Update rag_hf.py

rag_hf.py

@@ -33,7 +33,6 @@ st.set_page_config(
 # === CUSTOM CSS ===
 st.markdown("""
 <style>
-    /* Main styles */
     .header {
         color: #2c3e50;
         border-bottom: 2px solid #4f46e5;
@@ -61,8 +60,6 @@ st.markdown("""
         margin: 0.5rem 0;
         border: 1px solid #e5e7eb;
     }
-    
-    /* Sidebar styles */
     .sidebar-section {
         margin-bottom: 1.5rem;
     }
@@ -80,8 +77,6 @@ st.markdown("""
     .suggested-question:hover {
         background-color: #f1f5f9;
     }
-    
-    /* Metrics and badges */
     .metric-badge {
         display: inline-block;
         background-color: #e8f4fc;
@@ -120,7 +115,7 @@ def load_all_components():
         methods[label] = (matrix, id_map, G, rdf)
     return methods, embedder
 
-def get_top_k(matrix, id_map, query, k):
+def get_top_k(matrix, id_map, query, k, embedder):
     vec = embedder.encode(f"query: {query}", convert_to_tensor=True, device=DEVICE)
     vec = vec.cpu().numpy().astype("float32")
     sims = np.dot(matrix, vec) / (np.linalg.norm(matrix, axis=1) * np.linalg.norm(vec) + 1e-10)
@@ -148,8 +143,8 @@ def query_rdf(rdf, lang_id):
     except Exception as e:
         return [("error", str(e))]
 
-def generate_response(matrix, id_map, G, rdf, user_question, k=3):
-    ids = get_top_k(matrix, id_map, user_question, k)
+def generate_response(matrix, id_map, G, rdf, user_question, k, embedder):
+    ids = get_top_k(matrix, id_map, user_question, k, embedder)
     context = [get_context(G, i) for i in ids]
     rdf_facts = []
     for i in ids:
@@ -183,17 +178,14 @@ Answer:
 
 # === MAIN APP ===
 def main():
-    # Load components
     methods, embedder = load_all_components()
 
-    # Main header
     st.markdown("""
     <div class="header">
         <h1>🌍 Language Atlas: South American Indigenous Languages</h1>
     </div>
     """, unsafe_allow_html=True)
     
-    # Overview section
     with st.expander("📌 Overview", expanded=True):
         st.markdown("""
         This app provides **AI-powered analysis** of endangered indigenous languages in South America, 
@@ -226,13 +218,9 @@ def main():
             </div>
             """, unsafe_allow_html=True)
     
-    # Sidebar
     with st.sidebar:
-        # Logo and academic info
         st.markdown("### Departamento Académico de Humanidades")
         st.markdown("---")
-        
-        # Quick start guide
         st.markdown("### 🚀 Quick Start")
         st.markdown("""
         1. **Type a question** in the input box  
@@ -241,8 +229,6 @@ def main():
         """)
         
         st.markdown("---")
-        
-        # Suggested questions
         st.markdown("### 🔍 Example Queries")
         questions = [
             "What languages are endangered in Brazil?",
@@ -256,8 +242,6 @@ def main():
                 st.session_state.query = q
         
         st.markdown("---")
-        
-        # Technical details
         st.markdown("### ⚙️ Technical Details")
         st.markdown("""
         - <span class="tech-badge">Embeddings</span> Node2Vec vs. GraphSAGE  
@@ -266,8 +250,6 @@ def main():
         """, unsafe_allow_html=True)
         
         st.markdown("---")
-        
-        # Data sources
         st.markdown("### 📂 Data Sources")
         st.markdown("""
         - **Glottolog** (Language classification)  
@@ -276,18 +258,13 @@ def main():
         """)
         
         st.markdown("---")
-        
-        # Analysis parameters
         st.markdown("### 📊 Analysis Parameters")
         k = st.slider("Number of languages to analyze", 1, 10, 3)
         st.markdown("---")
-        
-        # Debug options
         st.markdown("### 🔧 Advanced Options")
         show_ctx = st.checkbox("Show context information", False)
         show_rdf = st.checkbox("Show structured facts", False)
 
-    # Main query interface
     st.markdown("### 📝 Ask About Indigenous Languages")
     query = st.text_input(
         "Enter your question:",
@@ -312,10 +289,9 @@ def main():
                 }[label])
                 
                 start = datetime.datetime.now()
-                response, lang_ids, context, rdf_data = generate_response(*method, query, k)
+                response, lang_ids, context, rdf_data = generate_response(*method, query, k, embedder)
                 duration = (datetime.datetime.now() - start).total_seconds()
                 
-                # Response display
                 st.markdown(f"""
                 <div class="response-card">
                     {response}
@@ -326,7 +302,6 @@ def main():
                 </div>
                 """, unsafe_allow_html=True)
                 
-                # Additional information
                 if show_ctx:
                     with st.expander(f"📖 Context from {len(lang_ids)} languages"):
                         for lang_id, ctx in zip(lang_ids, context):
@@ -336,7 +311,6 @@ def main():
                     with st.expander("🔗 Structured facts (RDF)"):
                         st.code("\n".join(rdf_data))
         
-        # Footer note
         st.markdown("---")
         st.markdown("""
         <div style="font-size: 0.8rem; color: #64748b; text-align: center;">