10.5. JSON Decoder

10.5.1. Problem

  • Problem with date, datetime, time, timedelta

  • Python does not decode values automatically

10.5.2. SetUp

>>> from pprint import pprint
>>> from datetime import date
>>> import json

10.5.3. Problem

>>> DATA = """{
...     "firstname": "Mark",
...     "lastname": "Watney",
...     "birthdate": "1994-10-12"
... }"""
>>>
>>> result = json.loads(DATA)
>>> print(result)   
{'firstname': 'Mark',
 'lastname': 'Watney',
 'birthdate': '1994-10-12'}

10.5.4. Function Decoder

  • This works for simple (flat) data

  • This works for nested data structures

>>> DATA = """{
...     "firstname": "Mark",
...     "lastname": "Watney",
...     "birthdate": "1994-10-12"
... }"""
>>>
>>>
>>> def decoder(data: dict) -> dict:
...     for field, value in data.items():
...         if field == 'birthdate':
...             data[field] = date.fromisoformat(value)
...     return data
>>>
>>>
>>> result = json.loads(DATA, object_hook=decoder)
>>> print(result)   
{'firstname': 'Mark',
 'lastname': 'Watney',
 'birthdate': datetime.date(1994, 10, 12)}

10.5.5. Context Dependency Injection

>>> DATA = """{
...     "firstname": "Mark",
...     "lastname": "Watney",
...     "birthdate": "1994-10-12"
... }"""
>>>
>>>
>>> class Decoder(json.JSONDecoder):
...     def __init__(self):
...         super().__init__(object_hook=self.default)
...
...     def default(self, data: dict) -> dict:
...         for field, value in data.items():
...             if field == 'birthdate':
...                 data[field] = date.fromisoformat(value)
...         return data
>>>
>>>
>>> result = json.loads(DATA, cls=Decoder)
>>> print(result)   
{'firstname': 'Mark',
 'lastname': 'Watney',
 'birthdate': datetime.date(1994, 10, 12)}

10.5.6. Use Case - 0x01

  • This works for simple (flat) data

  • This won't work for nested data structures

>>> import json
>>> from datetime import datetime, date, time, timedelta
>>> from pprint import pprint
>>>
>>>
>>> DATA = """{
...     "firstname": "Mark",
...     "lastname": "Watney",
...     "birthdate": "1994-10-12",
...     "launch": "1969-07-21T02:56:15",
...     "landing": "12:30:00",
...     "flight_time": 15552000,
...     "duration": 13
... }"""
>>>
>>>
>>> def decoder(obj: dict) -> dict:
...     obj['birthdate'] = date.fromisoformat(obj['birthdate'])
...     obj['launch'] = datetime.fromisoformat(obj['launch'])
...     obj['landing'] = time.fromisoformat(obj['landing'])
...     obj['flight_time'] = timedelta(seconds=obj['flight_time'])
...     obj['duration'] = timedelta(days=obj['duration'])
...     return obj
>>>
>>>
>>> result = json.loads(DATA, object_hook=decoder)
>>> pprint(result, sort_dicts=False)
{'firstname': 'Mark',
 'lastname': 'Watney',
 'birthdate': datetime.date(1994, 10, 12),
 'launch': datetime.datetime(1969, 7, 21, 2, 56, 15),
 'landing': datetime.time(12, 30),
 'flight_time': datetime.timedelta(days=180),
 'duration': datetime.timedelta(days=13)}

