Update src/modules/Forcasting.py

src/modules/Forcasting.py

@@ -0,0 +1,625 @@
+import streamlit as st
+import pandas as pd
+import sys
+import os
+import plotly.graph_objects as go
+from plotly.subplots import make_subplots
+import numpy as np
+
+# Import du module d'analyse
+sys.path.append(os.path.join(os.path.dirname(__file__), '..'))
+from Analytics.VortexOutFlux import VortexOutFlux
+
+# === CSS SPÉCIFIQUE MODULE FORECASTING ===
+def apply_forecasting_styles():
+    st.markdown("""
+    <style>
+    /* === STYLES SPÉCIFIQUES MODULE FORECASTING === */
+    
+    /* Wrapper pour isolation */
+    #forecasting-module {
+        font-family: 'Space Grotesk', sans-serif;
+    }
+    
+    /* Headers spécifiques */
+    #forecasting-module h1 {
+        font-size: 1.8rem !important;
+        margin-bottom: 24px !important;
+        color: #58a6ff !important;
+        border-bottom: 2px solid rgba(88, 166, 255, 0.3);
+        padding-bottom: 12px;
+    }
+    
+    #forecasting-module h2 {
+        font-size: 1.4rem !important;
+        margin-bottom: 16px !important;
+        color: #8b949e !important;
+    }
+    
+    #forecasting-module h3 {
+        font-size: 1.1rem !important;
+        margin-bottom: 12px !important;
+        color: #58a6ff !important;
+    }
+    
+    /* Metrics cards pour KPIs */
+    #forecasting-module [data-testid="stMetric"] {
+        background: rgba(22, 27, 34, 0.7);
+        border: 1px solid rgba(88, 166, 255, 0.4);
+        padding: 16px !important;
+        border-radius: 6px;
+        box-shadow: 0 2px 8px rgba(0, 0, 0, 0.3);
+    }
+    
+    #forecasting-module [data-testid="stMetric"] label {
+        color: #8b949e !important;
+        font-size: 0.75rem !important;
+        font-weight: 600 !important;
+        text-transform: uppercase;
+        letter-spacing: 0.8px;
+    }
+    
+    #forecasting-module [data-testid="stMetric"] [data-testid="stMetricValue"] {
+        color: #58a6ff !important;
+        font-size: 1.6rem !important;
+        font-weight: 700 !important;
+    }
+    
+    #forecasting-module [data-testid="stMetric"] [data-testid="stMetricDelta"] {
+        color: #54bd4b !important;
+        font-size: 0.9rem !important;
+    }
+    
+    /* Carte de prédiction style Gotham */
+    .forecast-card {
+        background: linear-gradient(135deg, rgba(22, 27, 34, 0.9) 0%, rgba(13, 17, 23, 0.9) 100%);
+        border: 2px solid rgba(88, 166, 255, 0.3);
+        border-radius: 8px;
+        padding: 20px;
+        margin: 16px 0;
+        transition: all 0.3s ease;
+        position: relative;
+        overflow: hidden;
+    }
+    
+    .forecast-card:hover {
+        border-color: rgba(88, 166, 255, 0.6);
+        transform: translateY(-2px);
+        box-shadow: 0 8px 16px rgba(0, 0, 0, 0.4);
+    }
+    
+    .forecast-card::before {
+        content: '';
+        position: absolute;
+        top: 0;
+        left: 0;
+        right: 0;
+        height: 3px;
+        background: linear-gradient(90deg, #58a6ff, #54bd4b);
+    }
+    
+    .forecast-card-header {
+        display: flex;
+        justify-content: space-between;
+        align-items: center;
+        margin-bottom: 16px;
+        padding-bottom: 12px;
+        border-bottom: 1px solid rgba(88, 166, 255, 0.2);
+    }
+    
+    .forecast-card-title {
+        font-size: 1.2rem;
+        font-weight: 700;
+        color: #58a6ff;
