更好的文本差异比较

目前的nonebug在测试时遇到和期望不相符的文本时，只会显示这样的格式：

f"Application got api call {api} with data {data} but expected {model.data}"

这经常会造成如下对人眼检查不友好的输出:

Application got api call ... with data {'foo': 1, 'bar': {'hello': 'world'}} but expected {'foo': 1, 'bar': {'hello': 'nonebug'}}

希望可以加入对 期望输出 和 实际输出 的比较，类似这样的格式:

  {
      "foo": 1,
      "bar": {
-         "hello": "world"
?                   ^ ^^^

+         "hello": "nonebot"
?                   ^ ^^^^^

      }
  }

这样的输出可以通过标准库difflib实现，大体类似于：

from difflib import ndiff
from json import dumps

print(
	"\n".join(
		ndiff(
			dumps(data, indent=4).splitlines(),
			dumps(model.data, indent=4).splitlines()
		)
	)
)

Locate the assertion or error-formatting path that currently emits the API call message with actual and expected data. Read how the existing pytest output is assembled, then use the issue's difflib example as the target format and run the relevant test suite to verify readable differences for nested data.