Upload app.py

app.py

@@ -21,7 +21,57 @@ def normalize_paragraph(text):
     text = re.sub(r'\s{2,}', ' ', text)
     return text.strip()
 
-def format_docx(file, chapter_keywords):
+def process_paragraphs_with_cleanup(doc, combined_pattern):
+    """處理段落並清理多餘空行，實現 ^p^p -> ^p 效果"""
+    content_list = []
+    prev_empty = False
+    
+    for para in doc.paragraphs:
+        text = para.text.strip()
+        if not text:
+            # 只有當前一個段落不是空的時候，才保留這個空段落
+            if not prev_empty:
+                content_list.append(('empty', ''))
+            prev_empty = True
+        elif re.search(combined_pattern, text):
+            content_list.append(('heading', text))
+            prev_empty = False
+        else:
+            content_list.append(('paragraph', text))
+            prev_empty = False
+    
+    return content_list
+
+def rebuild_document(doc, content_list):
+    """重建文檔內容"""
+    for kind, text in content_list:
+        if kind == 'heading':
+            heading = doc.add_paragraph(text, style='Heading 1')
+            heading.paragraph_format.page_break_before = True
+            heading.paragraph_format.space_before = Cm(0)
+            heading.paragraph_format.space_after = Cm(0.3)
+            heading.paragraph_format.line_spacing = 1.0
+            heading.paragraph_format.left_indent = Cm(0)
+            heading.paragraph_format.first_line_indent = Cm(0)
+            set_outline_level(heading, 0)
+            for run in heading.runs:
+                run.font.name = '新細明體'
+                run.font.size = Pt(16)
+        elif kind == 'paragraph':
+            clean_text = normalize_paragraph(text)
+            para = doc.add_paragraph(clean_text)
+            para.paragraph_format.space_before = Cm(0)
+            para.paragraph_format.space_after = Cm(0)
+            para.paragraph_format.line_spacing = 1.0
+            para.paragraph_format.left_indent = Cm(0)
+            para.paragraph_format.first_line_indent = Cm(0.7)
+            for run in para.runs:
+                run.font.name = '新細明體'
+        elif kind == 'empty':
+            # 只有在真正需要保留空段落時才創建
+            doc.add_paragraph('')
+
+def format_docx(file, chapter_keywords, remove_empty_paragraphs=True):
     if file is None:
         return None, "請上傳一個 Word 文件"
     if not chapter_keywords.strip():
@@ -43,56 +93,25 @@ def format_docx(file, chapter_keywords):
         patterns = [f'第\s*[0-9一二三四五六七八九十百千萬壹貳參肆伍陸柒捌玖拾佰仟萬IVXLCDMivxlcdm]+\s*{k}' for k in keywords]
         combined_pattern = '|'.join(patterns)
 
-        # 模擬 Word 的 ^p^p -> ^p 處理段落
-        content_list = []
-        prev_empty = False
-        for para in doc.paragraphs:
-            text = para.text.strip()
-            if not text:
-                if not prev_empty:
-                    content_list.append(('empty', ''))
-                prev_empty = True
-            elif re.search(combined_pattern, text):
-                content_list.append(('heading', text))
-                prev_empty = False
-            else:
-                content_list.append(('paragraph', text))
-                prev_empty = False
+        # 處理段落並清理多餘空行
+        content_list = process_paragraphs_with_cleanup(doc, combined_pattern)
+        
+        # 如果選擇移除空段落，過濾掉所有空段落
+        if remove_empty_paragraphs:
+            content_list = [item for item in content_list if item[0] != 'empty']
 
         # 清空原始內容
         for p in doc.paragraphs:
             p._element.getparent().remove(p._element)
 
         # 重建段落
-        for kind, text in content_list:
-            if kind == 'heading':
-                heading = doc.add_paragraph(text, style='Heading 1')
-                heading.paragraph_format.page_break_before = True
-                heading.paragraph_format.space_before = Cm(0)
-                heading.paragraph_format.space_after = Cm(0.3)
-                heading.paragraph_format.line_spacing = 1.0
-                heading.paragraph_format.left_indent = Cm(0)
-                heading.paragraph_format.first_line_indent = Cm(0)
-                set_outline_level(heading, 0)
-                for run in heading.runs:
-                    run.font.name = '新細明體'
-                    run.font.size = Pt(16)
-            elif kind == 'paragraph':
-                clean_text = normalize_paragraph(text)
-                para = doc.add_paragraph(clean_text)
-                para.paragraph_format.space_before = Cm(0)
-                para.paragraph_format.space_after = Cm(0)
-                para.paragraph_format.line_spacing = 1.0
-                para.paragraph_format.left_indent = Cm(0)
-                para.paragraph_format.first_line_indent = Cm(0.7)
-                for run in para.runs:
-                    run.font.name = '新細明體'
-            elif kind == 'empty':
-                doc.add_paragraph('')
+        rebuild_document(doc, content_list)
 
         output_path = tempfile.mktemp(suffix='.docx')
         doc.save(output_path)
-        return output_path, f"✅ 處理完成！找到章節關鍵字：{', '.join(keywords)}"
+        
+        empty_status = "已移除所有空段落" if remove_empty_paragraphs else "保留單個空段落"
+        return output_path, f"✅ 處理完成！找到章節關鍵字：{', '.join(keywords)}，{empty_status}"
     except Exception as e:
         return None, f"❌ 處理失敗：{str(e)}"
 
@@ -109,12 +128,21 @@ def create_interface():
             with gr.Column(scale=1):
                 file_input = gr.File(label="上傳 Word 文件 (.docx)", file_types=[".docx"], file_count="single")
                 chapter_input = gr.Textbox(label="章節分段方式", placeholder="章,節,話", value="章,節,話")
+                remove_empty_checkbox = gr.Checkbox(
+                    label="移除空段落", 
+                    value=True, 
+                    info="勾選時會移除所有空段落，取消勾選時會保留單個空段落（^p^p -> ^p）"
+                )
                 process_btn = gr.Button("🔄 開始處理", variant="primary", size="lg")
             with gr.Column(scale=1):
                 status_output = gr.Textbox(label="處理狀態", interactive=False, lines=3)
                 download_output = gr.File(label="下載處理後的文件", interactive=False)
 
-        process_btn.click(fn=format_docx, inputs=[file_input, chapter_input], outputs=[download_output, status_output])
+        process_btn.click(
+            fn=format_docx, 
+            inputs=[file_input, chapter_input, remove_empty_checkbox], 
+            outputs=[download_output, status_output]
+        )
 
     return demo