+        margin: 0;
+    }
+    
+    .forecast-card-badge {
+        background: rgba(88, 166, 255, 0.2);
+        color: #58a6ff;
+        padding: 4px 12px;
+        border-radius: 12px;
+        font-size: 0.75rem;
+        font-weight: 600;
+        text-transform: uppercase;
+    }
+    
+    .forecast-card-badge.positive {
+        background: rgba(84, 189, 75, 0.2);
+        color: #54bd4b;
+    }
+    
+    .forecast-card-badge.negative {
+        background: rgba(243, 156, 18, 0.2);
+        color: #f39c12;
+    }
+    
+    .forecast-card-body {
+        color: #c9d1d9;
+    }
+    
+    .forecast-card-row {
+        display: flex;
+        justify-content: space-between;
+        padding: 10px 0;
+        border-bottom: 1px solid rgba(48, 54, 61, 0.4);
+    }
+    
+    .forecast-card-row:last-child {
+        border-bottom: none;
+    }
+    
+    .forecast-card-label {
+        color: #8b949e;
+        font-size: 0.9rem;
+        font-weight: 500;
+    }
+    
+    .forecast-card-value {
+        color: #c9d1d9;
+        font-weight: 600;
+        font-size: 1rem;
+    }
+    
+    .forecast-card-value.highlight {
+        color: #58a6ff;
+        font-size: 1.2rem;
+    }
+    
+    /* Bloc d'interprétation */
+    .interpretation-box {
+        background: rgba(88, 166, 255, 0.05);
+        border-left: 4px solid #58a6ff;
+        border-radius: 4px;
+        padding: 16px;
+        margin: 16px 0;
+    }
+    
+    .interpretation-box h4 {
+        color: #58a6ff !important;
+        margin: 0 0 12px 0 !important;
+        font-size: 1rem !important;
+    }
+    
+    .interpretation-box p {
+        color: #c9d1d9;
+        margin: 8px 0;
+        line-height: 1.6;
+        font-size: 0.9rem;
+    }
+    
+    /* Bloc métriques de performance */
+    .metrics-performance {
+        background: rgba(22, 27, 34, 0.6);
+        border: 1px solid rgba(48, 54, 61, 0.8);
+        border-radius: 6px;
+        padding: 16px;
+        margin: 16px 0;
+    }
+    
+    .metrics-grid {
+        display: grid;
+        grid-template-columns: repeat(auto-fit, minmax(200px, 1fr));
+        gap: 16px;
+        margin-top: 12px;
+    }
+    
+    .metric-item {
+        background: rgba(13, 17, 23, 0.8);
+        padding: 12px;
+        border-radius: 4px;
+        border-left: 3px solid #58a6ff;
+    }
+    
+    .metric-item-label {
+        color: #8b949e;
+        font-size: 0.75rem;
+        text-transform: uppercase;
+        letter-spacing: 0.5px;
+        margin-bottom: 4px;
+    }
+    
+    .metric-item-value {
+        color: #58a6ff;
+        font-size: 1.3rem;
+        font-weight: 700;
+    }
+    
+    /* Divider */
+    #forecasting-module hr {
+        background: rgba(88, 166, 255, 0.3);
+        height: 2px;
+        margin: 24px 0;
+    }
+    
+    /* Alert boxes */
+    #forecasting-module .stAlert {
+        border-radius: 6px;
+        padding: 12px 16px !important;
+    }
+    
+    #forecasting-module .stAlert[kind="info"] {
+        background: rgba(88, 166, 255, 0.1) !important;
+        border-left: 4px solid #58a6ff !important;
+    }
+    
+    #forecasting-module .stAlert[kind="warning"] {
+        background: rgba(243, 156, 18, 0.1) !important;
+        border-left: 4px solid #f39c12 !important;
+    }
+    
+    #forecasting-module .stAlert[kind="success"] {
+        background: rgba(84, 189, 75, 0.1) !important;
+        border-left: 4px solid #54bd4b !important;
+    }
+    
+    /* Expanders */
+    #forecasting-module .streamlit-expanderHeader {
+        background: rgba(22, 27, 34, 0.7) !important;
+        border-left: 3px solid rgba(88, 166, 255, 0.6) !important;
+        padding: 12px 16px !important;
+        font-weight: 600 !important;
+        color: #8b949e !important;
+    }
+    
+    #forecasting-module .streamlit-expanderHeader:hover {
+        background: rgba(33, 38, 45, 0.9) !important;
+        border-left-color: rgba(88, 166, 255, 0.9) !important;
+    }
+    
+    #forecasting-module .streamlit-expanderContent {
+        background: rgba(13, 17, 23, 0.7);
+        border: 1px solid rgba(48, 54, 61, 0.6);
+        padding: 16px !important;
+    }
+    
+    /* Animation de chargement */
+    @keyframes pulse {
+        0%, 100% { opacity: 1; }
+        50% { opacity: 0.5; }
+    }
+    
+    .loading-indicator {
+        animation: pulse 2s ease-in-out infinite;
+        color: #58a6ff;
+        text-align: center;
+        padding: 20px;
+    }
+    </style>
+    """, unsafe_allow_html=True)
+
+def render_forecast_card(month_data):
+    """Génère une carte de prédiction mensuelle style Gotham"""
+    
+    date_str = month_data['Date'].strftime("%B %Y")
+    montant = month_data['Montant_Predit']
+    lower = month_data['Borne_Inf']
+    upper = month_data['Borne_Sup']
+    
+    # Calcul de la marge d'erreur
+    margin = ((upper - lower) / 2) / montant * 100 if montant > 0 else 0
+    
+    html = f"""
+    <div class="forecast-card">
+        <div class="forecast-card-header">
+            <h4 class="forecast-card-title">{date_str}</h4>
+            <span class="forecast-card-badge">Prédiction</span>
+        </div>
+        <div class="forecast-card-body">
+            <div class="forecast-card-row">
+                <span class="forecast-card-label">Montant Prédit</span>
+                <span class="forecast-card-value highlight">{montant:,.0f} XOF</span>
+            </div>
+            <div class="forecast-card-row">
+                <span class="forecast-card-label">Intervalle de Confiance (95%)</span>
+                <span class="forecast-card-value">{lower:,.0f} - {upper:,.0f} XOF</span>
+            </div>
+            <div class="forecast-card-row">
+                <span class="forecast-card-label">Marge d'Erreur</span>
+                <span class="forecast-card-value">± {margin:.1f}%</span>
+            </div>
+        </div>
+    </div>
+    """
+    return html
+
+def show_forecasting_module(client, sheet_name):
+    """Module principal de prévision des flux de sortie"""
+    
+    # Appliquer les styles
+    apply_forecasting_styles()
+    
+    # Wrapper pour isolation
+    st.markdown('<div id="forecasting-module">', unsafe_allow_html=True)
+    
+    st.header("JASMINE - PRÉDICTION DES FLUX DE SORTIE")
+    st.caption("Analyse prédictive en temps réel basée sur la régression linéaire")
+    
+    try:
+        # Connexion à Google Sheets
+        sh = client.open(sheet_name)
+        ws_forecasting = sh.worksheet("Forecasting")
+        
+        # Chargement des données
+        df_forecasting = pd.DataFrame(ws_forecasting.get_all_records())
+        
+        if df_forecasting.empty or len(df_forecasting) < 2:
+            st.warning("⚠️ Données insuffisantes pour effectuer une prédiction. Minimum 2 points de données requis.")
