Update app.py

app.py

@@ -145,49 +145,43 @@ def generate_music(transcribed_text, emotion_prediction):
         print("Error generating music:", e)
         return None
 
+# --- DeepAI Image Generation (Text2Img) ---
 api_key = os.getenv("DeepAI_api_key")
 
-# Function to generate an image using DeepAI Text to Image API
-def generate_image(emotion_prediction, transcribed_text, output_resolution=(1024, 1024)):
+def generate_image(emotion_prediction, transcribed_text):
     try:
         if not api_key:
-            return "API key not found"
+            # fallback white image if no API key
+            return Image.new('RGB', (512, 512), color='white')
             
-        url = "https://api.deepai.org/api/image-editor"
-        headers = {
-            'api-key': api_key
-        }
+        # Create the prompt for text2img
+        prompt = f"Generate Patagonian Monsters with a {emotion_prediction} attitude, representing the idea of: [{transcribed_text}]. Illustrate this using asemic writings in an old map style."
         
-        # Select a random image file from TerraIncognita0.jpg to TerraIncognita9.jpg
-        random_index = random.randint(0, 9)
-        image_file_path = f'TAI_Images/TerraIncognita{random_index}.jpg'
+        # Make request to DeepAI text2img API
+        response = requests.post(
+            "https://api.deepai.org/api/text2img",
+            data={
+                'text': prompt, 
+                'width': 512, 
+                'height': 512, 
+                'image_generator_version': 'hd'
+            },
+            headers={'api-key': api_key}
+        )
         
-        # Check if the file exists
-        if not os.path.exists(image_file_path):
-            return f"Image file not found: {image_file_path}"
-            
-        prompt_text = f"Generate Patagonian Monsters' with a {emotion_prediction} attitude, representing the idea of: [ {transcribed_text} ]. Illustrate this using asemic writings in an old map style."
-        
-        with open(image_file_path, 'rb') as image_file:
-            files = {
-                'image': image_file,
-            }
-            data = {
-                'text': prompt_text
-            }
-            response = requests.post(url, headers=headers, files=files, data=data)
-            
-        response_data = response.json()
-        if 'output_url' in response_data:
-            # Download the image and return it as a PIL Image
-            image_response = requests.get(response_data['output_url'])
-            return Image.open(BytesIO(image_response.content))
+        data = response.json()
+        if 'output_url' in data:
+            # Download the generated image
+            img_resp = requests.get(data['output_url'])
+            return Image.open(BytesIO(img_resp.content))
         else:
-            print("Error in DeepAI response:", response_data)
-            return None
+            print("Error in DeepAI response:", data)
+            # Return a fallback image
+            return Image.new('RGB', (512, 512), color='white')
     except Exception as e:
         print("Error generating image:", e)
-        return None
+        # Return a fallback image
+        return Image.new('RGB', (512, 512), color='white')
 
 # Function to get predictions
 def get_predictions(audio_input):
@@ -201,6 +195,7 @@ def get_predictions(audio_input):
     # Analyze sentiment of transcribed text
     sentiment, polarity = analyze_sentiment(transcribed_text)
     
+    # Generate image using text2img
     image = generate_image(emotion_prediction, transcribed_text)
     
     # Generate music based on transcription and emotion
@@ -217,7 +212,7 @@ interface = gr.Interface(
         gr.Label(label="Transcribed Text"),
         gr.Label(label="Sentiment Analysis"),
         gr.Image(type='pil', label="Generated Image"),
-        gr.Audio(label="Generated Music", type="filepath")  # Added music output
+        gr.Audio(label="Generated Music", type="filepath")
     ],
     title="Affective Virtual Environments",
     description="Create an AVE using your voice. Get emotion prediction, transcription, sentiment analysis, a generated image, and music."