Update app.py

app.py

@@ -490,30 +490,37 @@ def extract_entities_with_gemini(query):
         logger.error(f"Error en la extracción con Gemini (puede ser reintentado): {e}")
         raise e
 
-def reinforce_entities_with_keywords(entities, query, food_map, symptom_map):
+def reinforce_entities_with_keywords(entities, query, food_map, master_symptom_map):
     if not entities:
         entities = {"alimentos": [], "sintomas": []}
     
     query_sanitized = sanitize_text(query)
     
+    # --- Parte 1: Reforzar Alimentos (sin cambios en la lógica) ---
     current_foods = entities.get("alimentos", []) or []
     current_foods_sanitized = {sanitize_text(f) for f in current_foods}
     for food_keyword in food_map.keys():
         if food_keyword in query_sanitized and food_keyword not in current_foods_sanitized:
-            logger.info(f"Red de seguridad (Alimento): Añadiendo '{food_keyword}'.")
             current_foods.append(food_keyword)
     entities["alimentos"] = list(set(current_foods))
 
+    # --- Parte 2: Normalizar y Reforzar Síntomas usando el Diccionario Maestro ---
     current_symptoms = entities.get("sintomas", []) or []
     query_to_search_symptoms = " " + query_sanitized + " "
     
-    for main_symptom, synonyms in symptom_map.items():
-        for synonym in sorted(synonyms, key=len, reverse=True):
-            if (" " + synonym + " ") in query_to_search_symptoms:
+    # Iterar sobre el diccionario maestro para encontrar coincidencias
+    for main_symptom, details in master_symptom_map.items():
+        # Ordenar las frases de la más larga a la más corta para máxima precisión
+        phrases_es = details.get("frases_es", [])
+        for phrase in sorted(phrases_es, key=len, reverse=True):
+            if (" " + phrase + " ") in query_to_search_symptoms:
+                # Si encontramos una coincidencia, añadimos el término estandarizado (la clave)
                 if main_symptom not in current_symptoms:
-                    logger.info(f"Red de seguridad (Síntoma): Normalizando '{synonym}' a '{main_symptom}'.")
+                    logger.info(f"Sistema de respaldo: Normalizando '{phrase}' a '{main_symptom}'.")
                     current_symptoms.append(main_symptom)
-                query_to_search_symptoms = query_to_search_symptoms.replace(" " + synonym + " ", " ")
+                
+                # Reemplazamos la frase encontrada para evitar sub-coincidencias
+                query_to_search_symptoms = query_to_search_symptoms.replace(" " + phrase + " ", " ")
     
     entities["sintomas"] = list(set(current_symptoms))
     return entities
@@ -723,7 +730,8 @@ if st.session_state.start_analysis:
             except Exception as e:
                 logger.warning(f"La extracción con Gemini falló; se usará el sistema de respaldo: {e}")
             
-            entities = reinforce_entities_with_keywords(entities, query_to_analyze, FOOD_TO_COMPOUND_MAP, SYMPTOM_KEYWORD_MAP)
+            # --- Aquí se usa el nuevo Diccionario Maestro ---
+            entities = reinforce_entities_with_keywords(entities, query_to_analyze, FOOD_TO_COMPOUND_MAP, MASTER_SYMPTOM_MAP)
             st.session_state.entities = entities
         
         if entities and (entities.get("alimentos") or entities.get("sintomas")):
@@ -735,6 +743,7 @@ if st.session_state.start_analysis:
         else:
             st.error("No se pudieron identificar alimentos o síntomas claros en tu descripción. Intenta ser más específico.")
             st.session_state.search_results = []
+
 if st.session_state.search_results is not None:
     results = st.session_state.search_results
     
@@ -769,18 +778,18 @@ if st.session_state.search_results is not None:
                 score_col2.metric("Puntos por Síntomas", f"{best_match_data['score']['symptoms']}")
                 score_col3.metric("PUNTUACIÓN TOTAL", f"{best_match_data['score']['total']}", delta="Máxima coincidencia")
             
-            # --- LÓGICA DEL POPOVER INTELIGENTE ---
+            # --- POPOVER INTELIGENTE USANDO EL DICCIONARIO MAESTRO ---
             with col2:
                 st.write("") 
                 if foodb_index:
                     with st.popover("🔬 Componentes Moleculares Relevantes"):
-                        st.info("Compuestos en los alimentos mencionados que podrían estar relacionados con tus síntomas (datos de FoodB en inglés).")
+                        st.info("Compuestos en los alimentos mencionados que podrían estar relacionados con tus síntomas.")
                         
                         user_foods_mentioned = st.session_state.entities.get("alimentos", [])
-                        user_symptoms = st.session_state.entities.get("sintomas", [])
+                        user_symptoms_es = st.session_state.entities.get("sintomas", [])
                         
-                        if not user_foods_mentioned:
-                            st.warning("No se identificó un alimento específico para buscar.")
+                        if not user_foods_mentioned or not user_symptoms_es:
+                            st.warning("Se necesita al menos un alimento y un síntoma para el análisis molecular.")
                         else:
                             found_relevant_data = False
                             
@@ -790,13 +799,17 @@ if st.session_state.search_results is not None:
                                     if key_es in alimento_es.lower():
                                         search_terms_en.extend(value_en_list)
                             
+                            symptom_keywords_en = []
+                            for symptom_es in user_symptoms_es:
+                                if symptom_es in MASTER_SYMPTOM_MAP:
+                                    symptom_keywords_en.extend(MASTER_SYMPTOM_MAP[symptom_es].get('keywords_en', []))
+                            
                             for foodb_key, compounds_data in foodb_index.items():
                                 if any(term in foodb_key.lower() for term in set(search_terms_en)):
-                                    
                                     relevant_compounds = []
                                     for item in compounds_data:
-                                        effects = item.get("effects", [])
-                                        if any(symptom.lower() in effect.lower() for symptom in user_symptoms for effect in effects):
+                                        effects_en = item.get("effects", [])
+                                        if any(keyword_en in effect.lower() for keyword_en in symptom_keywords_en for effect in effects_en):
                                             relevant_compounds.append(item)
                                     
                                     if relevant_compounds:
@@ -805,13 +818,13 @@ if st.session_state.search_results is not None:
                                             st.subheader(f"Análisis de: {foodb_key.capitalize()}")
                                             for item in relevant_compounds[:5]:
                                                 st.write(f"**Compuesto:** {item['compound']}")
-                                                relevant_effects = [eff for eff in item['effects'] if any(symptom.lower() in eff.lower() for symptom in user_symptoms)]
-                                                st.write(f"**Efectos Relevantes:** {', '.join(relevant_effects)}")
+                                                relevant_effects = [eff for eff in item['effects'] if any(kw in eff.lower() for kw in symptom_keywords_en)]
+                                                st.write(f"**Efectos Relevantes (en inglés):** {', '.join(relevant_effects)}")
                                             st.markdown("---")
 
                             if not found_relevant_data:
                                 st.warning("No se encontraron compuestos moleculares directamente relacionados con tus síntomas para los alimentos mencionados.")
-            
+
             st.markdown("---")
 
             with st.container(border=True):