def upload_image(image_path, upload_url):
    try:
        with open(image_path, 'rb') as file:
            files = {'image': file}
            response = requests.post(upload_url, files=files)
            if response.status_code == 200:
                return response.json().get('url')
            else:
                print(f"上傳失敗，狀態(tài)碼: {response.status_code}")
    except Exception as e:
        print(f"上傳時發(fā)生錯誤: {e}")
    return None

def process_docx(docx_path, upload_url):
    doc = docx.Document(docx_path)
    sections = {}
    stack = []  
    temp_image_dir = 'temp_images'
    os.makedirs(temp_image_dir, exist_ok=True)

    for para in doc.paragraphs:
        if para.style.name.startswith('Heading'):
            level = int(para.style.name.split()[1])  

            while stack and stack[-1][2] >= level:
                prev_title, prev_paras, _ = stack.pop()
                if prev_title not in sections:
                    sections[prev_title] = prev_paras

            title = para.text
            stack.append((title, [], level))
        else:
            p_text = para.text
            image_count = 0
            for rel in para._element.xpath('.//w:drawing/wp:inline/a:graphic/a:graphicData/pic:pic'):
                for blip in rel.xpath('.//a:blip'):
                    rId = blip.get('{http://schemas.openxmlformats.org/officeDocument/2006/relationships}embed')
                    if rId in doc.part.rels:
                        image_part = doc.part.rels[rId].target_part
                        image_ext = image_part.content_type.split('/')[-1]
                        image_path = os.path.join(temp_image_dir, f'{rId}.{image_ext}')
                        with open(image_path, 'wb') as f:
                            f.write(image_part.blob)
                        image_url = upload_image(image_path, upload_url)
                        if image_url:
                            placeholder = f'[圖片 {image_count}]'
                            p_text = p_text + f' {placeholder}'
                            p_text = p_text.replace(placeholder, image_url)
                            image_count += 1

            if stack:
                stack[-1][1].append(p_text)

    while stack:
        prev_title, prev_paras, _ = stack.pop()
        if prev_title not in sections:
            sections[prev_title] = prev_paras

    for filename in os.listdir(temp_image_dir):
        file_path = os.path.join(temp_image_dir, filename)
        os.remove(file_path)
    os.rmdir(temp_image_dir)

    return sections

if __name__ == "__main__":
    docx_path = 'Z:/知識圖譜問答/黑龍江省一卡通平臺 軟件操作手冊 202501（財政）.docx'
    upload_url = 'http://localhost:5000/upload'
    result = process_docx(docx_path, upload_url)
    for section_title, paragraphs in result.items():
        print(f"章節(jié): {section_title}")
        for para in paragraphs:
            # print('-------------')
            print(para)
        print("-" * 50)

pip install python - docx requests