fix add inference client

app.py

@@ -4,6 +4,7 @@ import os
 import logging
 from utils.meldrx import MeldRxAPI
 from utils.oneclick import generate_discharge_paper_one_click
+from huggingface_hub import InferenceClient
 
 logging.basicConfig(level=logging.INFO)
 logger = logging.getLogger(__name__)
@@ -23,6 +24,13 @@ meldrx_api = MeldRxAPI(
     redirect_uri=REDIRECT_URI
 )
 
+
+HF_TOKEN = os.getenv("HF_TOKEN")
+if not HF_TOKEN:
+    raise ValueError("HF_TOKEN environment variable not set.")
+client = InferenceClient(api_key=HF_TOKEN)
+MODEL_NAME = "meta-llama/Llama-3.3-70B-Instruct"
+
 @app.route('/')
 def index():
     return render_template('index.html')
@@ -67,7 +75,7 @@ def one_click():
         logger.info(f"One-click request - ID: {patient_id}, First: {first_name}, Last: {last_name}, Action: {action}")
 
         pdf_path, status, basic_summary, ai_summary = generate_discharge_paper_one_click(
-            meldrx_api, grok_client, patient_id, first_name, last_name
+            meldrx_api, client, patient_id, first_name, last_name
         )
 
         if action == "Display Summary":