用 Python 的 difflib 模拟 GNU diff 的输出

项目中需要在不同操作系统上使用 GNU diff 命令，因此用 Python 自带的 difflib 库模拟它的输出格式。

代码

import difflib
import re


def diff(file1, file2, encoding='utf-8'):
    """
    Simulates the output of GNU diff.
    You can use `diff(f1, f2)` to simulate `diff -w f1 f2`
    """
    texts        = []
    for f in [file1, file2]:
        with open(f, 'r', encoding=encoding) as f:
            text = f.read()
        # Ignore whitespace characters
        for i in '\t\r\v\f':
            text = text.replace(i, '')
        texts   += [text.split('\n')]
    text1, text2 = texts

    output       = []
    new_part     = True
    num          = 0
    for line in difflib.unified_diff(text1, text2, fromfile=file1, tofile=file2, n=0, lineterm=''):
        num     += 1
        if num   < 3:
            continue
        flag             = line[0]
        if flag         == '-':   # line unique to sequence 1
            new_flag     = '< '
        elif flag       == '+':   # line unique to sequence 2
            new_flag     = '> '
            if new_part:
                new_part = False
                output  += ['---']
        elif flag       == ' ':   # line common to both sequences
            # new_flag   = b'  '
            continue
        elif flag       == '?':   # line not present in either input sequence
            new_flag     = '? '
        elif flag       == '@':
            output      += [re.sub(r'@@ -([^ ]+) \+([^ ]+) @@', r'\1c\2', line)]
            new_part     = True
            continue
        else:
            new_flag     = flag
        output          += [new_flag + line[1:]]

    return '\n'.join(output)

效果演示

GNU diff 的输出格式：

will@Will:/nt/d/1$ diff -w 1.txt 2.txt 
1,3c1,2
< 1
< 2
< 3
---
> Hello
> World

用 Python 模拟的输出格式：

>>> print(diff('1.txt', '2.txt'))
1,3c1,2
< 1
< 2
< 3
---
> Hello
> World

可见已经能成功模拟。