+            st.markdown('</div>', unsafe_allow_html=True)
+            return
+        
+        # Vérification des colonnes
+        if 'Date' not in df_forecasting.columns or 'Montant_Total_Sortie' not in df_forecasting.columns:
+            st.error("❌ Structure de données invalide. Colonnes requises : 'Date', 'Montant_Total_Sortie'")
+            st.markdown('</div>', unsafe_allow_html=True)
+            return
+        
+        # Initialisation de l'analyseur
+        analyzer = VortexOutFlux(df_forecasting)
+        
+        # Section 1 : Vue d'ensemble des données
+        st.divider()
+        st.subheader("📊 Vue d'Ensemble des Données Historiques")
+        
+        col1, col2, col3, col4 = st.columns(4)
+        
+        with col1:
+            nb_points = len(analyzer.df)
+            st.metric("Points de Données", nb_points)
+        
+        with col2:
+            flux_moyen = analyzer.df['Montant_Total_Sortie'].mean()
+            st.metric("Flux Moyen", f"{flux_moyen:,.0f} XOF")
+        
+        with col3:
+            flux_total = analyzer.df['Montant_Total_Sortie'].sum()
+            st.metric("Flux Total", f"{flux_total:,.0f} XOF")
+        
+        with col4:
+            date_debut = analyzer.df['Date'].min().strftime("%m/%Y")
+            date_fin = analyzer.df['Date'].max().strftime("%m/%Y")
+            st.metric("Période", f"{date_debut} - {date_fin}")
+        
+        # Section 2 : Prédictions
+        st.divider()
+        st.subheader("🔮 Prédictions pour les 3 Prochains Mois")
+        
+        with st.spinner("Calcul des prédictions en cours..."):
+            predictions_result = analyzer.predict_next_months(n_months=3)
+        
+        if 'error' in predictions_result:
+            st.error(f"❌ {predictions_result['error']}")
+            st.markdown('</div>', unsafe_allow_html=True)
+            return
+        
+        predictions_df = predictions_result['predictions']
+        models = predictions_result['models']
+        
+        # Affichage des cartes de prédiction
+        cols = st.columns(3)
+        for idx, (_, row) in enumerate(predictions_df.iterrows()):
+            with cols[idx]:
+                st.markdown(render_forecast_card(row), unsafe_allow_html=True)
+        
+        # Section 3 : Analyse de Tendance et Interprétation
+        st.divider()
+        st.subheader("📈 Analyse de Tendance")
+        
+        interpretations = analyzer.get_interpretation(models)
+        
+        # Affichage de la tendance principale
+        tendance_badge_class = "positive" if interpretations['tendance'] == "CROISSANCE" else "negative"
+        
+        st.markdown(f"""
+        <div class="forecast-card">
+            <div class="forecast-card-header">
+                <h4 class="forecast-card-title">Tendance Détectée</h4>
+                <span class="forecast-card-badge {tendance_badge_class}">{interpretations['tendance']}</span>
+            </div>
+            <div class="forecast-card-body">
+                <div class="forecast-card-row">
+                    <span class="forecast-card-label">Variation Mensuelle Moyenne</span>
+                    <span class="forecast-card-value highlight">{abs(interpretations['pente_mensuelle']):,.0f} XOF/mois ({interpretations['pct_variation']:.1f}%)</span>
+                </div>
+                <div class="forecast-card-row">
+                    <span class="forecast-card-label">Force de la Corrélation</span>
+                    <span class="forecast-card-value">{interpretations['force_correlation'].upper()}</span>
+                </div>
+                <div class="forecast-card-row">
+                    <span class="forecast-card-label">Qualité du Modèle</span>
+                    <span class="forecast-card-value">{interpretations['qualite_modele'].upper()}</span>
+                </div>
+            </div>
+        </div>
+        """, unsafe_allow_html=True)
+        
+        # Interprétation textuelle
+        st.markdown(f"""
+        <div class="interpretation-box">
+            <h4>💡 Interprétation</h4>
