Update app.py

app.py

@@ -18,6 +18,7 @@ OUTPUT_DIR = os.path.join(BASE_DIR, "mineru_outputs")
 ZIP_OUTPUT_PATH = os.path.join(BASE_DIR, "mineru_results.zip")
 
 os.makedirs(PAPERS_DIR, exist_ok=True)
+os.makedirs(OUTPUT_DIR, exist_ok=True)
 
 # ==========================================
 # --- 🎨 Custom Purple Theme Definition ---
@@ -93,17 +94,56 @@ purple_theme = PurpleTheme()
 # ==========================================
 # --- ⚙️ Backend Logic & Functions ---
 # ==========================================
+
+def get_tree_html(dir_path):
+    """递归生成可折叠的 HTML 目录树"""
+    if not os.path.exists(dir_path):
+        return "<div style='margin-left: 15px; color: #888;'><i>Directory missing</i></div>"
+        
+    def build_html(current_path):
+        html = ""
+        try:
+            items = sorted(os.listdir(current_path))
+        except Exception:
+            return ""
+        
+        if not items:
+            return ""
+            
+        for item in items:
+            item_path = os.path.join(current_path, item)
+            if os.path.isdir(item_path):
+                # 文件夹节点 (可折叠)
+                html += f'<details style="margin-left: 15px; cursor: pointer; margin-top: 4px;"><summary style="outline: none; color: #C084FC;">📁 <b>{item}</b></summary>'
+                inner_html = build_html(item_path)
+                html += inner_html if inner_html else "<div style='margin-left: 20px; color: #888; font-size: 12px;'><i>Empty</i></div>"
+                html += '</details>'
+            else:
+                # 文件节点
+                html += f'<div style="margin-left: 18px; padding-left: 12px; border-left: 1px dotted #A855F7; margin-top: 4px; color: #DAB2FF;">📄 {item}</div>'
+        return html
+
+    content = build_html(dir_path)
+    return content if content else "<div style='margin-left: 15px; color: #888;'><i>Empty directory</i></div>"
+
 def get_debug_info():
-    now = datetime.now().strftime("%H:%M:%S")
-    files = os.listdir(PAPERS_DIR) if os.path.exists(PAPERS_DIR) else "Directory missing"
-    output_detail = "Not generated"
-    if os.path.exists(OUTPUT_DIR):
-        all_output_items = []
-        for root, dirs, files_in_out in os.walk(OUTPUT_DIR):
-            for name in files_in_out:
-                all_output_items.append(os.path.join(os.path.relpath(root, OUTPUT_DIR), name))
-        output_detail = f"Found {len(all_output_items)} files: {all_output_items[:5]}..." if all_output_items else "Directory exists but is EMPTY"
-    return f"[{now}] 📁 papers/ Content: {files}\n\n[{now}] 📂 mineru_outputs Status: {output_detail}"
+    """返回渲染好的目录树 HTML 容器"""
+    papers_tree = get_tree_html(PAPERS_DIR)
+    output_tree = get_tree_html(OUTPUT_DIR)
+    
+    html = f"""
+    <div style="font-family: 'IBM Plex Mono', monospace; font-size: 13px; background-color: #1e1e1e; border: 1px solid #C084FC; border-radius: 8px; padding: 16px; max-height: 400px; overflow-y: auto;">
+        <details open style="margin-bottom: 12px; cursor: pointer;">
+            <summary style="outline: none; font-size: 15px; color: #A855F7;">📁 <b>papers/</b></summary>
+            {papers_tree}
+        </details>
+        <details open style="cursor: pointer;">
+            <summary style="outline: none; font-size: 15px; color: #A855F7;">📂 <b>mineru_outputs/</b></summary>
+            {output_tree}
+        </details>
+    </div>
+    """
+    return html
 
 def save_pdf(file):
     if file is None: return "❌ Please select a PDF first", get_debug_info()
@@ -121,16 +161,41 @@ def save_pdf(file):
         return f"❌ Error: {str(e)}", get_debug_info()
 