10.5.7. Use Case - 0x02

  • This works for simple (flat) data

  • This won't work for nested data structures

>>> import json
>>> from datetime import datetime, date, time, timedelta
>>> from pprint import pprint
>>>
>>>
>>> DATA = """{
...     "firstname": "Mark",
...     "lastname": "Watney",
...     "birthdate": "1994-10-12",
...     "launch": "1969-07-21T02:56:15",
...     "landing": "12:30:00",
...     "flight_time": 15552000,
...     "duration": 13
... }"""
>>>
>>>
>>> def decoder(obj: dict) -> dict:
...     return obj | {
...         'birthdate': date.fromisoformat(obj['birthdate']),
...         'launch': datetime.fromisoformat(obj['launch']),
...         'landing': time.fromisoformat(obj['landing']),
...         'flight_time': timedelta(seconds=obj['flight_time']),
...         'duration': timedelta(days=obj['duration']),
...     }
>>>
>>>
>>> result = json.loads(DATA, object_hook=decoder)
>>> pprint(result, sort_dicts=False)
{'firstname': 'Mark',
 'lastname': 'Watney',
 'birthdate': datetime.date(1994, 10, 12),
 'launch': datetime.datetime(1969, 7, 21, 2, 56, 15),
 'landing': datetime.time(12, 30),
 'flight_time': datetime.timedelta(days=180),
 'duration': datetime.timedelta(days=13)}

10.5.8. Use Case - 0x03

>>> from datetime import datetime, date, time, timedelta
>>> import json
>>> from pprint import pprint
>>>
>>>
>>> DATA = """{
...     "firstname": "Mark",
...     "lastname": "Watney",
...     "birthdate": "1994-10-12",
...     "launch": "1969-07-21T02:56:15",
...     "landing": "12:30:00",
...     "flight_time": 15552000,
...     "duration": 13
... }"""
>>>
>>>
>>> def decoder(x):
...     for key, value in x.items():
...         match key:
...             case 'birthdate':
...                 x[key] = date.fromisoformat(value)
...             case 'launch':
...                 x[key] = datetime.fromisoformat(value)
...             case 'landing':
...                 x[key] = time.fromisoformat(value)
...             case 'flight_time' | 'mission_time':
...                 x[key] = timedelta(seconds=float(value))
...             case 'duration':
...                 x[key] = timedelta(days=int(value))
...     return x
>>>
>>>
>>> result = json.loads(DATA, object_hook=decoder)
>>> pprint(result, sort_dicts=False)
{'firstname': 'Mark',
 'lastname': 'Watney',
 'birthdate': datetime.date(1994, 10, 12),
 'launch': datetime.datetime(1969, 7, 21, 2, 56, 15),
 'landing': datetime.time(12, 30),
 'flight_time': datetime.timedelta(days=180),
 'duration': datetime.timedelta(days=13)}

10.5.9. Use Case - 0x04

>>> from datetime import datetime, date, time, timedelta
>>> import json
>>> from pprint import pprint
>>>
>>>
>>> DATA = """{
...     "firstname": "Mark",
...     "lastname": "Watney",
...     "birthdate": "1994-10-12",
...     "launch": "1969-07-21T02:56:15",
...     "landing": "12:30:00",
...     "flight_time": 15552000,
...     "duration": 13
... }"""
>>>
>>>
>>> class Decoder(json.JSONDecoder):
...     def __init__(self) -> None:
...         super().__init__(object_hook=lambda data: {
...             field: getattr(self, field)(value)
...             for field, value in data.items()})
...
...     def firstname(self, value: str) -> str:
...         return value
...
...     def lastname(self, value: str) -> str:
...         return value
...
...     def birthdate(self, value: str) -> date:
...         return date.fromisoformat(value)
...
...     def launch(self, value: str) -> datetime:
...         return datetime.fromisoformat(value)
...
...     def landing(self, value: str) -> time:
...         return time.fromisoformat(value)
...
...     def flight_time(self, value: str) -> timedelta:
...         return timedelta(seconds=float(value))
...
...     def duration(self, value: str) -> timedelta:
...         return timedelta(days=int(value))
>>>
>>>
>>> result = json.loads(DATA, cls=Decoder)
>>> pprint(result, sort_dicts=False)
{'firstname': 'Mark',
 'lastname': 'Watney',
 'birthdate': datetime.date(1994, 10, 12),
 'launch': datetime.datetime(1969, 7, 21, 2, 56, 15),
 'landing': datetime.time(12, 30),
 'flight_time': datetime.timedelta(days=180),
 'duration': datetime.timedelta(days=13)}

