Update app.py

app.py

@@ -12,6 +12,11 @@ from scipy.io.wavfile import write
 import PIL
 from openai import OpenAI
 import time
+from PIL import Image
+import io
+import hashlib
+import datetime
+
 dotenv.load_dotenv()
 
 seamless_client = Client("facebook/seamless_m4t")
@@ -81,32 +86,91 @@ def process_speech(input_language, audio_input):
     except Exception as e :
         return f"{e}"
 
+def decode_image(encoded_image: str) -> Image:
+    decoded_bytes = base64.b64decode(encoded_image.encode("utf-8"))
+    buffer = io.BytesIO(decoded_bytes)
+    image = Image.open(buffer)
+    return image
+
+
+def encode_image(image: Image.Image, format: str = "PNG") -> str:
+    with io.BytesIO() as buffer:
+        image.save(buffer, format=format)
+        encoded_image = base64.b64encode(buffer.getvalue()).decode("utf-8")
+    return encoded_image
+
+
+def get_conv_log_filename():
+    t = datetime.datetime.now()
+    name = os.path.join(LOGDIR, f"{t.year}-{t.month:02d}-{t.day:02d}-conv.json")
+    return name
+
 
+def get_conv_image_dir():
+    name = os.path.join(LOGDIR, "images")
+    os.makedirs(name, exist_ok=True)
+    return name
 
 
-def process_image(image):
-    # Save the image to a file
-    img_name = f"{np.random.randint(0, 100)}.jpg"
-    PIL.Image.fromarray(image.astype('uint8'), 'RGB').save(img_name)
+def get_image_name(image, image_dir=None):
+    buffer = io.BytesIO()
+    image.save(buffer, format="PNG")
+    image_bytes = buffer.getvalue()
+    md5 = hashlib.md5(image_bytes).hexdigest()
+
+    if image_dir is not None:
+        image_name = os.path.join(image_dir, md5 + ".png")
+    else:
+        image_name = md5 + ".png"
+
+    return image_name
+
+def resize_image(image, max_size):
+    width, height = image.size
+    aspect_ratio = float(width) / float(height)
+
+    if width > height:
+        new_width = max_size
+        new_height = int(new_width / aspect_ratio)
+    else:
+        new_height = max_size
+        new_width = int(new_height * aspect_ratio)
 
-    # Prepare the payload for the Gradio client
+    resized_image = image.resize((new_width, new_height))
+    return resized_image
+
+
+
+def process_image(image_input, text_input):
+    # Resize the image if needed
+    max_image_size = 1024  # You can adjust this size
+    image = resize_image(image_input, max_image_size)
+
+    # Encode the image to base64
+    base64_image_str = encode_image(image)
+
+    # Prepare the payload for the HTTP request
     payload = {
-        "data": [
-            img_name,  # File path of the saved image
-            "Hello!!"  # Example text, replace with relevant text if needed
+        "content": [
+            {
+                "prompt": text_input,
+                "image": base64_image_str,
+            }
         ],
-        "api_name": "/http_bot"
+        "token": "sk-OtterHD",  # Replace with your actual token
     }
 
-    # Initialize the Gradio client
-    client = Client("https://teamtonic-otterhd-demo.hf.space/--replicas/rcng4/")
+    # Specify the URL for the HTTP request
+    url = "https://ensures-picture-choices-labels.trycloudflare.com/app/otter"
+    headers = {"Content-Type": "application/json"}
 
-    try:
-        # Send the request to the Gradio client
-        result = client.predict(payload)
-        return result
-    except Exception as e:
-        return f"Error: {e}"
+    # Make the HTTP request
+    response = requests.post(url, headers=headers, data=json.dumps(payload))
+    if response.status_code == 200:
+        results = response.json()
+        return results["result"]
+    else:
+        return f"Error: {response.status_code}, {response.text}"
 
 
 def query_vectara(text):