 def clear_pdf():
-    """Delete PDF from the backend directory when cleared in UI"""
+    """Delete PDF and its corresponding mineru output directory"""
     try:
         deleted_files = []
+        deleted_dirs = []
+        
+        # 1. 记录要被删除的 PDF 文件名（用于匹配输出目录）
+        pdf_names = [f for f in os.listdir(PAPERS_DIR) if f.endswith('.pdf')]
+        
+        # 2. 删除 papers 目录下的 PDF
         for f in os.listdir(PAPERS_DIR):
             file_to_del = os.path.join(PAPERS_DIR, f)
             if os.path.isfile(file_to_del):
                 os.remove(file_to_del)
                 deleted_files.append(f)
-        if deleted_files:
-            return "🗑️ File deleted from workspace", get_debug_info()
+                
+        # 3. 删除 mineru_outputs 下的同名文件夹
+        if os.path.exists(OUTPUT_DIR):
+            for pdf_name in pdf_names:
+                base_name = os.path.splitext(pdf_name)[0]
+                base_name_us = base_name.replace(" ", "_") # 兼容 Mineru 可能会把空格转为下划线
+                
+                for item in os.listdir(OUTPUT_DIR):
+                    # 匹配严格同名，或者下划线名称，或者带 .pdf 后缀的文件夹
+                    if item in [base_name, base_name_us, pdf_name]:
+                        dir_to_del = os.path.join(OUTPUT_DIR, item)
+                        if os.path.isdir(dir_to_del):
+                            shutil.rmtree(dir_to_del)
+                            deleted_dirs.append(item)
+                            
+        # 4. 删除可能遗留的打包压缩文件
+        if os.path.exists(ZIP_OUTPUT_PATH):
+            os.remove(ZIP_OUTPUT_PATH)
+            
+        if deleted_files or deleted_dirs:
+            return f"🗑️ Workspace cleared (Deleted: {len(deleted_files)} PDF, {len(deleted_dirs)} Output Folder)", get_debug_info()
         return "ℹ️ Workspace is already empty", get_debug_info()
     except Exception as e:
         return f"❌ Error deleting file: {str(e)}", get_debug_info()
@@ -268,7 +333,7 @@ body, .gradio-container {
     background-image: linear-gradient(rgba(168, 85, 247, .1) 1px, transparent 1px), linear-gradient(90deg, rgba(168, 85, 247, .1) 1px, transparent 1px) !important;
 }
 
-#col-container { margin: 0 auto; max-width: 100%; padding: 20px; }
+#col-container { margin: 0 auto; max-width: 90%; padding: 20px; }
 
 #main-title { text-align: center !important; padding: 1.5rem 0 0.5rem 0; }
 #main-title h1 {
@@ -285,20 +350,19 @@ body, .gradio-container {
 #subtitle p { margin: 0 auto; color: #666; font-size: 1.1rem; font-weight: 500; }
 .dark #subtitle p { color: #DAB2FF; }
 
-/* ======== 修改：模块卡片强化圆角和内外边距 ======== */
 .gradio-group {
     background: rgba(255, 255, 255, 0.9) !important;
     border: 2px solid #E9D5FF !important; 
-    border-radius: 24px !important; /* 更明显的圆角 */
+    border-radius: 24px !important; 
     box-shadow: 0 4px 24px rgba(168, 85, 247, 0.08) !important;
     backdrop-filter: blur(10px); 
     transition: all 0.3s ease; 
-    padding: 16px 16px 16px 16px !important; /* 边缘与内层内容的间距 */
+    padding: 16px 16px 16px 16px !important; 
     overflow: visible !important;
-    margin-bottom: 10px !important; /* 模块之间的垂直间距 */
+    margin-bottom: 10px !important; 
     display: flex !important;
     flex-direction: column !important;
-    gap: 16px !important; /* 内部组件之间的自然间距 */
+    gap: 16px !important; 
 }
 .gradio-group:hover { 
     box-shadow: 0 8px 32px rgba(168, 85, 247, 0.12) !important; 
@@ -309,19 +373,15 @@ body, .gradio-container {
     border-color: rgba(168, 85, 247, 0.3) !important; 
 }
 
-/* 恢复 Gradio 内部组件的圆角（Gradio默认会将 group 内的元素边缘变直） */
 .gradio-group > div, 
-.gradio-group > .form {
-    border-radius: 12px !important;
-}
+.gradio-group > .form { border-radius: 12px !important; }
 
-/* ======== 优化：文件上传框居中与背景清理 ======== */
 #pdf-upload-box {
     border: 2px dashed rgba(192, 132, 252, 0.6) !important;
     border-radius: 16px !important;
     background-color: rgba(250, 245, 255, 0.5) !important;
     transition: all 0.3s ease !important;
-    min-height: 220px !important; /* 增加高度提供居中空间 */
+    min-height: 220px !important; 
     position: relative !important;
     margin-top: 10px !important;
 }
@@ -332,16 +392,10 @@ body, .gradio-container {
     box-shadow: 0 4px 15px rgba(168, 85, 247, 0.15) !important;
 }
 
