如何使用pythonpandas從dict類型的數據幀中刪除重復項？

我有下面的數據幀與重復數據。我想從數據幀中刪除這些重復項。

df = pd.DataFrame({'test_id': [
        {'user_id':2, 'insert_date':'2020-12-23', 'is_admin': "true"},
        {'user_id':4, 'insert_date':'2020-12-23', 'is_admin': "true"},
        {'user_id':3, 'insert_date':'2020-12-21', 'is_admin': "false"},
        {'user_id':2, 'insert_date':'2020-12-23', 'is_admin': "true"}
    ], 'contact_id':[1,4,2,1]}
)

print(df)

                                             test_id  contact_id
0  {'user_id': 2, 'insert_date': '2020-12-23', 'i...           1
1  {'user_id': 4, 'insert_date': '2020-12-23', 'i...           4
2  {'user_id': 3, 'insert_date': '2020-12-21', 'i...           2
3  {'user_id': 2, 'insert_date': '2020-12-23', 'i...           1

我在下面試過消除重復

df = df.drop_duplicates(subset=['test_id', 'contact_id'], keep='first')
print(df)

但是低于錯誤

c:\python38\lib\site-packages\pandas\core\algorithms.py in _factorize_array(values, na_sentinel, size_hint, na_value)
    482 
    483     table = hash_klass(size_hint or len(values))
--> 484     uniques, codes = table.factorize(values, na_sentinel=na_sentinel, na_value=na_value)
    485 
    486     codes = ensure_platform_int(codes)

pandas\_libs\hashtable_class_helper.pxi in pandas._libs.hashtable.PyObjectHashTable.factorize()

pandas\_libs\hashtable_class_helper.pxi in pandas._libs.hashtable.PyObjectHashTable._unique()

TypeError: unhashable type: 'dict'

有誰能指導我如何刪除基于'test_id'，'contact_id'組合的重復數據嗎？

我要低于輸出

                                             test_id  contact_id
0  {'user_id': 2, 'insert_date': '2020-12-23', 'i...           1
1  {'user_id': 4, 'insert_date': '2020-12-23', 'i...           4
2  {'user_id': 3, 'insert_date': '2020-12-21', 'i...           2