customXml/
├── item1.xml
├── item2.xml
├── _rels/
│   ├── item1.xml.rels
│   └── item2.xml.rels
├── itemProps1.xml   ← 關(guān)鍵文件
└── itemProps2.xml

<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
<ds:datastoreItem ds:itemID="{B1977F7D-...}" xmlns:ds="...">
  <ds:schemaRefs>
    <ds:schemaRef ds:uri="http://www.wps.cn/officeDocument/2013/wpsCustomData"/>
  </ds:schemaRefs>
</ds:datastoreItem>

<s:customData xmlns="http://www.wps.cn/officeDocument/2013/wpsCustomData">
  <!-- WPS 私有擴展信息：形狀、段落屬性等 -->
</s:customData>

word/_rels/document.xml.rels
  └─ rId12 → ../customXml/item1.xml   （WPS 私有數(shù)據(jù)）
  └─ rId14 → ../customXml/item2.xml   （APA 書目，正常）
[Content_Types].xml
  └─ <Override PartName="/customXml/itemProps1.xml" ContentType="...customXmlProperties+xml"/>

http://www.wps.cn/officeDocument/2013/wpsCustomData

def clean_wps_custom_data(docx_path):
    """
    從生成的 docx 文件中移除 WPS 私有 customXml 數(shù)據(jù)。
    模板由 WPS 創(chuàng)建時會嵌入 wpsCustomData 私有 XML，其 schema URI 為
    http://www.wps.cn/officeDocument/2013/wpsCustomData，Microsoft Word
    無法識別此私有 schema，導(dǎo)致打開時彈出"無法讀取內(nèi)容"警告。
    此函數(shù)在 doc.save() 之后運行，直接對 ZIP 包進行手術(shù)，移除該私有數(shù)據(jù)。
    """
    import zipfile, re, io
    buffer = io.BytesIO()
    try:
        with zipfile.ZipFile(docx_path, 'r') as z_in:
            with zipfile.ZipFile(buffer, 'w', compression=zipfile.ZIP_DEFLATED) as z_out:
                file_list = z_in.namelist()
                # 第一步：找出含有 WPS schema 引用的 itemProps 文件編號
                wps_item_nums = set()
                for fname in file_list:
                    if fname.startswith('customXml/itemProps') and fname.endswith('.xml'):
                        with z_in.open(fname) as f:
                            content = f.read().decode('utf-8', errors='replace')
                        if 'wps.cn' in content or 'wpsCustomData' in content:
                            m = re.search(r'itemProps(\d+)\.xml', fname)
                            if m:
                                wps_item_nums.add(m.group(1))
                # 第二步：構(gòu)建需要整體刪除的文件集合
                files_to_delete = set()
                for num in wps_item_nums:
                    files_to_delete.add(f'customXml/item{num}.xml')
                    files_to_delete.add(f'customXml/itemProps{num}.xml')
                    files_to_delete.add(f'customXml/_rels/item{num}.xml.rels')
                if not wps_item_nums:
                    print('clean_wps_custom_data: 未發(fā)現(xiàn) WPS 私有數(shù)據(jù)，無需清理')
                    return
                print(f'clean_wps_custom_data: 發(fā)現(xiàn) WPS 私有數(shù)據(jù)（item 編號 {wps_item_nums}），正在清理...')
                # 第三步：逐文件處理，重新打包 ZIP
                for fname in file_list:
                    if fname in files_to_delete:
                        continue  # 跳過 WPS 私有文件，不寫入新 zip
                    with z_in.open(fname) as f:
                        data = f.read()
                    if fname == '[Content_Types].xml':
                        # 刪除 WPS itemProps 的 Override 聲明
                        content = data.decode('utf-8')
                        for num in wps_item_nums:
                            content = re.sub(
                                r'<Override[^>]*PartName="/customXml/itemProps' + num + r'\.xml"[^>]*/>', '',
                                content
                            )
                        z_out.writestr(fname, content.encode('utf-8'))
                    elif fname == 'word/_rels/document.xml.rels':
                        # 刪除指向 WPS item 文件的 Relationship 條目
                        content = data.decode('utf-8')
                        for num in wps_item_nums:
                            content = re.sub(
                                r'<Relationship[^>]*Target="\.\./customXml/item' + num + r'\.xml"[^>]*/>', '',
                                content
                            )
                        z_out.writestr(fname, content.encode('utf-8'))
                    else:
                        z_out.writestr(fname, data)
        # 將處理后的內(nèi)容寫回原文件
        buffer.seek(0)
        with open(docx_path, 'wb') as f:
            f.write(buffer.read())
        print('clean_wps_custom_data: WPS 私有數(shù)據(jù)清理完成')
    except Exception as e:
        print(f'clean_wps_custom_data: 清理失?。ú挥绊懳臋n內(nèi)容）: {e}')

doc = Document(template_path)
# ... 填充內(nèi)容 ...
doc.save(output_path)
clean_wps_custom_data(output_path)   # ← 緊接在 save 之后調(diào)用

# ? 錯誤：[^/]* 會在 ContentType 屬性值中的 / 處提前停止
r'<Override[^>]*PartName="..."[^/]*/>'

# ? 正確：[^>]* 只排除 >，允許屬性值中包含 /
r'<Override[^>]*PartName="..."[^>]*/>'

import zipfile

def verify_clean(docx_path):
    with zipfile.ZipFile(docx_path, 'r') as z:
        file_list = z.namelist()

        # 1. 檢查是否還有 WPS schema 引用
        for fname in file_list:
            if fname.startswith('customXml/itemProps') and fname.endswith('.xml'):
                with z.open(fname) as f:
                    content = f.read().decode('utf-8', errors='replace')
                if 'wps.cn' in content or 'wpsCustomData' in content:
                    print(f"? 仍有 WPS schema: {fname}")
                    return False

        # 2. 檢查 document.xml.rels
        with z.open('word/_rels/document.xml.rels') as f:
            rels = f.read().decode('utf-8')
        if 'wpsCustomData' in rels or ('customXml' in rels and 'wps.cn' in rels):
            print("? document.xml.rels 中仍有 WPS 引用")
            return False

        print("? 清理驗證通過，無 WPS 私有數(shù)據(jù)殘留")
        return True