Python 判断两个字典类型值列以分配标签

我有一个数据帧，看起来是这样的:

import pandas as pd
import numpy as np

data = {
    'api_spec_id': [117455, 117455, 117455, 117456, 117456],
    'commit_date': ['2023-06-01', '2023-06-02', '2023-06-03', '2023-06-01', '2023-06-02'],
    'diff_2': [
        [{
            "id": "response-media-type-removed",
            "text": "removed the media type 'application/json' for the response with the status '400'",
            "level": 0,
            "operation": "GET",
            "path": "/continuous/filters"
        },
        {
            "id": "response-media-type-removed",
            "text": "removed the media type 'application/json' for the response with the status '406'",
            "level": 0,
            "operation": "GET",
            "path": "/continuous/filters"
        },
        {
            "id": "response-media-type-removed",
            "text": "removed the media type 'application/json' for the response with the status '501'",
            "level": 0,
            "operation": "GET",
            "path": "/continuous/filters"
        }],
        [],
        [],
        [{
            "id": "response-media-type-removed",
            "text": "removed the media type 'application/xml' for the response with the status '200'",
            "level": 0,
            "operation": "GET",
            "path": "/continuous/filters"
        }],
        []
    ],
    'nonBreakingChanges': [
        {
            "api title added": 0,
            "api title modified": 0,
            "api description added": 0,
            "api description modified": 0,
            "api version added": 0,
            "api version modified": 0,
            "api contact deleted": 0,
            "api contact added": 0,
            "api contact modified": 0,
            "api license deleted": 0,
            "api license added": 0,
            "api license modified": 0,
            "server deleted": 0,
            "server added": 0,
            "server modified": 0,
            "path added": 0,
            "path parameter added": 0,
            "path parameter deleted": 0,
            "path parameter modified": 0,
            "desc schema property of resp": 57
        },
        {},
        {},
        {},
        {
            "api title added": 0,
            "api title modified": 0,
            "api description added": 0,
            "api description modified": 1,
            "api version added": 0,
            "api version modified": 0,
            "api contact deleted": 0,
            "api contact added": 0,
            "api contact modified": 0,
            "api license deleted": 0,
            "api license added": 0,
            "api license modified": 0,
            "server deleted": 0,
            "server added": 0,
            "server modified": 0,
            "path added": 0,
            "path parameter added": 0,
            "path parameter deleted": 0,
            "path parameter modified": 0,
            "desc schema property of resp": 0
        }
    ]
}

df = pd.DataFrame(data)

每个id都有多个提交日期，所以我想遍历所有提交日期，实际上是这样做的:

我有两个列diff_2和nonBreakingChanges，我想通过它们基于四个条件创建新的列type:

首先，如果diff_2和nonBreakingChanges都有值并且不为空，则分配Both

如果只有diff_2具有值，则分配B(在这种情况下，另一列可以是{}、[]、NaN或空字符串)

如果nonbreakingChanges具有值，则分配NB(在这种情况下，另一列可以是{}、[]、NaN或空字符串)

我不确定这如何可能与字典类型的数据，任何建议或 idea 将非常感激.