refactor(app.py): improve token handling and error management for Hugging Face integration

Enhance the robustness of the Hugging Face token handling by introducing a fallback mechanism using Streamlit secrets or environment variables. Additionally, improve error management during model loading to provide better user feedback in case of failures.

app.py

@@ -676,14 +676,37 @@ from transformers import pipeline, AutoModelForSequenceClassification, AutoToken
 import os
 from huggingface_hub import login
 
-login(token=os.environ.get("HF_TOKEN"))
+# login(token=os.environ.get("HF_TOKEN"))
+
+try:
+    # For Streamlit Cloud or Spaces deployment
+    hf_token = st.secrets["HF_TOKEN"]
+except:
+    # Fallback to environment variables for local development
+    hf_token = os.environ.get("HF_TOKEN")
+
+if hf_token:
+    login(token=hf_token)
+else:
+    st.warning("Hugging Face token not found. Some features may not work correctly.")
+
+
+# if selected == 'Mental-Analysis':
+#     # Load the Hugging Face model
+#     model_name = "mental/mental-roberta-base"
+#     tokenizer = AutoTokenizer.from_pretrained(model_name)
+#     model = AutoModelForSequenceClassification.from_pretrained(model_name)
+
 
 if selected == 'Mental-Analysis':
     # Load the Hugging Face model
-    model_name = "mental/mental-roberta-base"
-    tokenizer = AutoTokenizer.from_pretrained(model_name)
-    model = AutoModelForSequenceClassification.from_pretrained(model_name)
-
+    try:
+        model_name = "mental/mental-roberta-base"
+        tokenizer = AutoTokenizer.from_pretrained(model_name)
+        model = AutoModelForSequenceClassification.from_pretrained(model_name)
+    except Exception as e:
+        st.error(f"Error loading mental health model: {e}")
+        st.info("Please check your Hugging Face token configuration.")
     # Sidebar with title and markdown
     st.sidebar.title("🧠 Mental Health Analysis")
     st.sidebar.markdown("""