Update core/plot.py

core/plot.py

@@ -1,3 +1,4 @@
+# core/plot.py
 import plotly.graph_objects as go
 import plotly.express as px
 import pandas as pd
@@ -39,20 +40,20 @@ def plot_indicators(df, ticker):
             logging.warning("Missing columns for candlestick: 'Date', 'Open', 'High', 'Low', 'value'")
 
         for ma in ['sma_10', 'sma_20', 'sma_50', 'ema_12', 'ema_26', 'ema_50']:
-            if ma in df:
+            if ma in df and not df[ma].isna().all():
                 logging.debug(f"Adding {ma} trace")
                 fig.add_trace(
                     go.Scatter(x=df['Date'], y=df[ma], name=ma.upper(), line=dict(width=1.5)),
                     row=1, col=1
                 )
             else:
-                logging.warning(f"{ma} not found in DataFrame")
+                logging.warning(f"{ma} not found or all NaN in DataFrame")
 
         # Bollinger Bands
         bb_u = 'bbu_20_2.0' if 'bbu_20_2.0' in df else ('bbu_20_2' if 'bbu_20_2' in df else None)
         bb_m = 'bbm_20_2.0' if 'bbm_20_2.0' in df else ('bbm_20_2' if 'bbm_20_2' in df else None)
         bb_l = 'bbl_20_2.0' if 'bbl_20_2.0' in df else ('bbl_20_2' if 'bbl_20_2' in df else None)
-        if bb_u and bb_m and bb_l:
+        if bb_u and bb_m and bb_l and not df[bb_u].isna().all() and not df[bb_m].isna().all() and not df[bb_l].isna().all():
             logging.debug("Adding Bollinger Bands traces")
             fig.add_trace(
                 go.Scatter(x=df['Date'], y=df[bb_u], name='BB Upper', line=dict(color='gray', dash='dot')),
@@ -67,40 +68,52 @@ def plot_indicators(df, ticker):
                 row=1, col=1
             )
         else:
-            logging.warning(f"Bollinger Bands columns missing: {bb_u}, {bb_m}, {bb_l}")
+            logging.warning(f"Bollinger Bands columns missing or all NaN: {bb_u}, {bb_m}, {bb_l}")
 
         # Signals
-        if 'Signal' in df:
+        if 'Signal' in df and not df['Signal'].isna().all():
             logging.debug("Adding signal traces")
             buy_signals = df[df['Signal'] == 'Buy']
             sell_signals = df[df['Signal'] == 'Sell']
             hold_signals = df[df['Signal'] == 'Hold']
-            fig.add_trace(
-                go.Scatter(
-                    x=buy_signals['Date'], y=buy_signals['value'], mode='markers+text',
-                    name='Buy', marker=dict(symbol='triangle-up', size=12, color='green'),
-                    text=['Buy'] * len(buy_signals), textposition='top center'
-                ), row=1, col=1
-            )
-            fig.add_trace(
-                go.Scatter(
-                    x=sell_signals['Date'], y=sell_signals['value'], mode='markers+text',
-                    name='Sell', marker=dict(symbol='triangle-down', size=12, color='red'),
-                    text=['Sell'] * len(sell_signals), textposition='bottom center'
-                ), row=1, col=1
-            )
-            fig.add_trace(
-                go.Scatter(
-                    x=hold_signals['Date'], y=hold_signals['value'], mode='markers',
-                    name='Hold', marker=dict(symbol='circle', size=8, color='gray'),
-                    opacity=0.5
-                ), row=1, col=1
-            )
+            if not buy_signals.empty:
+                logging.debug(f"Adding {len(buy_signals)} buy signal markers")
+                fig.add_trace(
+                    go.Scatter(
+                        x=buy_signals['Date'], y=buy_signals['value'], mode='markers+text',
+                        name='Buy', marker=dict(symbol='triangle-up', size=12, color='green'),
+                        text=['Buy'] * len(buy_signals), textposition='top center'
+                    ), row=1, col=1
+                )
+            else:
+                logging.warning("No buy signals found")
+            if not sell_signals.empty:
+                logging.debug(f"Adding {len(sell_signals)} sell signal markers")
+                fig.add_trace(
+                    go.Scatter(
+                        x=sell_signals['Date'], y=sell_signals['value'], mode='markers+text',
+                        name='Sell', marker=dict(symbol='triangle-down', size=12, color='red'),
+                        text=['Sell'] * len(sell_signals), textposition='bottom center'
+                    ), row=1, col=1
+                )
+            else:
+                logging.warning("No sell signals found")
+            if not hold_signals.empty:
+                logging.debug(f"Adding {len(hold_signals)} hold signal markers")
+                fig.add_trace(
+                    go.Scatter(
+                        x=hold_signals['Date'], y=hold_signals['value'], mode='markers',
+                        name='Hold', marker=dict(symbol='circle', size=8, color='gray'),
+                        opacity=0.5
+                    ), row=1, col=1
+                )
+            else:
+                logging.warning("No hold signals found")
         else:
-            logging.warning("Signal column not found in DataFrame")
+            logging.warning("Signal column not found or all NaN in DataFrame")
 