-/* 隐藏原生的中/英文上传提示、图标及背景 */
-#pdf-upload-box .upload-container {
-    background: transparent !important;
-}
+#pdf-upload-box .upload-container { background: transparent !important; }
 #pdf-upload-box .upload-container > span,
-#pdf-upload-box .upload-container > svg {
-    display: none !important; 
-}
+#pdf-upload-box .upload-container > svg { display: none !important; }
 
-/* 使用伪元素在中央注入自定义文字和图标 */
 #pdf-upload-box .upload-container::before {
     content: "📤\\A Click here to select a PDF\\A or Drag & Drop the file here";
     white-space: pre-wrap;
@@ -356,28 +410,25 @@ body, .gradio-container {
     width: 100%;
     height: 100%;
     position: absolute;
-    top: 0;
-    left: 0;
-    pointer-events: none; /* 防止遮挡拖拽事件 */
+    top: 0; left: 0;
+    pointer-events: none; 
 }
 
-/* ======== 修改：减小按钮尺寸，保持字体一致 ======== */
-/* 主要操作按钮 (Start Mineru & Generate All) */
 .primary-action-btn {
-    border-radius: 25px !important; /* 调小 */
+    border-radius: 25px !important; 
     background: linear-gradient(135deg, #9333EA, #7E22CE) !important;
     color: white !important; font-weight: 700 !important; border: none !important;
-    height: 50px !important; /* 调小：60 -> 50 */
+    height: 50px !important; 
     width: 80% !important;
-    margin-left: auto !important;  /* 新增：自动左边距 */
-    margin-right: auto !important; /* 新增：自动右边距 */
+    margin-left: auto !important;  
+    margin-right: auto !important; 
     margin-top: 10px !important;
     margin-bottom: 10px !important;
-    display: block !important;      /* 新增：确保块级元素 */
+    display: block !important;      
     transition: transform 0.3s cubic-bezier(0.175, 0.885, 0.32, 1.275), box-shadow 0.3s ease !important;
     box-shadow: 0 4px 15px rgba(126, 34, 206, 0.3) !important;
     cursor: pointer !important; 
-    font-size: 1.15rem !important; /* 字体保持一致：略调小以配合新尺寸 */
+    font-size: 1.15rem !important; 
 }
 .primary-action-btn:hover {
     transform: translateY(-5px) scale(1.02) !important;
@@ -386,16 +437,14 @@ body, .gradio-container {
 }
 .primary-action-btn:active { transform: translateY(2px) scale(0.98) !important; box-shadow: 0 2px 10px rgba(126, 34, 206, 0.2) !important; }
 
-/* 3个小按钮的样式 */
 .action-row { display: flex !important; justify-content: center !important; gap: 10px !important; margin-bottom: 10px !important; margin-top: 10px !important;}
 .action-btn {
-    border-radius: 24px !important; /* 调小 */
+    border-radius: 24px !important; 
     background: linear-gradient(135deg, #A855F7, #9333EA) !important;
     color: white !important; font-weight: 600 !important; border: none !important;
-    height: 40px !important; /* 调小：55 -> 48 */
+    height: 40px !important; 
     width: 120px !important;
     flex: none !important;
-    /*文字垂直居中*/
     display: flex !important;
     align-items: center !important;
     justify-content: center !important;
@@ -403,11 +452,10 @@ body, .gradio-container {
     padding: 0 !important;
     margin-top: 10px !important;
     margin-bottom: 10px !important;
-    
     transition: transform 0.3s cubic-bezier(0.175, 0.885, 0.32, 1.275), box-shadow 0.3s ease !important;
     box-shadow: 0 4px 15px rgba(147, 51, 234, 0.2) !important;
-    cursor: pointer !important; display: flex !important; align-items: center !important; justify-content: center !important;
-    font-size: 1.05rem !important; /* 字体保持一致：显式设置 */
+    cursor: pointer !important; 
+    font-size: 1.05rem !important; 
 }
 .action-btn:hover {
     transform: translateY(-5px) scale(1.03) !important;
@@ -416,7 +464,6 @@ body, .gradio-container {
 }
 .action-btn:active { transform: translateY(2px) scale(0.98) !important; box-shadow: 0 2px 10px rgba(147, 51, 234, 0.2) !important; }
 