10.5.10. Use Case - 0x05

>>> from datetime import datetime, date, time, timedelta
>>> import json
>>>
>>>
>>> DATA = """{
...     "firstname": "Mark",
...     "lastname": "Watney",
...     "birthdate": "1994-10-12",
...     "launch": "1969-07-21T02:56:15",
...     "landing": "12:30:00",
...     "flight_time": 15552000,
...     "duration": 13
... }"""
>>>
>>>
>>> class Decoder(json.JSONDecoder):
...     def __init__(self):
...         super().__init__(object_hook=lambda data: {
...                 field: self.default(field, value)
...                 for field, value in data.items()})
...
...     def default(self, field, value):
...         result = {
...             'birthdate': lambda x: date.fromisoformat(x),
...             'launch': lambda x: datetime.fromisoformat(x),
...             'landing': lambda x: time.fromisoformat(x),
...             'duration': lambda x: timedelta(days=x),
...             'flight_time': lambda x: timedelta(seconds=x),
...         }.get(field, value)
...         return result(value) if callable(result) else result
>>>
>>>
>>> result = json.loads(DATA, cls=Decoder)
>>> print(result)   
{'firstname': 'Mark',
 'lastname': 'Watney',
 'birthdate': datetime.date(1994, 10, 12),
 'launch': datetime.datetime(1969, 7, 21, 2, 56, 15),
 'landing': datetime.time(12, 30),
 'flight_time': datetime.timedelta(days=180),
 'duration': datetime.timedelta(days=13)}

10.5.11. Assignments

"""
* Assignment: JSON Decoder Function
* Complexity: easy
* Lines of code: 8 lines
* Time: 5 min

English:
    1. Define `result: dict` with decoded `DATA` from JSON
    2. Use decoder function
    3. Run doctests - all must succeed

Polish:
    1. Zdefiniuj `result: dict` z odkodowanym `DATA` z JSON
    2. Użyj dekodera funkcyjnego
    3. Uruchom doctesty - wszystkie muszą się powieść

Hints:
    * json.loads(object_hook=...)
    * dict.items()
    * datetime.fromisoformat()
    * date.fromisoformat()

Tests:
    >>> import sys; sys.tracebacklimit = 0
    >>> from inspect import isfunction
    >>> assert isfunction(decoder), \
    'Decoder must be a function'
    >>> assert type(result) is dict, \
    'Result must be a dict'
    >>> assert len(result) > 0, \
    'Result cannot be empty'
    >>> assert all(type(key) is str
    ...            and type(value) in (str, datetime, list)
    ...            for key, value in result.items()), \
    'All keys must be str and all values must be either str, datetime or list'


    >>> result  # doctest: +NORMALIZE_WHITESPACE
    {'mission': 'Ares 3',
     'launch_date': datetime.datetime(2035, 6, 29, 0, 0),
     'destination': 'Mars',
     'destination_landing': datetime.datetime(2035, 11, 7, 0, 0),
     'destination_location': 'Acidalia Planitia',
     'crew': [{'name': 'Melissa Lewis', 'birthdate': datetime.date(1995, 7, 15)},
              {'name': 'Rick Martinez', 'birthdate': datetime.date(1996, 1, 21)},
              {'name': 'Alex Vogel', 'birthdate': datetime.date(1994, 11, 15)},
              {'name': 'Chris Beck', 'birthdate': datetime.date(1999, 8, 2)},
              {'name': 'Beth Johanssen', 'birthdate': datetime.date(2006, 5, 9)},
              {'name': 'Mark Watney', 'birthdate': datetime.date(1994, 10, 12)}]}
"""