         # Position Size and Risk Annotation
-        if 'atr_14' in df:
+        if 'atr_14' in df and not df['atr_14'].isna().all():
             atr = df['atr_14'].iloc[-1]
             stop_distance = atr * 2
             position_size = (10000 * 0.01) / stop_distance if stop_distance != 0 else 0
@@ -111,20 +124,20 @@ def plot_indicators(df, ticker):
                 font=dict(color="black", size=12)
             )
         else:
-            logging.warning("atr_14 not found for position size annotation")
+            logging.warning("atr_14 not found or all NaN for position size annotation")
 
         # Volume
-        if 'Volume' in df:
+        if 'Volume' in df and not df['Volume'].isna().all():
             logging.debug("Adding volume trace")
             fig.add_trace(
                 go.Bar(x=df['Date'], y=df['Volume'], name='Volume', marker_color='blue', opacity=0.5),
                 row=2, col=1
             )
         else:
-            logging.warning("Volume column not found in DataFrame")
+            logging.warning("Volume column not found or all NaN in DataFrame")
 
         # MACD & RSI
-        if 'macd_12_26_9' in df and 'macds_12_26_9' in df and 'macdh_12_26_9' in df:
+        if all(col in df for col in ['macd_12_26_9', 'macds_12_26_9', 'macdh_12_26_9']) and not df['macd_12_26_9'].isna().all():
             logging.debug("Adding MACD traces")
             fig.add_trace(
                 go.Scatter(x=df['Date'], y=df['macd_12_26_9'], name='MACD', line=dict(color='blue')),
@@ -139,9 +152,9 @@ def plot_indicators(df, ticker):
                 row=3, col=1
             )
         else:
-            logging.warning("MACD columns (macd_12_26_9, macds_12_26_9, macdh_12_26_9) not found")
+            logging.warning("MACD columns (macd_12_26_9, macds_12_26_9, macdh_12_26_9) not found or all NaN")
 
-        if 'rsi_14' in df:
+        if 'rsi_14' in df and not df['rsi_14'].isna().all():
             logging.debug("Adding RSI 14 trace")
             fig.add_trace(
                 go.Scatter(x=df['Date'], y=df['rsi_14'], name='RSI 14', line=dict(color='purple')),
@@ -152,20 +165,20 @@ def plot_indicators(df, ticker):
             fig.add_trace(go.Scatter(x=[df['Date'].min(), df['Date'].max()], y=[30, 30],
                                     showlegend=False, line=dict(color='green', dash='dash', width=1)), row=3, col=1)
         else:
-            logging.warning("rsi_14 not found in DataFrame")
+            logging.warning("rsi_14 not found or all NaN in DataFrame")
 
