Update app.py

app.py

@@ -17,14 +17,23 @@ from streamlit_extras.add_vertical_space import add_vertical_space
 
 UPLOAD_FOLDER = "uploaded_nmr"
 
+def is_running_on_huggingface():
+    """Checks if the app is running on Hugging Face Spaces."""
+    return "HF_SPACE_ID" in os.environ
+    
 # ----------------------------- CACHED FUNCTIONS -----------------------------
 @st.cache_data
 def build_library():
     """Build NMR library from HF."""
+    if is_running_on_huggingface():
+        token = os.getenv("HF_TOKEN") # If on huggingface
+    else:
+        token = st.secrets["HF_TOKEN"] # If on streamlit community cloud
+
     nmr_dataset = load_dataset(
         "mahynski/bmrb-hsqc-nmr-1H13C",
         split="train",
-        token=os.getenv("HF_TOKEN"),
+        token=token,
         trust_remote_code=True,
     )
     substances = [
@@ -36,8 +45,6 @@ def build_library():
     lib = finchnmr.library.Library(substances)
     return lib
 
-
-# @st.cache_data
 def build_model(_target, _lib, _param_grid, _nmr_model, _model_kw):
     """Build model for target."""
     optimized_models, analyses = finchnmr.model.optimize_models(
@@ -49,7 +56,6 @@ def build_model(_target, _lib, _param_grid, _nmr_model, _model_kw):
     )
     return optimized_models, analyses
 
-
 # --------------------------------- SIDEBAR ----------------------------------
 st.set_page_config(layout="wide")
 st.header("Analyze an HSQC NMR Spectra with FINCHnmr")
@@ -112,8 +118,6 @@ if uploaded_file is not None:
         warning="ignore",
     )
 
-    # optimized_models = []
-
     tab1_, tab2_ = st.tabs(["Configure Model", "Analyze Results"])
     with tab1_:
         st.subheader("Configure Model")
@@ -258,6 +262,8 @@ if uploaded_file is not None:
             model_ = st.session_state['optimized_models'][0]  # We only fit one model
             analysis_ = st.session_state['analyses'][0]
 
+            st.metric(label="Model's score (R^2)", value=model_.score())
+
             # Plot original vs. reconstructed and residual
             col3_, col4_ = st.columns(2)
             with col3_:
@@ -317,7 +323,6 @@ if uploaded_file is not None:
             max_n_ = len(analysis_._model.importances())
             default_n_ = np.min([10, max_n_])
             with col6_:
-                st.metric(label="Model's score (R^2)", value=model_.score())
                 n_imp_ = st.slider(
                     label="Visualize the most important N spectra in the library",
                     value=default_n_,