import json
from datetime import datetime, date


DATA = """
    {"mission": "Ares 3",
     "launch_date": "2035-06-29T00:00:00",
     "destination": "Mars",
     "destination_landing": "2035-11-07T00:00:00",
     "destination_location": "Acidalia Planitia",
     "crew": [{"name": "Melissa Lewis", "birthdate": "1995-07-15"},
              {"name": "Rick Martinez", "birthdate": "1996-01-21"},
              {"name": "Alex Vogel", "birthdate": "1994-11-15"},
              {"name": "Chris Beck", "birthdate": "1999-08-02"},
              {"name": "Beth Johanssen", "birthdate": "2006-05-09"},
              {"name": "Mark Watney", "birthdate": "1994-10-12"}]}"""


# JSON decoder
def decoder(obj):
    ...


# JSON decoded DATA
# type: dict[str, str|list|datetime]
result = ...


"""
* Assignment: JSON Decoder Class
* Complexity: medium
* Lines of code: 12 lines
* Time: 5 min

English:
    1. Define `result: dict` with decoded `DATA` from JSON
    2. Use decoder class
    3. Run doctests - all must succeed

Polish:
    1. Zdefiniuj `result: dict` z odkodowanym `DATA` z JSON
    2. Użyj dekodera klasowego
    3. Uruchom doctesty - wszystkie muszą się powieść

Hints:
    * json.loads(object_hook=...)
    * dict.items()
    * super().__init__(object_hook=...)
    * datetime.fromisoformat()
    * date.fromisoformat()

Tests:
    >>> import sys; sys.tracebacklimit = 0
    >>> from inspect import isclass
    >>> assert isclass(Decoder), \
    'Decoder must be a class'
    >>> assert issubclass(Decoder, json.JSONDecoder), \
    'Decoder must inherit from `json.JSONDecoder`'
    >>> assert type(result) is dict, \
    'Result must be a dict'
    >>> assert len(result) > 0, \
    'Result cannot be empty'
    >>> assert all(type(key) is str
    ...            and type(value) in (str, datetime, list)
    ...            for key, value in result.items()), \
    'All keys must be str and all values must be either str, datetime or list'

    >>> result  # doctest: +NORMALIZE_WHITESPACE
    {'mission': 'Ares 3',
     'launch_date': datetime.datetime(2035, 6, 29, 0, 0),
     'destination': 'Mars',
     'destination_landing': datetime.datetime(2035, 11, 7, 0, 0),
     'destination_location': 'Acidalia Planitia',
     'crew': [{'name': 'Melissa Lewis', 'birthdate': datetime.date(1995, 7, 15)},
              {'name': 'Rick Martinez', 'birthdate': datetime.date(1996, 1, 21)},
              {'name': 'Alex Vogel', 'birthdate': datetime.date(1994, 11, 15)},
              {'name': 'Chris Beck', 'birthdate': datetime.date(1999, 8, 2)},
              {'name': 'Beth Johanssen', 'birthdate': datetime.date(2006, 5, 9)},
              {'name': 'Mark Watney', 'birthdate': datetime.date(1994, 10, 12)}]}
"""

import json
from datetime import date, datetime


DATA = """
    {"mission": "Ares 3",
     "launch_date": "2035-06-29T00:00:00",
     "destination": "Mars",
     "destination_landing": "2035-11-07T00:00:00",
     "destination_location": "Acidalia Planitia",
     "crew": [{"name": "Melissa Lewis", "birthdate": "1995-07-15"},
              {"name": "Rick Martinez", "birthdate": "1996-01-21"},
              {"name": "Alex Vogel", "birthdate": "1994-11-15"},
              {"name": "Chris Beck", "birthdate": "1999-08-02"},
              {"name": "Beth Johanssen", "birthdate": "2006-05-09"},
              {"name": "Mark Watney", "birthdate": "1994-10-12"}]}"""


# JSON decoder
class Decoder:
    ...


# JSON decoded DATA
# type: dict[str, str|list|datetime]
result = ...