conversations, responding to voice inputs, analyzing visual content, processing documents, generating and editing images, and seamlessly switching between different GPT models using slash commands. Whether for advanced medical revisions, casual interactions, or exploring the capabilities of AI, DecryptGPT is designed to enrich the Discord chat experience.
 
 ## Features
 
 - **Multiple GPT Modes**: Choose between GPT-4o or GPT-4 Vision for diverse interactions.
 - **Voice Message Understanding**: DecryptGPT can transcribe voice messages and respond to them, making interaction more seamless.
 - **Image Recognition**: With GPT-4 Vision, the bot can interpret images sent in the chat, adding a new dimension to AI conversations.
+- **Image Generation and Editing**: Using GPT-Image-1, the bot can:
+  - **Generate** completely new images based on text prompts.
+  - **Edit** existing uploaded images according to a provided text instruction.
 - **Document Processing**: The bot can analyze and extract text from various file formats:
   - PDF documents
   - Text files (.txt)
@@ -15,6 +18,16 @@ DecryptGPT is an advanced Discord bot developed by Decrypt, utilizing the `disco
 - **Customizable Configuration**: Tailor the bot's behavior through the `config.js` file, including setting the AI name, choosing the default GPT model, and more.
 - **Detailed Logging**: Enable detailed logging for debugging and monitoring interactions.
 
+## Slash Commands
+
+DecryptGPT supports several slash commands to interact with the AI:
+
+- `/chat` — Start a conversation with the selected GPT model.
+- `/generate` — Generate a new image from a text prompt using **GPT-Image-1**.
+- `/edit` — Upload an image and provide a prompt to modify it using **GPT-Image-1**.
+
+Console logs are included at every major step to ensure visibility on operations (e.g., request sent, response received, errors).
+
 ## Configuration
 
 Before running DecryptGPT, some configurations are required:
@@ -24,62 +37,60 @@ Before running DecryptGPT, some configurations are required:
   ```bash
   API_KEY=your_openai_api_key
   TOKEN=your_discord_bot_token
-  ```
+  CLIENT_ID=your_discord_app_client_id
 
-  Replace `your_openai_api_key` and `your_discord_bot_token` with your respective OpenAI API key and Discord bot token.
+Replace your_openai_api_key and your_discord_bot_token with your respective OpenAI API key and Discord bot token.
+ • channels.mjs: Edit channels.mjs to include the IDs of the Discord channels where the bot is allowed to operate. Channel IDs can be obtained by enabling Developer Mode in Discord and right-clicking on the channel to copy its ID.
+ • config.js: Customize the bot’s settings in config.js.
 
-- **channels.mjs**: Edit `channels.mjs` to include the IDs of the Discord channels where the bot is allowed to operate. Channel IDs can be obtained by enabling Developer Mode in Discord and right-clicking on the channel to copy its ID.
+Installation
 
-- **config.js**: Customize the bot's settings in `config.js`.
+ 1. Clone the repository:
 
-## Installation
+git clone <https://github.com/decryptu/DecryptGPT.git>
+ 
+ 2. Navigate to the project directory:
 
-1. Clone the repository:
+cd DecryptGPT
 
-   ```bash
-   git clone https://github.com/decryptu/DecryptGPT.git
-   ```
+ 3. Install dependencies:
 
-2. Navigate to the project directory:
+npm install
 
-   ```bash
-   cd DecryptGPT
-   ```
+ 4. Install additional packages for document processing:
 
-3. Install dependencies:
+npm install pdf.js-extract xlsx
 
-   ```bash
-   npm install
-   ```
+Running the Bot
+ • To run the bot:
 
-4. Install additional packages for document processing:
+npm start
 
-   ```bash
-   npm install pdf.js-extract xlsx
-   ```
+ • For development with hot reloads:
 
-## Running the Bot
+npm run dev
 
-- To run the bot, use:
+Example Usage
+ • Generate an image:
 
-  ```bash
-  npm start
-  ```
+/image prompt: "A futuristic city floating above the clouds"
 
-- For development, with hot reloads, use:
+The bot will reply with an original AI-generated image.
 
-  ```bash
-  npm run dev
-  ```
+ • Edit an uploaded image:
+
+/image-edit image: [Upload your image] prompt: "Make the sky sunset-themed with flying cars"
+
+The bot will modify the uploaded image according to the prompt.
 
-## Contributing
+Contributing
 
-Contributions to DecryptGPT are welcome! Please feel free to submit pull requests, create issues, or suggest new features.
+Contributions to DecryptGPT are welcome! Feel free to submit pull requests, create issues, or suggest new features.
 
-## License
+License
 
-This project is licensed under the [ISC License](LICENSE).
+This project is licensed under the ISC License.
 
