Update app.py

app.py

@@ -57,9 +57,9 @@ def generate_forecast(model, df):
     forecast_df = model.predict(df=df)
     return forecast_df
 
-def determine_frequency(df):
-    df['ds'] = pd.to_datetime(df['ds'])
-    df = df.set_index('ds')
+def determine_frequency(df, ds_col):
+    df[ds_col] = pd.to_datetime(df[ds_col])
+    df = df.set_index(ds_col)
     freq = pd.infer_freq(df.index)
     return freq
 
@@ -155,18 +155,18 @@ def select_model(horizon, model_type, max_steps=200):
     else:
         raise ValueError(f"Unsupported model type: {model_type}")
 
-def model_train(df,model):
-    df['ds'] = pd.to_datetime(df['ds'])
+def model_train(df,model, ds_col):
+    df[ds_col] = pd.to_datetime(df[ds_col])
     model.fit(df)
     return model
 
-def forecast_time_series(df, model_type, horizon, max_steps=200):
+def forecast_time_series(df, model_type, horizon, max_steps=200, ds_col='ds'):
     start_time = time.time()  # Start timing
-    freq = determine_frequency(df)
+    freq = determine_frequency(df, ds_col)
     st.sidebar.write(f"Data frequency: {freq}")
     
     selected_model = select_model(horizon, model_type, max_steps)
-    model = model_train(df, selected_model)
+    model = model_train(df, selected_model, ds_col)
     
     forecast_results = {}
     st.sidebar.write(f"Generating forecast using {model_type} model...")
@@ -187,22 +187,35 @@ def load_default():
 def transfer_learning_forecasting():
     st.title("Transfer Learning Forecasting")
 
-    # Upload dataset
-    uploaded_file = st.sidebar.file_uploader("Upload your time series data (CSV)", type=["csv"])
-    if uploaded_file:
-        df = pd.read_csv(uploaded_file)
-    else:
-        df = load_default()  
+    with st.sidebar.expander("Upload and Configure Dataset", expanded=True):
+        uploaded_file = st.file_uploader("Upload your time series data (CSV)", type=["csv"])
+        if uploaded_file:
+            df = pd.read_csv(uploaded_file)
+            st.session_state.df = df
+        else:
+            df = load_default()  
+            st.session_state.df = df
+        
+        # Column selection
+        columns = df.columns if not df.empty else []
+        ds_col = st.selectbox("Select Date/Time column", options=columns, index=columns.index('ds') if 'ds' in columns else 0)
+        y_col = st.selectbox("Select Target column", options=columns, index=columns.index('y') if 'y' in columns else 1)
+        unique_id_col = st.text_input("Unique ID column (default: '1')", value="1")
+
+    st.session_state.ds_col = ds_col
+    st.session_state.y_col = y_col
+    st.session_state.unique_id_col = unique_id_col
 
-    nhits_models, timesnet_models, lstm_models, tft_models = load_all_models()
-    
     # Model selection and forecasting
     st.sidebar.subheader("Model Selection and Forecasting")
     model_choice = st.sidebar.selectbox("Select model", ["NHITS", "TimesNet", "LSTM", "TFT"])
     horizon = st.sidebar.number_input("Forecast horizon", value=18)
 
+    df = df.rename(columns={ds_col: 'ds', y_col: 'y'})
+    st.session_state.df = df
+
     # Determine frequency of data
-    frequency = determine_frequency(df)
+    frequency = determine_frequency(df, 'ds')
     st.sidebar.write(f"Detected frequency: {frequency}")
 
     # Load pre-trained models
@@ -229,20 +242,35 @@ def transfer_learning_forecasting():
 def dynamic_forecasting():
     st.title("Dynamic Forecasting")
     
-    # Upload dataset
-    uploaded_file = st.sidebar.file_uploader("Upload your time series data (CSV)", type=["csv"])
-    if uploaded_file:
-        df = pd.read_csv(uploaded_file)
-    else:
-        df = load_default()
-    
+    with st.sidebar.expander("Upload and Configure Dataset", expanded=True):
+        uploaded_file = st.file_uploader("Upload your time series data (CSV)", type=["csv"])
+        if uploaded_file:
+            df = pd.read_csv(uploaded_file)
+            st.session_state.df = df
+        else:
+            df = load_default()
+            st.session_state.df = df
+        
+        # Column selection
+        columns = df.columns if not df.empty else []
+        ds_col = st.selectbox("Select Date/Time column", options=columns, index=columns.index('ds') if 'ds' in columns else 0)
+        y_col = st.selectbox("Select Target column", options=columns, index=columns.index('y') if 'y' in columns else 1)
+        unique_id_col = st.text_input("Unique ID column (default: '1')", value="1")
+
+    st.session_state.ds_col = ds_col
+    st.session_state.y_col = y_col
+    st.session_state.unique_id_col = unique_id_col
+
     # Dynamic forecasting
     st.sidebar.subheader("Dynamic Model Selection and Forecasting")
     dynamic_model_choice = st.sidebar.selectbox("Select model for dynamic forecasting", ["NHITS", "TimesNet", "LSTM", "TFT"], key="dynamic_model_choice")
     dynamic_horizon = st.sidebar.number_input("Forecast horizon", value=18)
     dynamic_max_steps = st.sidebar.number_input('Max steps', value=200)
 
-    forecast_time_series(df, dynamic_model_choice, dynamic_horizon, dynamic_max_steps)
+    df = df.rename(columns={ds_col: 'ds', y_col: 'y'})
+    st.session_state.df = df
+    
+    forecast_time_series(df, dynamic_model_choice, dynamic_horizon, dynamic_max_steps, ds_col='ds')
 
 pg = st.navigation({
     "Overview": [