Update app.py

app.py

@@ -44,6 +44,7 @@ def get_best_match(query, choices, limit=15):
     
 def filtrar_itens_similares(df, termo_pesquisa, limit=15):
     titulos = df['Title'].tolist()
+    # Use RapidFuzz for improved performance and fuzzy matching
     titulos_similares = get_best_match(termo_pesquisa, titulos, limit=limit)
     df_filtrado = df[df['Title'].isin(titulos_similares)]
     return df_filtrado
@@ -91,17 +92,36 @@ def select_nearest_items(df):
             break
 
     return pd.DataFrame(nearest_items)
-
+    
+def search_with_fallback(query, df, limit=15):
+    # Split the query into parts
+    query_parts = query.split()
+
+    # Start with the most specific search (full query)
+    specificities = [
+        " ".join(query_parts[i:]) for i in range(len(query_parts))
+    ]
+
+    for specificity in specificities:
+        df_filtrado = filtrar_itens_similares(df, specificity, limit=limit)
+        if not df_filtrado.empty:
+            # If we find results at this level of specificity, return them
+            return df_filtrado
+
+    # If no results are found at any level of specificity, return an empty DataFrame
+    return pd.DataFrame()
+    
 def integrated_app(query, titulo, EC, PU):
     df_mercadolibre = fetch_data_to_dataframe(query, 50, "mercadolibre")
     df_combined = pd.concat([df_mercadolibre, data_crawler], ignore_index=True)
-    
+
     if df_combined.empty:
         return "Nenhum dado encontrado. Tente uma consulta diferente.", pd.DataFrame()
 
     df_refined = refinar_resultados(df_combined)
-    df_similares = filtrar_itens_similares(df_refined, query)
-       
+    # Use the new search_with_fallback function
+    df_similares = search_with_fallback(query, df_refined)
+
     if df_similares.empty:
         return "Nenhum item similar encontrado.", pd.DataFrame()
     else: