Update app.py

app.py

@@ -54,6 +54,22 @@ def load_data():
         st.error(f"Error cargando los archivos de datos: {e}")
         return None, None
 alimentos_data, foodb_index = load_data()
+COMPOUND_SYNONYM_MAP = {
+    "purinas": ["purine", "hypoxanthine", "adenine", "guanine", "xanthine"],
+    "fodmaps": ["fructan", "oligosaccharide", "polyol", "sorbitol", "mannitol"],
+    "aminas": ["amine", "histamine", "tyramine", "phenylethylamine", "putrescine", "cadaverine"],
+    "gluten": ["gluten", "gliadin"],
+    "caseína": ["casein", "casomorphin"],
+    "lactosa": ["lactose"],
+    "fructosa": ["fructose"],
+    "oxalatos": ["oxalate", "oxalic acid"],
+    "salicilatos": ["salicylate", "salicylic acid"],
+    "lectinas": ["lectin", "agglutinin"],
+    "sulfitos": ["sulfite", "sulphite"],
+    "glutamato": ["glutamate", "glutamic acid"],
+    # Añade más sinónimos aquí según sea necesario
+}
+
 KNOWN_TRIGGERS_MAP = {
     # --- Aminas Vasoactivas y Biógenas ---
     "Histamine": ["dolor de cabeza", "migraña", "inflamación", "congestión nasal", "picazón", "erupción", "ansiedad", "diarrea"],
@@ -756,20 +772,20 @@ def find_best_foodb_matches(user_foods_es, foodb_index_keys, food_name_map, limi
     
     found_matches = []
     
-    # Búsqueda por prioridades
+    # Búsqueda por prioridades para encontrar varias coincidencias relevantes
+    all_possible_matches = []
     for term in sorted_terms:
         # Prioridad 1: Coincidencia perfecta
-        for key in foodb_index_keys:
-            if term == key.lower() and key not in found_matches:
-                found_matches.append(key)
+        all_possible_matches.extend([key for key in foodb_index_keys if term == key.lower()])
         # Prioridad 2: Coincidencia de inicio
-        for key in foodb_index_keys:
-            if key.lower().startswith(term) and key not in found_matches:
-                found_matches.append(key)
+        all_possible_matches.extend([key for key in foodb_index_keys if key.lower().startswith(term)])
         # Prioridad 3: Coincidencia general
-        for key in foodb_index_keys:
-            if term in key.lower() and key not in found_matches:
-                found_matches.append(key)
+        all_possible_matches.extend([key for key in foodb_index_keys if term in key.lower()])
+
+    # Eliminar duplicados manteniendo el orden
+    for match in all_possible_matches:
+        if match not in found_matches:
+            found_matches.append(match)
 
     return found_matches[:limit]
     
@@ -957,38 +973,47 @@ 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")
             
-            with col2_expander:
+with col2_expander:
                 st.write("") 
                 if foodb_index:
                     with st.popover("🔬 Componentes Moleculares del Diagnóstico"):
+                        
                         st.info("Análisis de los compuestos en los alimentos mencionados que están directamente implicados en el diagnóstico principal.")
+                        
                         user_foods_mentioned = st.session_state.entities.get("alimentos", [])
                         
                         if not user_foods_mentioned:
                             st.warning("No se identificó un alimento específico para buscar.")
                         else:
-                            main_diagnosis_symptoms = set(s.lower() for s in best_match.get("sintomas_clave", []))
-                            target_compounds = set()
+                            # --- NUEVA LÓGICA DE ENRIQUECIMIENTO DE PISTAS ---
                             
-                            target_compounds_text = best_match.get("compuesto_alimento", "").lower()
-                            cleaned_text = re.sub(r'\(.*?\)', '', target_compounds_text)
-                            direct_keywords = set(re.findall(r'\b[a-zA-Z-]+\b', cleaned_text))
+                            # 1. RECOPILACIÓN DE PISTAS INICIALES
+                            initial_clues = set()
                             
