ocr_verify/test/unified_diff.py

import difflib

def generate_unified_diff(file1_path, file2_path):
    """
    生成类似git diff的统一差异格式输出
    """
    with open(file1_path, 'r', encoding='utf-8') as f1, open(file2_path, 'r', encoding='utf-8') as f2:
        file1_lines = f1.readlines()
        file2_lines = f2.readlines()

    # 使用unified_diff生成差异
    diff = difflib.unified_diff(
        file1_lines, 
        file2_lines, 
        fromfile=file1_path, 
        tofile=file2_path,
        lineterm=''  # 确保每行末尾不添加额外字符
    )
    
    # 将差异生成器转换为列表并返回，或直接打印
    diff_output = list(diff)
    return diff_output

import difflib

def generate_html_diff(file1_path, file2_path, output_html_path):
    """
    生成HTML格式的差异报告并保存到文件
    """
    with open(file1_path, 'r', encoding='utf-8') as f1, open(file2_path, 'r', encoding='utf-8') as f2:
        file1_lines = f1.readlines()
        file2_lines = f2.readlines()

    # 创建HtmlDiff对象并生成差异HTML内容
    diff_html = difflib.HtmlDiff().make_file(
        file1_lines, 
        file2_lines, 
        fromdesc=file1_path, 
        todesc=file2_path,
        context=True  # 显示上下文
    )

    # 将HTML内容写入文件
    with open(output_html_path, 'w', encoding='utf-8') as html_file:
        html_file.write(diff_html)
    print(f"HTML差异报告已生成: {output_html_path}")

if __name__ == "__main__":
	# 使用示例
	file1 = '/Users/zhch158/workspace/data/至远彩色印刷工业有限公司/data_DotsOCR_Results/2023年度报告母公司_page_001.md'
	file2 = '/Users/zhch158/workspace/repository.git/ocr_verify/output/pre_validation/2023年度报告母公司_page_001.md'
	diff_result = generate_unified_diff(file1, file2)

	# 打印差异结果
	if diff_result:
		print("\n".join(diff_result))
	else:
		print("文件内容相同，无差异。")

	# 使用示例
	generate_html_diff(file1, file2, 'diff_report.html')