         for rsi in ['rsi_21', 'rsi_50']:
-            if rsi in df:
+            if rsi in df and not df[rsi].isna().all():
                 logging.debug(f"Adding {rsi} trace")
                 fig.add_trace(
                     go.Scatter(x=df['Date'], y=df[rsi], name=rsi.upper(), line=dict(color='magenta' if rsi == 'rsi_21' else 'cyan', dash='dash' if rsi == 'rsi_21' else 'dot')),
                     row=3, col=1
                 )
             else:
-                logging.warning(f"{rsi} not found in DataFrame")
+                logging.warning(f"{rsi} not found or all NaN in DataFrame")
 
         # Stochastic & Williams %R
-        if 'stochk_14_3_3' in df and 'stochd_14_3_3' in df:
+        if all(col in df for col in ['stochk_14_3_3', 'stochd_14_3_3']) and not df['stochk_14_3_3'].isna().all():
             logging.debug("Adding Stochastic traces")
             fig.add_trace(
                 go.Scatter(x=df['Date'], y=df['stochk_14_3_3'], name='Stoch %K', line=dict(color='blue')),
@@ -180,9 +193,9 @@ def plot_indicators(df, ticker):
             fig.add_trace(go.Scatter(x=[df['Date'].min(), df['Date'].max()], y=[20, 20],
                                     showlegend=False, line=dict(color='green', dash='dash', width=1)), row=4, col=1)
         else:
-            logging.warning("Stochastic columns (stochk_14_3_3, stochd_14_3_3) not found")
+            logging.warning("Stochastic columns (stochk_14_3_3, stochd_14_3_3) not found or all NaN")
 
-        if 'willr_14' in df:
+        if 'willr_14' in df and not df['willr_14'].isna().all():
             logging.debug("Adding Williams %R trace")
             fig.add_trace(
                 go.Scatter(x=df['Date'], y=df['willr_14'], name='Williams %R', line=dict(color='green')),
@@ -193,10 +206,10 @@ def plot_indicators(df, ticker):
             fig.add_trace(go.Scatter(x=[df['Date'].min(), df['Date'].max()], y=[-80, -80],
                                     showlegend=False, line=dict(color='green', dash='dash', width=1)), row=4, col=1)
         else:
-            logging.warning("willr_14 not found in DataFrame")
+            logging.warning("willr_14 not found or all NaN in DataFrame")
 
         # ADX & DI
-        if all(col in df for col in ['adx_14', 'pdi_14', 'mdi_14']):
+        if all(col in df for col in ['adx_14', 'pdi_14', 'mdi_14']) and not df['adx_14'].isna().all():
             logging.debug("Adding ADX and DI traces")
             fig.add_trace(
                 go.Scatter(x=df['Date'], y=df['adx_14'], name='ADX', line=dict(color='blue')),
@@ -213,19 +226,19 @@ def plot_indicators(df, ticker):
             fig.add_trace(go.Scatter(x=[df['Date'].min(), df['Date'].max()], y=[25, 25],
                                     showlegend=False, line=dict(color='black', dash='dash', width=1)), row=5, col=1)
         else:
-            logging.warning("ADX/DI columns (adx_14, pdi_14, mdi_14) not found")
+            logging.warning("ADX/DI columns (adx_14, pdi_14, mdi_14) not found or all NaN")
 
         # ATR & CCI
-        if 'atr_14' in df:
+        if 'atr_14' in df and not df['atr_14'].isna().all():
             logging.debug("Adding ATR trace")
             fig.add_trace(
                 go.Scatter(x=df['Date'], y=df['atr_14'], name='ATR', line=dict(color='orange')),
                 row=6, col=1
             )
         else:
-            logging.warning("atr_14 not found in DataFrame")
+            logging.warning("atr_14 not found or all NaN in DataFrame")
 
