Dump headers of parent nested tables

エントリーポイント tomlkit.dumps(my_dict) から開始し、現在の出力を要求されている親テーブルのヘッダーと比較します。望ましいフォーマットをどのように公開できるかを判断し、その後、ネストされたテーブル apps.test.whatever_1 と apps.test.whatever_2 がヘッダーとインデントを保持していることを確認します。

Hello,

1. I was trying to convert this python dictionnary

   my_dict = {
       "apps": {
           "test": {
               "whatever_1": {
                   "date": "2025-12-16",
               },
               "whatever_2": {
                   "date": "2025-12-16",
               },
           },
       },
   }
   

   ...to this TOML output:

   [apps]
       [apps.test]
           [apps.test.whatever_1]
           date = "2025-12-16"
           [apps.test.whatever_2]
           date = "2025-12-16"
   
   

2. I first had great hopes with:

   print(tomlkit.dumps(my_dict))
   

   but that outputs

   [apps.test.whatever_1]
   date = "2025-12-16"
   
   [apps.test.whatever_2]
   date = "2025-12-16"
   

3. With a bit of pain, I eventually manage to get the expected output using:

   my_dict_new = tomlkit.document()
   for category_name, apps in my_dict.items():
       cat = tomlkit.table(False)
       for app_name, app_reports in apps.items():
           app = tomlkit.table(False).indent(4)
           for report_name, report_data in app_reports.items():
               rep = tomlkit.table(False).indent(4)
               for key, val in report_data.items():
                   rep.add(key, val).indent(4) 
               app.append(report_name, rep)
           cat.append(app_name, app)
       my_dict_new.append(category_name, cat)
   
   print(tomlkit.dumps(my_dict_new))
   

As this the reverse scenario from #47 and #107 in which the target output here seemed to be the default .dumps() output at the time, I wonder whether there would be a quicker way to access the old behavior again (i.e. dump headers for parent nested headers) ?