Best Practice for Populating Objects in Python

Question

So I am pulling data (list of JSON) from an API and want to parse it into Python objects. However the JSON objects of my list returned from my API need to be transformed a bit to fit into my object. I don't know if this transformation should happen inside of the __init__ function or outside of the object.

Data pulled from API:

{
 "data": [
 {
 "league": {
 "id": 62,
 "name": "Ligue 2",
 "type": "League"
 },
 "country": {
 "name": "France",
 "code": "FR"
 }
 }
 ]
}

Option 1 (transformation happens inside the object):

class League:
 def __init__(self, data):
 self.id = data["league"]["id"]
 self.name = data["league"]["name"]
 self.country_code = data["country"]["code"]
for d in data["data"]:
 league = League(data=d)

Option 2:

class League:
 def __init__(self, id, name, country_code):
 self.id = id
 self.name = name
 self.country_code = country_code
for d in data["data"]:
 league = League(
 id=d["league"]["id"],
 name=d["league"]["name"],
 country_code=d["country"]["code"])

The transformation will be more complex, but you get the idea hopefully. Looking for advice to figure out which of the 2 options makes the most sense (leaning towards option 2).

Answer 1

Consider that you might want to create League objects from other sources. Then you'd need to construct a dictionary in the specific shape from your JSON source in order to do so. So, option 2 would be preferred.

If your class has to be tightly coupled to the dictionary representation read from JSON at all, consider using a classmethod as an alternative constructor:

class League:
 def __init__(self, id, name, country_code):
 self.id = id
 self.name = name
 self.country_code = country_code
 @classmethod
 def from_json(cls, data):
 league = data['league']
 country = data['country']
 return cls(id=league['id'], name=league['name'], country_code=country['code'])
for d in data["data"]:
 league = League.from_json(d)