Optimizing code for this function that filters a JSON file

Question

I made a function that turns all child from a JSON file to parents (or brothers?). It adds the parent name to the key, so there are no duplicates.

It works ok, but I would like to ask if there are better ways to do this, in a logic, pythonic and even syntaxic way. I feel I am stuck with if-else logic.

Thank you!

I have this JSON file:

{
   "update_id":1,
   "message":{
      "message_id":1,
      "from":{
         "id":1,
         "is_bot":false,
         "first_name":"Albert",
         "language_code":"es"
      },
      "chat":{
         "id":1,
         "first_name":"Albert",
         "type":"private"
      },
      "date":1,
      "text":"HI"
   }
}

Filtering and convertion from child to parents is done this code:

MF_FIELDS = (
    "message_from_id",
    "message_from_is_bot",
    "message_from_first_name",
    "message_from_username",
    "message_chat_id",
    "message_chat_title",
    "message_chat_type",
    "message_date",
    "message_text",
    "message_entities_type"
)

def dict_to_first_level(d, parent_name=None, rec_dict=None):
    ret_dict = {}

    if rec_dict: ret_dict = {**ret_dict, **rec_dict} 

    for k, v in d.items():

        if isinstance(v, dict):
            if parent_name: k = f'{parent_name}_{k}'

            _recursive = dict_to_first_level(v, k, ret_dict)
            ret_dict = {**ret_dict, **_recursive[0]} 

        else:
            if parent_name:
                k = f'{parent_name}_{k}'
            ret_dict[k] = v

    return ret_dict, parent_name, rec_dict

def filter_json(encoded_json, filter):
    ret_dict = {}
    json_dict = json.loads(encoded_json)
    first_level_dict, _, _ = dict_to_first_level(json_dict)

    for k, v in first_level_dict.items():
        if k in filter:
            ret_dict[k] = v
    return ret_dict

Producing this result:

{
   "message_from_id":1,
   "message_from_is_bot":false,
   "message_from_first_name":"Albert",
   "message_chat_id":1,
   "message_chat_type":"private",
   "message_date":1,
   "message_text":"HI"
}

Answer 1

The main issues are naming and formatting.

MF_FIELDS - what is mf?
ret_dict - what is ret?
rec_dict - what is rec? it's easy to mix it up with ret_dict
k and v can be key and value
what is _recursive?
filter shadows a built-in name

if rec_dict: ret_dict = {**ret_dict, **rec_dict}
Split this into 2 lines, it's much easier to read this way. In this case there is no reason to add **ret_dict since ret_dict is empty as was stated on the previous line.

for k, v in d.items():
    if isinstance(v, dict):
        if parent_name: k = f'{parent_name}_{k}'

The line (it should be on a new line) k = f'{parent_name}_{k}' looks like it's trying to change the value of one of the keys in d. But in fact it creates a new object for k to reference without changing the dictionary in any way. To make it clear, don't name this new object k.

In general when you write something like

for element in some_collection:
    element = something

your collection never changes, the name element just stops referring to one of its elements and now points to the same object as something.

---

I'd like to know what task this algorithm is trying to solve? If you want specific values from the dictionary, you can get them like so:

message_from_first_name = your_dict["message"]["from"]["first_name"]