+            <p><strong>Tendance :</strong> {interpretations['message_principal']}</p>
+            <p><strong>Fiabilité :</strong> {interpretations['message_fiabilite']}</p>
+            <p><strong>Précision :</strong> {interpretations['message_precision']}</p>
+        </div>
+        """, unsafe_allow_html=True)
+        
+        # Section 4 : Métriques de Performance du Modèle
+        st.divider()
+        st.subheader("🎯 Performance du Modèle")
+        
+        col1, col2, col3, col4, col5 = st.columns(5)
+        
+        with col1:
+            st.metric("R² (R-Carré)", f"{models['r_squared']:.3f}", 
+                     help="Coefficient de détermination - Pourcentage de variance expliquée")
+        
+        with col2:
+            st.metric("R (Corrélation)", f"{models['r_coefficient']:.3f}",
+                     help="Coefficient de corrélation")
+        
+        with col3:
+            st.metric("MAE", f"{models['mae']:,.0f} XOF",
+                     help="Mean Absolute Error - Erreur moyenne absolue")
+        
+        with col4:
+            st.metric("RMSE", f"{models['rmse']:,.0f} XOF",
+                     help="Root Mean Squared Error - Erreur quadratique moyenne")
+        
+        with col5:
+            pente_jour = models['pente']
+            st.metric("Pente", f"{pente_jour:.2f} XOF/jour",
+                     help="Croissance par jour ordinal")
+        
+        # Section 5 : Visualisations
+        st.divider()
+        st.subheader("📉 Visualisations Analytiques")
+        
+        # Préparation des données de visualisation
+        viz_data = analyzer.generate_visualization_data(predictions_result)
+        
+        # Graphique principal : Historique + Prédictions
+        fig_main = go.Figure()
+        
+        # Données historiques
+        fig_main.add_trace(go.Scatter(
+            x=viz_data['historical']['dates'],
+            y=viz_data['historical']['values'],
+            mode='lines+markers',
+            name='Flux Historiques',
+            line=dict(color='#58a6ff', width=2),
+            marker=dict(size=8, symbol='circle')
+        ))
+        
+        # Prédictions futures
+        fig_main.add_trace(go.Scatter(
+            x=viz_data['future']['dates'],
+            y=viz_data['future']['predictions'],
+            mode='lines+markers',
+            name='Prédictions',
+            line=dict(color='#54bd4b', width=2, dash='dash'),
+            marker=dict(size=10, symbol='diamond')
+        ))
+        
+        # Intervalle de confiance
+        fig_main.add_trace(go.Scatter(
+            x=viz_data['future']['dates'] + viz_data['future']['dates'][::-1],
+            y=viz_data['future']['upper_bound'] + viz_data['future']['lower_bound'][::-1],
+            fill='toself',
+            fillcolor='rgba(84, 189, 75, 0.2)',
+            line=dict(color='rgba(84, 189, 75, 0)'),
+            name='Intervalle de Confiance (95%)',
+            showlegend=True
+        ))
+        
+        fig_main.update_layout(
+            title=dict(
+                text="Flux de Sortie Mensuels : Historique & Prédictions",
+                font=dict(size=16, color='#58a6ff')
+            ),
+            xaxis=dict(
+                title="Date",
+                gridcolor='rgba(48, 54, 61, 0.3)',
+                color='#8b949e'
+            ),
+            yaxis=dict(
+                title="Montant (XOF)",
+                gridcolor='rgba(48, 54, 61, 0.3)',
+                color='#8b949e'
+            ),
+            plot_bgcolor='rgba(13, 17, 23, 0.8)',
+            paper_bgcolor='rgba(22, 27, 34, 0.9)',
+            font=dict(color='#c9d1d9', family='Space Grotesk'),
+            hovermode='x unified',
+            legend=dict(
+                bgcolor='rgba(22, 27, 34, 0.8)',
+                bordercolor='rgba(88, 166, 255, 0.3)',
+                borderwidth=1
+            ),
+            height=500
+        )
+        
+        st.plotly_chart(fig_main, use_container_width=True)
+        
+        # Graphiques secondaires : Résidus
