refactor: implement Priority 1 UI improvements from review

Priority 1 Improvements (High Impact, Low Effort):

1. Remove redundant Model Details accordions
- Removed Extraction Model Details accordion
- Removed Embedding Model Details accordion
- Removed Synthesis Model Details accordion
- Removed event handlers for individual model info updates
- Users now rely on unified Model Information panel in right column

2. Rename conflicting Global Settings sections
- Renamed first 'Global Settings' to 'Output Settings' (clearer purpose)
- Renamed Advanced Mode 'Global Settings' to 'Pipeline Settings' (hardware/logging config)
- Eliminates naming confusion

3. Add visual indicator for active mode
- Added CSS styling for visible mode groups
- Active mode shows blue left border (3px solid)
- Subtle blue gradient background for active mode
- Better visual feedback for which mode is selected

Benefits:
- Reduced clutter in Advanced Mode (removed 3 redundant accordions)
- Clearer section naming (no more duplicate 'Global Settings')
- Better visual UX (active mode clearly indicated)
- Simplified UI with -30 lines of code

app.py

@@ -2530,6 +2530,19 @@ custom_css = """
         font-size: 1.8rem;
     }
 }
+
+/* ===== MODE VISUAL INDICATORS ===== */
+/* Style for visible mode groups to indicate they are active */
+.gradio-group:not([style*="display: none"]) {
+    position: relative;
+}
+
+/* Add subtle highlight border to active mode group */
+.gradio-group:not([style*="display: none"]) > .form {
+    border-left: 3px solid var(--primary-color);
+    padding-left: 12px;
+    background: linear-gradient(90deg, rgba(99, 102, 241, 0.03) 0%, transparent 100%);
+}
 """
 
 
@@ -2573,10 +2586,10 @@ def create_interface():
             with gr.Column(scale=1):
                 
                 # ==========================================
-                # Section 1: Input Configuration (Language + Source)
+                # Section 1: Output Configuration
                 # ==========================================
                 with gr.Group():
-                    gr.HTML('<div class="section-header"><span class="section-icon">🌐</span> Global Settings</div>')
+                    gr.HTML('<div class="section-header"><span class="section-icon">🌐</span> Output Settings</div>')
                     
                     language_selector = gr.Dropdown(
                         choices=[("English", "en"), ("Traditional Chinese (zh-TW)", "zh-TW")],
@@ -2777,11 +2790,6 @@ def create_interface():
                         info="Thinking before JSON extraction (Qwen3 hybrid models only)"
                     )
                     
-                    with gr.Accordion("📊 Extraction Model Details", open=False):
-                        extraction_model_info = gr.Markdown(
-                            value=get_extraction_model_info(DEFAULT_EXTRACTION_MODEL)
-                        )
-                    
                     # ========== STAGE 2: DEDUPLICATION ==========
                     gr.HTML('<div class="section-header" style="margin-top: 20px;"><span class="section-icon">🧬</span> Stage 2: Deduplication</div>')
                     
@@ -2802,11 +2810,6 @@ def create_interface():
                         info="Higher = stricter duplicate detection (items with similarity above this are merged)"
                     )
                     
-                    with gr.Accordion("📊 Embedding Model Details", open=False):
-                        embedding_model_info = gr.Markdown(
-                            value=get_embedding_model_info("granite-107m")
-                        )
-                    
                     # ========== STAGE 3: SYNTHESIS ==========
                     gr.HTML('<div class="section-header" style="margin-top: 20px;"><span class="section-icon">✨</span> Stage 3: Synthesis</div>')
                     
@@ -2860,13 +2863,8 @@ def create_interface():
                             info="Token selection limit"
                         )
                     
-                    with gr.Accordion("📊 Synthesis Model Details", open=False):
-                        synthesis_model_info = gr.Markdown(
-                            value=get_synthesis_model_info(DEFAULT_SYNTHESIS_MODEL)
-                        )
-                    
-                    # ========== GLOBAL SETTINGS ==========
-                    gr.HTML('<div class="section-header" style="margin-top: 20px;"><span class="section-icon">⚙️</span> Global Settings</div>')
+                    # ========== PIPELINE SETTINGS ==========
+                    gr.HTML('<div class="section-header" style="margin-top: 20px;"><span class="section-icon">⚙️</span> Pipeline Settings</div>')
                     
                     adv_thread_config_dropdown = gr.Dropdown(
                         choices=[
@@ -3208,25 +3206,6 @@ def create_interface():
             outputs=[enable_synthesis_reasoning]
         )
         
-        # Wire up model info display updates
-        extraction_model.change(
-            fn=get_extraction_model_info,
-            inputs=[extraction_model],
-            outputs=[extraction_model_info]
-        )
-        
-        embedding_model.change(
-            fn=get_embedding_model_info,
-            inputs=[embedding_model],
-            outputs=[embedding_model_info]
-        )
-        
-        synthesis_model.change(
-            fn=get_synthesis_model_info,
-            inputs=[synthesis_model],
-            outputs=[synthesis_model_info]
-        )
-
         # Debounced auto-discovery for custom repo ID (500ms delay)
         import time as time_module