feat(ai): smarter batch analysis with revenue, history, nomenclature and strict retail rules

src/features/grid/components/bulk-ai-analyzer.tsx

@@ -47,13 +47,18 @@ export function BulkAiAnalyzer() {
             for (const chunk of chunks) {
                 setProgress(prev => ({ ...prev, current: completed, total: chunks.length, message: `Analyse du rayon: ${chunk.rayon}` }));
 
-                // Simplify payload to save tokens
+                // Enrich payload with more context to improve AI decision quality
                 const payloadProducts = chunk.items.map(r => ({
                     codein: r.codein,
-                    gtin: r.gtin,
                     nom: r.libelle1,
-                    ventes: r.totalQuantite,
+                    ca: r.totalCa || 0,
+                    ventes: r.totalQuantite || 0,
                     marge: r.totalMarge ? parseFloat(((r.totalMarge / (r.totalCa || 1)) * 100).toFixed(1)) : 0,
+                    gammeInit: r.codeGammeInit || "N/A",
+                    // Simplified history: last 12 values joined (e.g., "5,0,10...")
+                    historique: Object.values(r.sales12m || {}).join(","),
+                    // Contextual nomenclature
+                    nomenclature: `${r.libelleNiveau1 || ""} > ${r.libelleNiveau2 || ""} > ${r.libelle3 || ""}`
                 }));
 
                 try {