-                            generic_words = {'y', 'o', 'en', 'con', 'lácteos', 'carnes', 'pescados'}
-                            direct_keywords -= generic_words
-                            
-                            if direct_keywords:
-                                target_compounds.update(direct_keywords)
-                            
-                            if not target_compounds:
-                                for compound, triggered_symptoms in KNOWN_TRIGGERS_MAP.items():
-                                    if main_diagnosis_symptoms.intersection(triggered_symptoms):
-                                        target_compounds.add(compound.lower())
+                            # Pista Directa: del campo 'compuesto_alimento' del diagnóstico
+                            direct_text = best_match.get("compuesto_alimento", "").lower()
+                            cleaned_text = re.sub(r'\(.*?\)', '', direct_text)
+                            initial_clues.update(re.findall(r'\b[a-zA-Z-]+\b', cleaned_text))
+
+                            # Pistas Indirectas: de los síntomas del diagnóstico via KNOWN_TRIGGERS_MAP
+                            main_diagnosis_symptoms = set(s.lower() for s in best_match.get("sintomas_clave", []))
+                            for compound, triggered_symptoms in KNOWN_TRIGGERS_MAP.items():
+                                if main_diagnosis_symptoms.intersection(triggered_symptoms):
+                                    initial_clues.add(compound.lower())
+
+                            # 2. ENRIQUECIMIENTO DE PISTAS con sinónimos moleculares
+                            final_search_keywords = set()
+                            for clue in initial_clues:
+                                final_search_keywords.add(clue) # Añadir la pista original
+                                if clue in COMPOUND_SYNONYM_MAP:
+                                    final_search_keywords.update(COMPOUND_SYNONYM_MAP[clue])
                             
-                            if not target_compounds:
+                            if not final_search_keywords:
                                 st.warning(f"No se pudieron determinar los compuestos moleculares clave para '{best_match.get('condicion_asociada')}'.")
                             else:
+                                logger.info(f"Buscando compuestos moleculares con las palabras clave: {final_search_keywords}")
                                 best_food_matches = find_best_foodb_matches(user_foods_mentioned, foodb_index.keys(), FOOD_NAME_TO_FOODB_KEY)
+                                
                                 if not best_food_matches:
                                     st.warning("No se encontraron datos moleculares para los alimentos específicos mencionados.")
                                 else:
@@ -998,19 +1023,23 @@ if st.session_state.search_results is not None:
                                         relevant_compounds = []
                                         for item in compounds_data:
                                             compound_name_lower = item['compound'].lower()
-                                            if any(target in compound_name_lower for target in target_compounds):
+                                            if any(target in compound_name_lower for target in final_search_keywords):
                                                 relevant_compounds.append(item)
+                                        
                                         if relevant_compounds:
                                             found_any_data = True
                                             with st.container(border=True):
                                                 st.subheader(f"Análisis de: {food_key.capitalize()}")
                                                 st.markdown("###### 🔬 Compuestos Relevantes para el Diagnóstico:")
-                                                for item in relevant_compounds[:5]:
-                                                    st.write(f"**Compuesto:** {item['compound']}")
-                                                    st.caption(f"Este compuesto está directamente relacionado con '{best_match.get('condicion_asociada')}'.")
+                                                unique_compounds_shown = set()
+                                                for item in relevant_compounds:
+                                                    if item['compound'] not in unique_compounds_shown:
+                                                        st.write(f"**Compuesto:** {item['compound']}")
+                                                        st.caption(f"Este compuesto está directamente relacionado con '{best_match.get('condicion_asociada')}'.")
+                                                        unique_compounds_shown.add(item['compound'])
+                                    
                                     if not found_any_data:
                                         st.warning(f"No se encontraron los compuestos específicos de '{best_match.get('condicion_asociada')}' en los alimentos analizados en la base de datos FoodB.")
-            
             st.markdown("---")
             with st.container(border=True):
                 st.markdown("##### 🧠 Posibles Efectos Neuropsicológicos de los Componentes")