+        with st.expander("📊 Analyse des Résidus (Diagnostic du Modèle)", expanded=False):
+            fig_residuals = make_subplots(
+                rows=1, cols=2,
+                subplot_titles=("Graphique des Résidus", "Distribution des Résidus"),
+                specs=[[{"type": "scatter"}, {"type": "histogram"}]]
+            )
+            
+            # Graphique des résidus
+            fig_residuals.add_trace(
+                go.Scatter(
+                    x=viz_data['residuals']['predictions'],
+                    y=viz_data['residuals']['values'],
+                    mode='markers',
+                    marker=dict(color='#58a6ff', size=8),
+                    name='Résidus'
+                ),
+                row=1, col=1
+            )
+            
+            # Ligne à zéro
+            fig_residuals.add_hline(
+                y=0, line_dash="dash", line_color="#f39c12",
+                row=1, col=1
+            )
+            
+            # Distribution des résidus
+            fig_residuals.add_trace(
+                go.Histogram(
+                    x=viz_data['residuals']['values'],
+                    marker=dict(color='#58a6ff', line=dict(color='#c9d1d9', width=1)),
+                    name='Distribution',
+                    nbinsx=15
+                ),
+                row=1, col=2
+            )
+            
+            fig_residuals.update_xaxes(title_text="Valeurs Prédites (XOF)", row=1, col=1, gridcolor='rgba(48, 54, 61, 0.3)', color='#8b949e')
+            fig_residuals.update_yaxes(title_text="Résidus (XOF)", row=1, col=1, gridcolor='rgba(48, 54, 61, 0.3)', color='#8b949e')
+            fig_residuals.update_xaxes(title_text="Résidus (XOF)", row=1, col=2, gridcolor='rgba(48, 54, 61, 0.3)', color='#8b949e')
+            fig_residuals.update_yaxes(title_text="Fréquence", row=1, col=2, gridcolor='rgba(48, 54, 61, 0.3)', color='#8b949e')
+            
+            fig_residuals.update_layout(
+                plot_bgcolor='rgba(13, 17, 23, 0.8)',
+                paper_bgcolor='rgba(22, 27, 34, 0.9)',
+                font=dict(color='#c9d1d9', family='Space Grotesk'),
+                showlegend=False,
+                height=400
+            )
+            
+            st.plotly_chart(fig_residuals, use_container_width=True)
+            
+            st.caption("**Note :** Les résidus doivent être aléatoirement distribués autour de zéro pour un bon modèle.")
+        
+        # Section 6 : Tableau des données
+        with st.expander("📋 Tableau Détaillé des Prédictions", expanded=False):
+            # Formatage du DataFrame pour l'affichage
+            display_df = predictions_df.copy()
+            display_df['Date'] = display_df['Date'].dt.strftime('%B %Y')
+            display_df['Montant_Predit'] = display_df['Montant_Predit'].apply(lambda x: f"{x:,.0f} XOF")
+            display_df['Borne_Inf'] = display_df['Borne_Inf'].apply(lambda x: f"{x:,.0f} XOF")
+            display_df['Borne_Sup'] = display_df['Borne_Sup'].apply(lambda x: f"{x:,.0f} XOF")
+            
+            display_df.columns = ['Date', 'Montant Prédit', 'Borne Inférieure (95%)', 'Borne Supérieure (95%)']
+            
+            st.dataframe(display_df, use_container_width=True, hide_index=True)
+        
+        # Section 7 : Informations complémentaires
+        st.divider()
+        st.info("""
+        **ℹ️ À Propos de ce Modèle**
+        
+        - **Modèle utilisé :** Régression Linéaire avec Intervalles de Confiance à 95%
+        - **Actualisation :** Les prédictions se mettent à jour automatiquement à chaque ajout de données dans la feuille "Forecasting"
+        - **Utilisation :** Ce modèle est adapté pour des tendances linéaires. Pour des patterns saisonniers complexes, envisagez des modèles avancés (SARIMA, Prophet)
+        - **Recommandation :** Vérifiez régulièrement les résidus et le R² pour assurer la qualité du modèle
+        """)
+        
+    except Exception as e:
+        st.error(f"❌ Erreur lors du chargement des données : {e}")
+        st.exception(e)
+    
+    st.markdown('</div>', unsafe_allow_html=True)