Update app.py

app.py

@@ -6,6 +6,7 @@ from datetime import datetime
 from simple_salesforce import Salesforce
 from dotenv import load_dotenv
 import plotly.express as px
+import plotly.graph_objects as go
 
 # Load environment variables from .env
 load_dotenv()
@@ -26,7 +27,7 @@ except Exception as e:
     sf = None
     print(f"Error connecting to Salesforce: {str(e)}")
 
-# Function to fetch Project ID from Salesforce automatically
+# Function to fetch Project ID from Salesforce
 def get_project_id():
     if not sf:
         return None, "Salesforce connection failed. Check credentials."
@@ -39,9 +40,10 @@ def get_project_id():
     except Exception as e:
         return None, f"Error fetching Project ID: {str(e)}"
 
-# Simple moving average forecast
+# Simple moving average forecast (works with 1+ days of data)
 def simple_forecast(df):
     df['Date'] = pd.to_datetime(df['Date'], dayfirst=True)
+    # Use rolling mean with min_periods=1 to allow single-day data
     df['Forecast'] = df['Attendance'].rolling(window=3, min_periods=1).mean()
     future_dates = pd.date_range(df['Date'].max(), periods=4, freq='D')[1:]
     future_preds = np.repeat(df['Forecast'].iloc[-1], 3)
@@ -66,10 +68,14 @@ def create_chart(df, predictions_dict):
     combined_df = pd.DataFrame()
     for trade, predictions in predictions_dict.items():
         trade_df = df[df['Trade'] == trade].copy()
+        if trade_df.empty:
+            continue
         trade_df['Type'] = 'Historical'
         trade_df['Trade'] = trade
 
         forecast_df = pd.DataFrame(predictions)
+        if forecast_df.empty:
+            continue
         forecast_df['Date'] = pd.to_datetime(forecast_df['date'])
         forecast_df['Attendance'] = forecast_df['headcount']
         forecast_df['Type'] = 'Forecast'
@@ -81,6 +87,9 @@ def create_chart(df, predictions_dict):
             forecast_df[['Date', 'Attendance', 'Type', 'Trade']]
         ])
 
+    if combined_df.empty:
+        return go.Figure().update_layout(title="Labour Attendance Forecast (No Data)")
+
     fig = px.line(
         combined_df,
         x='Date',
@@ -105,7 +114,7 @@ def format_output(trade_results):
                 value = ', '.join(str(item) for item in value)
             output.append(f"  • {key}: {value}")
         output.append("")
-    return "\n".join(output)
+    return "\n".join(output) if trade_results else "No valid trade data available."
 
 # Forecast function for Gradio
 def forecast_labour(csv_file):
@@ -119,37 +128,54 @@ def forecast_labour(csv_file):
             except UnicodeDecodeError:
                 continue
         if df is None:
-            return "Error: Could not decode CSV file with any supported encoding (utf-8, latin1, iso-8859-1, utf-16). Please ensure the file is properly encoded.", None
+            return "Error: Could not decode CSV file with any supported encoding (utf-8, latin1, iso-8859-1, utf-16).", None
 
         df.columns = df.columns.str.strip().str.capitalize()
-
         required_columns = ['Date', 'Attendance', 'Trade', 'Weather', 'Alert_status', 'Shortage_risk', 'Suggested_actions']
         missing_columns = [col for col in required_columns if col not in df.columns]
         if missing_columns:
             return f"Error: CSV missing required columns: {', '.join(missing_columns)}", None
 
-        df['Date'] = pd.to_datetime(df['Date'], dayfirst=True)
-        df['Attendance'] = df['Attendance'].astype(int)
+        df['Date'] = pd.to_datetime(df['Date'], dayfirst=True, errors='coerce')
+        if df['Date'].isna().all():
+            return "Error: All dates in CSV are invalid.", None
+        df['Attendance'] = pd.to_numeric(df['Attendance'], errors='coerce').fillna(0).astype(int)
         df['Shortage_risk'] = df['Shortage_risk'].replace('%', '', regex=True).astype(float) / 100
+        df['Shortage_risk'] = df['Shortage_risk'].fillna(0.5)
 
         unique_trades = df['Trade'].unique()
         if len(unique_trades) < 10:
-            return f"Error: CSV contains only {len(unique_trades)} trades, but a minimum of 10 trades is required.", None
+            return f"Error: CSV contains only {len(unique_trades)} trades, minimum 10 required.", None
+
+        # Check data sufficiency per trade
+        trade_data_counts = {trade: len(df[df['Trade'] == trade]) for trade in unique_trades}
+        insufficient_trades = [trade for trade, count in trade_data_counts.items() if count < 1]
+        if insufficient_trades:
+            return (f"Error: The following trades have no data: {', '.join(insufficient_trades)}"), None
 
         selected_trades = unique_trades[:10]
         trade_results = {}
         predictions_dict = {}
+        processed_trades = set()  # Track processed trades to avoid duplicates
+        errors = []
 
         project_id, error = get_project_id()
         if error:
             return f"Error: {error}", None
 
         for trade in selected_trades:
+            if trade in processed_trades:
+                continue  # Skip duplicates
+            processed_trades.add(trade)
             trade_df = df[df['Trade'] == trade].copy()
             if trade_df.empty:
+                errors.append(f"No data for trade: {trade}")
                 continue
 
             predictions = simple_forecast(trade_df)
+            if not predictions:
+                errors.append(f"No forecast generated for trade: {trade}")
+                continue
             predictions_dict[trade] = predictions
 
             latest_record = trade_df.sort_values(by='Date').iloc[-1]
@@ -190,6 +216,12 @@ def forecast_labour(csv_file):
             result_data.update(sf_result)
             trade_results[trade] = result_data
 
+        if not trade_results:
+            error_msg = "No valid trade data processed."
+            if errors:
+                error_msg += " Errors: " + "; ".join(errors)
+            return error_msg, None
+
         chart = create_chart(df, predictions_dict)
         return format_output(trade_results), chart
 
@@ -208,8 +240,8 @@ def gradio_interface():
             gr.Plot(label="Forecast Chart")
         ],
         title="Labour Attendance Forecast",
-        description="Upload a CSV file with columns: Date, Attendance, Trade, Weather, Alert_Status, Shortage_Risk (e.g. 22%), Suggested_Actions. The file must contain data for at least 10 trades. "
+        description="Upload a CSV file with columns: Date, Attendance, Trade, Weather, Alert_Status, Shortage_Risk (e.g. 22%), Suggested_Actions. The file must contain data for at least 10 trades."
     ).launch(share=False)
 
 if __name__ == '__main__':
-    gradio_interface()
+    gradio_interface()