halbot 1993.2.65 → 1993.2.66

package/index.mjs

@@ -19,31 +19,30 @@ const init = async (options = {}) => {
     }
     // use openai embedding, dall-e, tts if openai is enabled
     if (options.openaiApiKey) {
-        const apiKey = { apiKey: options.openaiApiKey };
+        const apiKey = { apiKey: options.openaiApiKey, provider: 'OPENAI' };
         await alan.init({
-            provider: 'OPENAI', model: options.openaiModel || '*',
-            ...apiKey, priority: options.openaiPriority, ...options
+            ...apiKey, model: options.openaiModel || '*',
+            priority: options.openaiPriority, ...options
         });
         await image.init(apiKey);
-        await speech.init({ ...apiKey, provider: 'OPENAI', ...speechOptions });
+        await speech.init({ ...apiKey, ...speechOptions });
         _speech.tts = speech.tts;
     }
     // use gemini embedding if gemini is enabled and chatgpt is not enabled
     // use google tts if google api key is ready
     if (options.googleApiKey) {
-        const apiKey = { apiKey: options.googleApiKey };
+        const apiKey = { apiKey: options.googleApiKey, provider: 'GOOGLE' };
         await alan.init({
-            provider: 'GEMINI', model: options.geminiModel || '*',
-            ...apiKey, priority: options.geminiPriority, ...options
+            ...apiKey, provider: 'GEMINI', model: options.geminiModel || '*',
+            priority: options.geminiPriority, ...options
         });
+        await image.init({ ...apiKey, provider: 'GEMINI' });
         if (!_speech.tts) {
-            await speech.init({
-                provider: 'GOOGLE', ...apiKey, ...speechOptions,
-            });
+            await speech.init({ ...apiKey, ...speechOptions });
             _speech.tts = speech.tts;
         }
         options.googleCx && await web.initSearch({
-            provider: 'GOOGLE', ...apiKey, cx: options.googleCx
+            ...apiKey, cx: options.googleCx,
         });
     }
     if (options.anthropicApiKey) {
@@ -92,10 +91,13 @@ const init = async (options = {}) => {
         });
     }
     const { ais } = await alan.initChat({ sessions: options?.storage });
+    const cmds = options?.cmds || [];
     // config multimodal engines
-    const supportedMimeTypes = new Set(Object.values(ais).map(
-        x => x.model
-    ).map(x => [
+    const supportedMimeTypes = new Set(Object.values(ais).map(x => {
+        // init instant ai selection
+        cmds.push(bot.newCommand(`ai_${x.id}`, `${x.name}: ${x.features}`));
+        return x.model;
+    }).map(x => [
         ...x.supportedMimeTypes || [], ...x.supportedAudioTypes || [],
     ]).flat().map(x => x.toLowerCase()));
     // init bot
@@ -104,7 +106,7 @@ const init = async (options = {}) => {
         auth: options?.auth,
         botToken: options?.telegramToken,
         chatType: options?.chatType,
-        cmds: options?.cmds,
+        cmds,
         database: options?.storage?.client && options?.storage,
         embedding: ais.find(x => x.embedding)?.embedding,
         supportedMimeTypes,

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "halbot",
   "description": "Just another `ChatGPT` / `Gemini` / `Ollama` Telegram bob, which is simple design, easy to use, extendable and fun.",
-  "version": "1993.2.65",
+  "version": "1993.2.66",
   "private": false,
   "homepage": "https://github.com/Leask/halbot",
   "type": "module",
@@ -36,7 +36,7 @@
     "@google-cloud/speech": "^7.0.1",
     "@google-cloud/text-to-speech": "^6.0.1",
     "@google-cloud/vision": "^5.1.0",
-    "@google/genai": "^0.4.0",
+    "@google/genai": "^0.6.0",
     "@mozilla/readability": "^0.6.0",
     "fluent-ffmpeg": "^2.1.3",
     "ioredis": "^5.6.0",
@@ -51,7 +51,7 @@
     "pgvector": "^0.2.0",
     "telegraf": "^4.16.3",
     "tesseract.js": "^6.0.0",
-    "utilitas": "^1999.1.29",
+    "utilitas": "^1999.1.37",
     "youtube-transcript": "^1.2.1"
   }
 }

package/skills/40_dream.mjs

@@ -4,13 +4,20 @@ const action = async (ctx, next) => {
     if (!ctx.cmd.args) {
         return await ctx.ok('Please input your prompt.');
     }
-    const objMsg = (await ctx.ok('💭'))[0];
-    const images = await ctx._.image.generate(ctx.cmd.args, { expected: 'URL' });
+    let [objMsg, tts, images] = [(await ctx.ok('💭'))[0], null, null];
+    try {
+        images = await ctx._.image.generate(ctx.cmd.args, { expected: 'URL' });
+    } catch (err) {
+        return await ctx.er(err.message || 'Error generating image.',
+            { lastMessageId: objMsg.message_id });
+    }
     await ctx.deleteMessage(objMsg.message_id);
-    for (let image of images) {
-        await ctx.image(image.url, { caption: image.revised_prompt });
-        await ctx.shouldSpeech(image.revised_prompt)
+    for (let image of images || []) {
+        tts = image.tts || '';
+        await ctx.image(image.data, { caption: image.caption || '' });
+        await ctx.timeout();
     }
+    await ctx.shouldSpeech(tts);
 };
 
 export const { name, run, priority, func, cmds, help } = {
@@ -19,10 +26,16 @@ export const { name, run, priority, func, cmds, help } = {
     priority: 40,
     func: action,
     help: bot.lines([
-        'Use DALL-E to generate images.',
-        'Example: /dream a cat',
+        '¶ Use Google `Imagen` (default) or OpenAI `DALL-E` to generate images.',
+        'Example 1: /dream a cat in a rocket',
+        '¶ Use `Imagen` to generate images.',
+        'Example 2: /imagen a cat in a car',
+        '¶ Use `DALL-E` to generate images.',
+        'Example: /dalle a cat on a bike',
     ]),
     cmds: {
-        dream: 'Use DALL-E to generate images: /dream `PROMPT`',
+        dream: 'Generate images with default model: /dream `PROMPT`',
+        imagen: 'Generate images with `Imagen`: /imagen `PROMPT`',
+        dalle: 'Generate images with `DALL-E`: /dalle `PROMPT`',
     },
 };