Update app.py

app.py

@@ -1,8 +1,8 @@
-import gradio as gr
+import requests
 import numpy as np
 import tensorflow as tf
-from tensorflow import keras
 import tensorflow_hub as hub
+import gradio as gr
 from PIL import Image
 
 # Load models
@@ -17,8 +17,43 @@ from PIL import Image
 #)
 #model_alzheimer = keras.models.load_model(
 #    "models/model_alzheimer.h5", custom_objects={'KerasLayer': hub.KerasLayer}
-#)
 
+# API key and user ID for the new API
+api_key = 'KGSjxB1uptfSk8I8A7ciCuNT9Xa3qWC3'
+external_user_id = 'plugin-1717464304'
+
+# Step 1: Create a chat session
+def create_chat_session():
+    create_session_url = 'https://api.on-demand.io/chat/v1/sessions'
+    create_session_headers = {
+        'apikey': api_key
+    }
+    create_session_body = {
+        "pluginIds": [],
+        "externalUserId": external_user_id
+    }
+    # Make the request to create a chat session
+    response = requests.post(create_session_url, headers=create_session_headers, json=create_session_body)
+    response_data = response.json()
+    session_id = response_data['data']['id']
+    return session_id
+
+# Step 2: Submit a query to the API
+def submit_query(session_id, query):
+    submit_query_url = f'https://api.on-demand.io/chat/v1/sessions/{session_id}/query'
+    submit_query_headers = {
+        'apikey': api_key
+    }
+    submit_query_body = {
+        "endpointId": "predefined-openai-gpt4o",
+        "query": query,
+        "pluginIds": ["plugin-1712327325", "plugin-1713962163"],
+        "responseMode": "sync"
+    }
+    response = requests.post(submit_query_url, headers=submit_query_headers, json=submit_query_body)
+    return response.json()['data']['response']
+
+# CNN Model for MRI Image Diagnosis (mockup since actual model code isn't available)
 class CombinedDiseaseModel(tf.keras.Model):
     def __init__(self, model_initial, model_alzheimer, model_tumor, model_stroke):
         super(CombinedDiseaseModel, self).__init__()
@@ -60,37 +95,50 @@ class CombinedDiseaseModel(tf.keras.Model):
 
         return f"The MRI image shows {main_disease} with a probability of {main_disease_prob*100:.2f}%.\nThe subcategory of {main_disease} is {sub_category} with a probability of {sub_category_prob*100:.2f}%."
 
-
-# Initialize the combined model
-#cnn_model = CombinedDiseaseModel(
-#    model_initial=model_initial,
-#    model_alzheimer=model_alzheimer,
-#    model_tumor=model_tumor,
-#    model_stroke=model_stroke
-#)
-
+# Example CNN models (mockup)
+# cnn_model = CombinedDiseaseModel(model_initial, model_alzheimer, model_tumor, model_stroke)
 
 def process_image(image):
+    """
+    Processes the uploaded MRI image and makes predictions using the combined CNN model.
+    """
     image = image.resize((256, 256))
-    image.convert("RGB")
+    image = image.convert("RGB")
     image_array = np.array(image) / 255.0
     image_array = np.expand_dims(image_array, axis=0)
-    predictions = cnn_model(image_array)
+    
+    predictions = cnn_model(image_array)  # Call the model to get predictions (replace with actual model call)
     return predictions
 
-
-def gradio_interface(patient_info, query_type, image):
+def query_llm_via_on_demand(patient_info, query_type):
+    """
+    Sends patient information and query type to the on-demand API and returns the generated response.
+    """
+    session_id = create_chat_session()
+    query = f"Patient Information: {patient_info}\nQuery Type: {query_type}\nPlease provide additional insights."
+    
+    response = submit_query(session_id, query)
+    return response
+
+def gradio_interface(patient_info, query_type, image=None):
+    """
+    Gradio interface function that processes patient info, query type, and an optional MRI image, 
+    and provides results from both the CNN model and the on-demand LLM API.
+    """
+    image_response = ""
     if image is not None:
         image_response = process_image(image)
-        response = f"Patient Info: {patient_info}\nQuery Type: {query_type}\n{image_response}"
-        return response
-    else:
-        return "Please upload an image."
-
-def gradio_interface(patient_info, query_type):
-    return f"Patient Info: {patient_info}\nQuery Type: {query_type}"
-
-# Create Gradio app
+    
+    llm_response = query_llm_via_on_demand(patient_info, query_type)
+    
+    response = f"Patient Info: {patient_info}\nQuery Type: {query_type}\n\nLLM Response:\n{llm_response}"
+    
+    if image_response:
+        response += f"\n\nImage Diagnosis:\n{image_response}"
+    
+    return response
+
+# Create Gradio interface with MRI image upload option
 iface = gr.Interface(
     fn=gradio_interface,
     inputs=[
@@ -101,16 +149,19 @@ iface = gr.Interface(
             max_lines=10
         ),
         gr.Textbox(
-            label="Query Type"
-        )#,
-        #gr.Image(
-        #    type="pil",
-        #    label="Upload an Image",
-        #)
+            label="Query Type",
+            placeholder="Enter the query type here..."
+        ),
+        gr.Image(
+            type="pil",
+            label="Upload an MRI Image",
+            optional=True  # Allow the image input to be optional
+        )
     ],
     outputs=gr.Textbox(label="Response", placeholder="The response will appear here..."),
-    title="Medical Diagnosis with MRI",
-    description="Upload MRI images and provide patient information for diagnosis.",
+    title="Medical Diagnosis with MRI and On-Demand LLM Insights",
+    description="Upload MRI images and provide patient information for diagnosis. The system integrates MRI diagnosis with insights from the on-demand LLM API."
 )
 
-iface.launch()
+# Launch the Gradio app
+iface.launch()