-/* Terminal Log Style */
 .log-box textarea { font-family: 'IBM Plex Mono', monospace !important; font-size: 13px !important; background-color: #1e1e1e !important; color: #DAB2FF !important; border: 1px solid #C084FC !important; border-radius: 8px !important; }
 .status-text textarea { background-color: transparent !important; border: none !important; box-shadow: none !important; font-weight: 600 !important; color: #6B21A8 !important; }
 .dark .status-text textarea { color: #C084FC !important; }
@@ -424,6 +471,10 @@ body, .gradio-container {
 ::-webkit-scrollbar-track { background: rgba(168, 85, 247, 0.05); border-radius: 4px; }
 ::-webkit-scrollbar-thumb { background: linear-gradient(135deg, #A855F7, #C084FC); border-radius: 4px; }
 ::-webkit-scrollbar-thumb:hover { background: linear-gradient(135deg, #9333EA, #A855F7); }
+
+/* 使折叠树中的 details 和 summary 更加平滑 */
+details > summary { transition: color 0.2s ease; }
+details > summary:hover { color: #E9D5FF !important; }
 """
 
 with gr.Blocks(theme=purple_theme, css=custom_css) as demo:
@@ -439,7 +490,7 @@ with gr.Blocks(theme=purple_theme, css=custom_css) as demo:
                 with gr.Group(elem_classes="gradio-group"):
                     gr.Markdown("### ⚙️ 1. Global API Configuration")
                     with gr.Row():
-                        key_input = gr.Textbox(label="Gemini API Key", type="password", placeholder="sk-...", scale=1)
+                        key_input = gr.Textbox(label="API Key", type="password", placeholder="sk-...", scale=1)
                         api_base_url_input = gr.Textbox(label="Base URL (Optional)", placeholder="https://api.example.com", scale=1)
                     key_btn = gr.Button("💾 Save API Configuration")
                 
@@ -447,14 +498,12 @@ with gr.Blocks(theme=purple_theme, css=custom_css) as demo:
                 with gr.Group(elem_classes="gradio-group"):
                     gr.Markdown("### 📄 2. Document Parsing")
                     
-                    # 简化了 Label，因为 CSS 已经接管了内部文字渲染
                     pdf_input = gr.File(
                         label="Upload Document", 
                         file_types=[".pdf"], 
                         elem_id="pdf-upload-box"
                     )
                     
-                    # 应用全新的 primary-action-btn 悬浮/圆角类
                     parse_btn = gr.Button("🚀 Start Mineru & DAG Extraction", elem_classes="primary-action-btn")
                     
                     parse_status = gr.Textbox(
@@ -471,7 +520,6 @@ with gr.Blocks(theme=purple_theme, css=custom_css) as demo:
                         gen_poster_btn = gr.Button("🖼️ Gen Poster", elem_classes="action-btn")
                         gen_pr_btn = gr.Button("📰 Gen Article", elem_classes="action-btn")
                     
-                    # 应用全新的 primary-action-btn 悬浮/圆角类
                     gen_all_btn = gr.Button("✨ Generate All Assets (ALL)", elem_classes="primary-action-btn")
 
             # ================= RIGHT COLUMN: OUTPUTS & LOGS =================
@@ -496,15 +544,16 @@ with gr.Blocks(theme=purple_theme, css=custom_css) as demo:
                         
                         with gr.Tab("🔍 System Snapshot"):
                             refresh_btn = gr.Button("🔄 Refresh Directory Tree") 
-                            debug_view = gr.Textbox(
-                                label="Workspace Files", lines=13, interactive=False, value=get_debug_info(), elem_classes="log-box"
+                            # 这里改为 HTML 组件，以便渲染我们的 details/summary 折叠树
+                            debug_view = gr.HTML(
+                                value=get_debug_info()
                             )
 
     # ================= LOGIC BINDINGS =================
     key_btn.click(fn=save_api_settings, inputs=[key_input, api_base_url_input], outputs=[parse_status, debug_view])
     
     pdf_input.upload(fn=save_pdf, inputs=pdf_input, outputs=[parse_status, debug_view])
-    # 增加清除逻辑，点击 "x" 后彻底删除后端文件
+    # 彻底清除：不仅清除 PDF，还清除 mineru output 中的对应文件夹
     pdf_input.clear(fn=clear_pdf, outputs=[parse_status, debug_view])
     
     parse_btn.click(fn=run_mineru_parsing_and_dag_gen, outputs=[parse_status, debug_view, cmd_logs])
@@ -522,4 +571,4 @@ with gr.Blocks(theme=purple_theme, css=custom_css) as demo:
     refresh_btn.click(fn=get_debug_info, outputs=debug_view)
 
 if __name__ == "__main__":
-    demo.launch()
+    demo.launch(debug=True)