-        if 'cci_20' in df:
+        if 'cci_20' in df and not df['cci_20'].isna().all():
             logging.debug("Adding CCI trace")
             fig.add_trace(
                 go.Scatter(x=df['Date'], y=df['cci_20'], name='CCI', line=dict(color='purple')),
@@ -236,10 +249,10 @@ def plot_indicators(df, ticker):
             fig.add_trace(go.Scatter(x=[df['Date'].min(), df['Date'].max()], y=[-100, -100],
                                     showlegend=False, line=dict(color='green', dash='dash', width=1)), row=6, col=1)
         else:
-            logging.warning("cci_20 not found in DataFrame")
+            logging.warning("cci_20 not found or all NaN in DataFrame")
 
         # Signal Strength
-        if all(col in df for col in ['RSI_Signal', 'MACD_Signal', 'ADX_Signal', 'Sentiment_Signal', 'Model_Signal']):
+        if all(col in df for col in ['RSI_Signal', 'MACD_Signal', 'ADX_Signal', 'Sentiment_Signal', 'Model_Signal']) and not df['RSI_Signal'].isna().all():
             logging.debug("Adding signal strength trace")
             signal_strength = (
                 df['RSI_Signal'].abs() +
@@ -257,7 +270,7 @@ def plot_indicators(df, ticker):
             fig.add_trace(go.Scatter(x=[df['Date'].min(), df['Date'].max()], y=[3, 3],
                                     showlegend=False, line=dict(color='orange', dash='dash', width=1)), row=7, col=1)
         else:
-            logging.warning("Signal strength columns (RSI_Signal, MACD_Signal, ADX_Signal, Sentiment_Signal, Model_Signal) not found")
+            logging.warning("Signal strength columns (RSI_Signal, MACD_Signal, ADX_Signal, Sentiment_Signal, Model_Signal) not found or all NaN")
 
         fig.update_layout(
             title=f"{ticker} Price and Technical Indicators",
@@ -282,11 +295,11 @@ def plot_future_forecast(df, result, timeframe):
     try:
         logging.debug(f"Starting plot_future_forecast for timeframe: {timeframe}")
         fig = go.Figure()
-        if 'Date' in df and 'value' in df:
+        if 'Date' in df and 'value' in df and not df['value'].isna().all():
             logging.debug("Adding historical close trace")
             fig.add_trace(go.Scatter(x=df['Date'], y=df['value'], name='Historical Close', line=dict(color='blue')))
         else:
-            logging.warning("Missing 'Date' or 'value' columns for historical close")
+            logging.warning("Missing 'Date' or 'value' columns or all NaN for historical close")
 