----
+⸻
 
 Developed with ❤️ by Decrypt

channels.mjs

@@ -1,5 +1,5 @@
 const allowedChannels = [
-  "1185568440392617984", // decrypt
+  "1088435060170051714", // decrypt
   "1097824934153170984", // idkzp
   "1272543925286211606", // idksmp
 ];

commands/dalle.js

@@ -0,0 +1,106 @@
+import { EmbedBuilder, AttachmentBuilder } from "discord.js";
+import axios from "axios";
+import { toFile } from "openai";
+
+async function imageEdit(interaction, client) {
+  // Immediately defer the reply to avoid timeout issues
+  await interaction.deferReply();
+  
+  try {
+    // Log the interaction for debugging
+    console.log("Image edit command received:", {
+      user: interaction.user.tag,
+      options: interaction.options.data
+    });
+    
+    const description = interaction.options.getString("description");
+    const imageAttachment = interaction.options.getAttachment("image");
+    
+    // Validate inputs after deferring to ensure we can respond properly
+    if (!description) {
+      return await interaction.editReply({
+        content: "Please provide an image description."
+      });
+    }
+
+    if (!imageAttachment) {
+      return await interaction.editReply({
+        content: "Please provide a valid image attachment."
+      });
+    }
+
+    // Log the attachment details
+    console.log("Attachment details:", {
+      name: imageAttachment.name,
+      contentType: imageAttachment.contentType,
+      size: imageAttachment.size,
+      url: imageAttachment.url
+    });
+
+    if (!imageAttachment.contentType?.startsWith("image/")) {
+      return await interaction.editReply({
+        content: "The provided attachment is not an image. Please upload a PNG, JPEG, or WebP image."
+      });
+    }
+    
+    // Download the image from Discord
+    console.log("Downloading image from Discord...");
+    const response = await axios.get(imageAttachment.url, { responseType: 'arraybuffer' });
+    const imageBuffer = Buffer.from(response.data);
+    console.log(`Image downloaded, size: ${imageBuffer.length} bytes`);
+    
+    // Convert to proper format for OpenAI
+    console.log("Converting image for OpenAI...");
+    const imageFile = await toFile(imageBuffer, imageAttachment.name || "image.png", {
+      type: imageAttachment.contentType
+    });
+    console.log("Image converted successfully");
+
+    // Call the edit endpoint
+    console.log("Calling OpenAI edit API with prompt:", description);
+    const result = await client.openai.images.edit({
+      model: "gpt-image-1",
+      image: imageFile,
+      prompt: description,
+    });
+
+    console.log("OpenAI API response received:", result);
+
+    if (!result?.data?.[0]?.b64_json) {
+      throw new Error("No image data returned from API");
+    }
+
+    // Convert base64 to buffer for Discord attachment
+    const outputBuffer = Buffer.from(result.data[0].b64_json, "base64");
+    const attachment = new AttachmentBuilder(outputBuffer, { name: "edited-image.png" });
+
+    const embed = new EmbedBuilder()
+      .setTitle("Image edited successfully")
+      .setDescription(`Prompt: ${description}`)
+      .setImage("attachment://edited-image.png")
+      .setTimestamp();
+
+    await interaction.editReply({
+      content: null,
+      embeds: [embed],
+      files: [attachment]
+    });
+    
+    console.log("Successfully sent edited image response");
+  } catch (error) {
+    console.error("Error in imageEdit function:", error);
+    
+    // Check for specific API errors
+    const errorMessage = error.response?.data?.error?.message || error.message || "Unknown error";
+    console.error("Detailed error:", errorMessage);
+    
+    // Always use editReply since we deferred at the beginning
+    await interaction.editReply({
+      content: `Failed to edit the image. Error: ${errorMessage}`,
+      embeds: [],
+      files: []
+    });
+  }
+}
+
+export default imageEdit;

commands/image-edit.js

@@ -0,0 +1,53 @@
+import { EmbedBuilder, AttachmentBuilder, MessageFlags } from "discord.js";
+
+async function imageGpt(interaction, client) {
+  const description = interaction.options.getString("description");
+  if (!description) {
+    await interaction.reply({
+      content: "Please provide an image description.",
+      flags: MessageFlags.Ephemeral
+    });
+    return;
+  }
+
+  try {
+    // Acknowledge the interaction immediately to prevent timeouts
+    await interaction.deferReply();
+    
+    const imageBase64 = await client.createGptImage(description);
+
+    if (!imageBase64) {
+      throw new Error("No image data returned");
+    }
+
+    // Convert base64 to buffer for Discord attachment
+    const buffer = Buffer.from(imageBase64, "base64");
+    const attachment = new AttachmentBuilder(buffer, { name: "generated-image.png" });
+
+    const embed = new EmbedBuilder()
+      .setTitle("Votre image")
+      .setImage("attachment://generated-image.png");
+
+    await interaction.editReply({
+      embeds: [embed],
+      files: [attachment]
+    });
+  } catch (error) {
+    console.error("Error in imageGpt function:", error);
+    
+    // Check if the interaction has already been replied to
+    if (interaction.deferred || interaction.replied) {
+      await interaction.editReply({
+        content: "Échec de la génération d'une image. Veuillez réessayer ultérieurement.",
+        files: []
+      });
+    } else {
+      await interaction.reply({
+        content: "Échec de la génération d'une image. Veuillez réessayer ultérieurement.",
+        flags: MessageFlags.Ephemeral
+      });
+    }
+  }
+}
+
+export default imageGpt;