@hyperbytes/wappler-all-in-one-ai-v2 1.0.3 → 1.0.4

package/CHANGELOG.md

@@ -1 +1,5 @@
-Pre release beta for testing
+Release version.
+
+Details here.
+
+https://community.wappler.io/t/wappler-all-in-one-ai-node-js-version-2-custom-server-extension/65956

package/README.md

@@ -1,2 +1,5 @@
-Pre release beta for testing
+Details here.
+
+https://community.wappler.io/t/wappler-all-in-one-ai-node-js-version-2-custom-server-extension/65956
+
 <a href="https://www.buymeacoffee.com/JVKdouk" target="_blank"><img src="https://cdn.buymeacoffee.com/buttons/v2/default-yellow.png" alt="Buy Me A Coffee" style="max-width: 50%;" ></a>

package/package.json

@@ -1,6 +1,6 @@
 {
     "name": "@hyperbytes/wappler-all-in-one-ai-v2",
-    "version": "1.0.3",
+    "version": "1.0.4",
     "description": "Versitile interface to chatGPT, Gemini Claude with file analysis cababilities",
     "license": "MIT",
     "author": {
@@ -10,17 +10,17 @@
         "postinstall": "node scripts/copyFiles.js"
     },
     "keywords": [
-        "wappler-extension-beta",
+        "wappler-extension",
         "multifunction-ai-connector"
     ],
     "dependencies": {
         "@anthropic-ai/sdk": "^0.71.2",
         "@google/generative-ai": "^0.24.1",
+        "fs": "^0.0.1-security",
         "openai": "^6.15.0",
-        "sharp": "^0.34.5",
-        "xlsx": "^0.18.5",
-        "pdf-parse": "^2.4.5",
         "path": "^0.12.7",
-        "fs": "^0.0.0"
+        "pdf-parse": "^2.4.5",
+        "sharp": "^0.34.5",
+        "xlsx": "^0.18.5"
     }
 }

package/server_connect/modules/multiaiv2.js

@@ -14,7 +14,10 @@ const clamp = (val, min, max) => Math.max(min, Math.min(max, val));
  */
 async function getCompressedImageBase64(filePath) {
     try {
-        const fullPath = path.join(process.cwd(), filePath);
+        // Remove leading slash if present to avoid path joining issues
+        const cleanPath = filePath.startsWith('/') ? filePath.substring(1) : filePath;
+        const fullPath = path.join(process.cwd(), cleanPath);
+
         console.log("Checking for image at:", fullPath);
 
         if (!fs.existsSync(fullPath)) {
@@ -23,6 +26,7 @@ async function getCompressedImageBase64(filePath) {
         }
 
         const buffer = await sharp(fullPath)
+            .rotate() // <--- CRITICAL: Fixes portrait/landscape flip
             .resize(1500, 1500, { fit: 'inside', withoutEnlargement: true })
             .jpeg({ quality: 80 })
             .toBuffer();
@@ -58,13 +62,20 @@ async function getFileContent(filePath) {
 }
 
 exports.multiaiv2 = async function (options) {
+
+    // 1. Ensure we grab the question correctly
     let question = this.parseRequired(options.question, "*", 'No Question passed');
+
+    // 2. Fix the file input assignment
+    // Use the exact key from your console log
+    const fileInput = options.image_url;
+
     const engine = this.parse(options.engine) || 'gpt-4o-mini';
     const maxTokens = parseInt(this.parseOptional(options.maxtokens, "*", 2000));
     const temp = parseFloat(this.parseOptional(options.temperature, "*", 0.7));
-    const fileInput = this.parse(options.image_url);
 
     const modelName = engine.toLowerCase();
+
     const isReasoningModel = modelName.includes('gpt-5') || modelName.startsWith('o1') || modelName.startsWith('o3');
 
     console.log(`Starting Request - Model: ${engine}, File: ${fileInput}`);
@@ -80,14 +91,15 @@ exports.multiaiv2 = async function (options) {
 
     try {
         // --- OpenAI (GPT-4o, GPT-5, o1) ---
+
         if (modelName.includes('gpt') || modelName.includes('o1') || modelName.includes('o3')) {
             const client = new OpenAI({ apiKey: process.env.OPENAI_API_KEY });
+            let userContent = [];
 
-            let userContent = [{ type: "text", text: question }];
-
-            if (fileInput && /\.(jpg|jpeg|png|webp)$/i.test(fileInput) && !isReasoningModel) {
+            if (fileInput && /\.(jpg|jpeg|png|webp)$/i.test(fileInput)) {
                 const b64 = await getCompressedImageBase64(fileInput);
                 if (b64) {
+                    // Push image first
                     userContent.push({
                         type: "image_url",
                         image_url: { url: b64, detail: "high" }
@@ -96,11 +108,16 @@ exports.multiaiv2 = async function (options) {
                 }
             }
 
+            // Add text instructions after the image
+            userContent.push({ type: "text", text: question });
+
             const payload = {
                 model: engine,
                 messages: [{ role: 'user', content: userContent }]
             };
 
+
+
             if (isReasoningModel) {
                 payload.max_completion_tokens = Math.max(maxTokens, 5000);
             } else {