         if "latest_prediction" in result:
             last_date = df['Date'].iloc[-1]
@@ -507,12 +520,12 @@ def plot_model_architecture(result):
         dummy_input = torch.randn(1, result['arch']['window'], result['arch']['input_dim'])
         graph = make_dot(model(dummy_input), params=dict(model.named_parameters()))
         graph.format = 'png'
-        graph.render("model_arch", cleanup=True)
+        graph.render("/tmp/model_arch", cleanup=True)
         logging.debug("Model architecture graph rendered")
         fig = go.Figure()
         fig.add_layout_image(
             dict(
-                source="data:image/png;base64," + base64.b64encode(open("model_arch.png", "rb").read()).decode(),
+                source="data:image/png;base64," + base64.b64encode(open("/tmp/model_arch.png", "rb").read()).decode(),
                 xref="paper", yref="paper",
                 x=0, y=1,
                 sizex=1, sizey=1,
@@ -522,7 +535,9 @@ def plot_model_architecture(result):
         fig.update_layout(
             title="Model Architecture Graph",
             template="plotly_dark",
-            showlegend=False
+            showlegend=False,
+            height=800,
+            width=1200
         )
         logging.info("Model architecture plot generated")
         return fig
@@ -536,11 +551,11 @@ def plot_signals(signals_df, ticker):
         logging.debug(f"Signals DataFrame columns: {signals_df.columns.tolist()}")
         fig = go.Figure()
         x_col = 'Date' if 'Date' in signals_df.columns else signals_df.index
-        if 'Price' in signals_df:
+        if 'Price' in signals_df and not signals_df['Price'].isna().all():
             logging.debug("Adding price trace")
             fig.add_trace(go.Scatter(x=signals_df[x_col], y=signals_df['Price'], mode='lines', name='Price', line=dict(color='blue')))
         else:
-            logging.warning("Price column not found in signals_df")
+            logging.warning("Price column not found or all NaN in signals_df")
         buy_signals = signals_df[signals_df['Signal'] == 'Buy']
         sell_signals = signals_df[signals_df['Signal'] == 'Sell']
         if not buy_signals.empty:
@@ -575,20 +590,43 @@ def plot_backtest(result, df, ticker):
             logging.warning("Actual or forecast data missing")
             return None
         logging.debug(f"Actual length: {len(actual)}, Forecast length: {len(forecast)}")
-        last_historical_date = df['Date'].iloc[-len(actual) - 1]
-        historical_dates = df['Date'].iloc[-len(actual) - 1: -len(actual) + len(actual)]
-        forecast_dates = pd.date_range(start=last_historical_date + timedelta(days=1), periods=len(forecast))
-        historical_values = df['value'].iloc[-len(actual) - 1: -len(actual) + len(actual)]
-        logging.debug("Adding historical and forecast traces")
+        # Use the last N actual points plus forecast
+        n_historical = len(actual)
+        historical_dates = df['Date'].iloc[-n_historical:]
+        historical_values = df['value'].iloc[-n_historical:]
+        forecast_dates = pd.date_range(start=historical_dates.iloc[-1] + timedelta(days=1), periods=len(forecast), freq='D')
+        # Combine historical and forecast for continuous plot
+        all_dates = pd.concat([pd.Series(historical_dates), pd.Series(forecast_dates)]).reset_index(drop=True)
+        all_values = np.concatenate([historical_values, forecast])
+        all_types = ['Historical'] * len(historical_values) + ['Forecast'] * len(forecast)
+        plot_df = pd.DataFrame({'Date': all_dates, 'Price': all_values, 'Type': all_types})
+        logging.debug(f"Combined plot data: {plot_df.head().to_dict()}")
         fig = go.Figure()
-        fig.add_trace(go.Scatter(x=historical_dates, y=historical_values, mode='lines', name='Historical', line=dict(color='blue')))
-        fig.add_trace(go.Scatter(x=forecast_dates, y=forecast, mode='lines', name='Forecast', line=dict(color='orange', dash='dash')))
+        fig.add_trace(go.Scatter(
+            x=plot_df[plot_df['Type'] == 'Historical']['Date'],
+            y=plot_df[plot_df['Type'] == 'Historical']['Price'],
+            mode='lines',
+            name='Historical',
+            line=dict(color='blue')
+        ))
+        fig.add_trace(go.Scatter(
+            x=plot_df[plot_df['Type'] == 'Forecast']['Date'],
+            y=plot_df[plot_df['Type'] == 'Forecast']['Price'],
+            mode='lines',
+            name='Forecast',
+            line=dict(color='orange', dash='dash')
+        ))
         fig.update_layout(
             title=f"{ticker} Backtest: Historical and Prediction",
             xaxis_title="Date",
             yaxis_title="Price",
             template="plotly_dark",
-            showlegend=True
+            showlegend=True,
+            height=600,
+            xaxis=dict(tickformat="%Y-%m-%d", minor=dict(ticks="inside", showgrid=True), gridcolor="lightgrey"),
+            yaxis=dict(gridcolor="lightgrey"),
+            plot_bgcolor="white",
+            paper_bgcolor="white"
         )
         logging.info(f"Backtest plot generated for {ticker}")
         return fig