Update app.py

app.py

@@ -5,7 +5,7 @@ import torch
 import os
 import spaces
 
-# --- Initialize the Model Pipeline ---
+# --- Initialize the Model Pipeline (No changes here) ---
 print("Loading MedGemma model...")
 try:
     pipe = pipeline(
@@ -25,23 +25,17 @@ except Exception as e:
 @spaces.GPU()
 def analyze_symptoms(symptom_image, symptoms_text):
     """
-    Analyzes user's symptoms using the correct prompt format for MedGemma.
+    Analyzes user's symptoms using the correct prompt format and keyword arguments for MedGemma.
     """
     if not model_loaded:
         return "Error: The AI model could not be loaded. Please check the Space logs."
 
-    # Standardize input to avoid issues with None or whitespace
     symptoms_text = symptoms_text.strip() if symptoms_text else ""
-
     if symptom_image is None and not symptoms_text:
         return "Please describe your symptoms or upload an image for analysis."
 
     try:
-        # --- CORRECTED PROMPT LOGIC ---
-        # MedGemma expects a specific prompt format with special tokens.
-        # We build this prompt string dynamically.
-        
-        # This is the instruction part of the prompt
+        # --- PROMPT LOGIC (Unchanged) ---
         instruction = (
             "You are an expert, empathetic AI medical assistant. "
             "Analyze the potential medical condition based on the following information. "
@@ -49,47 +43,42 @@ def analyze_symptoms(symptom_image, symptoms_text):
             "actionable next-steps plan. Start your analysis by describing the user-provided "
             "information (text and/or image)."
         )
-
-        # Build the final prompt based on user inputs
         prompt_parts = ["<start_of_turn>user"]
         if symptoms_text:
             prompt_parts.append(symptoms_text)
-        
-        # The <image> token is a placeholder that tells the model where to "look" at the image.
         if symptom_image:
             prompt_parts.append("<image>")
-            
         prompt_parts.append(instruction)
         prompt_parts.append("<start_of_turn>model")
-        
         prompt = "\n".join(prompt_parts)
         
         print("Generating pipeline output...")
         
-        # --- CORRECTED PIPELINE CALL ---
-        # The pipeline expects the prompt string and an 'images' argument (if an image is provided).
-        # We create a dictionary for keyword arguments to pass to the pipeline.
-        pipeline_kwargs = {
+        # --- CORRECTED & ROBUST PIPELINE CALL ---
+        # We build a dictionary of all arguments to pass to the pipeline.
+        # This avoids the TypeError by ensuring all arguments are passed explicitly by keyword.
+        
+        pipeline_args = {
+            "prompt": prompt,
             "max_new_tokens": 512,
             "do_sample": True,
             "temperature": 0.7
         }
 
         # The `images` argument should be a list of PIL Images.
+        # We only add it to our arguments dictionary if an image is provided.
         if symptom_image:
-            output = pipe(prompt, images=[symptom_image], **pipeline_kwargs)
-        else:
-            # If no image is provided, we do not include the `images` argument in the call.
-            output = pipe(prompt, **pipeline_kwargs)
+            pipeline_args["images"] = [symptom_image]
+            
+        # We use the ** syntax to unpack the dictionary into keyword arguments.
+        # This results in a call like: pipe(prompt=..., images=..., max_new_tokens=...)
+        output = pipe(**pipeline_args)
             
         print("Pipeline Output:", output)
 
-        # --- SIMPLIFIED OUTPUT PROCESSING ---
-        # The pipeline returns a list with one dictionary. The result is in the 'generated_text' key.
+        # --- SIMPLIFIED OUTPUT PROCESSING (Unchanged) ---
         if output and isinstance(output, list) and 'generated_text' in output[0]:
-            # We extract just the model's response part of the generated text.
             full_text = output[0]['generated_text']
-            # The model output includes the prompt, so we split it to get only the new part.
             result = full_text.split("<start_of_turn>model\n")[-1]
         else:
             result = "The model did not return a valid response. Please try again."
@@ -100,10 +89,9 @@ def analyze_symptoms(symptom_image, symptoms_text):
 
     except Exception as e:
         print(f"An exception occurred during analysis: {type(e).__name__}: {e}")
-        # Provide a more user-friendly error message
         return f"An error occurred during analysis. Please check the logs for details: {str(e)}"
 
-# --- Create the Gradio Interface (No changes needed here) ---
+# --- Gradio Interface (No changes needed) ---
 with gr.Blocks(theme=gr.themes.Soft(), title="AI Symptom Analyzer") as demo:
     gr.HTML("""
         <div style="text-align: center; background: linear-gradient(135deg, #667eea 0%, #764ba2 100%); color: white; padding: 2rem; border-radius: 10px; margin-bottom: 2rem;">