works

src/app.py

@@ -1,9 +1,18 @@
 import os
 import gradio as gr
-from llama_index.llms.huggingface import HuggingFaceLLM
-from parse_tabular import symptom_index
+from llama_index.core import Settings
+from llama_index.embeddings.huggingface import HuggingFaceEmbedding
+from parse_tabular import create_symptom_index  # Change this import
 import json
 
+# Configure embeddings globally
+Settings.embed_model = HuggingFaceEmbedding(
+    model_name="sentence-transformers/all-MiniLM-L6-v2"
+)
+
+# Create the index at startup
+symptom_index = create_symptom_index()
+
 # --- System prompt ---
 SYSTEM_PROMPT = """
 You are a medical assistant helping a user narrow down to the most likely ICD-10 code.
@@ -47,20 +56,9 @@ with gr.Blocks() as demo:
     
     microphone = gr.Microphone(
         streaming=True,
-        type="filepath"  # Return transcriptions as strings
+        type="filepath"  # Use 'filepath' to get the path to the recorded audio file
     )
     
-    # MCP metadata (store in a separate variable, not as an attribute of demo)
-    mcp_metadata = {
-        "title": "Medical Symptom to ICD-10 Code Assistant",
-        "description": "Convert spoken symptoms to ICD-10 codes",
-        "version": "1.0.0",
-        "capabilities": {
-            "speech_input": True,
-            "streaming": True
-        }
-    }
-    
     # Connect the streaming microphone to the chat
     microphone.stream(
         fn=process_speech,