Skip to content
/ ntgen Public

Generate NamedTuple code based on data samples

License

Notifications You must be signed in to change notification settings

mrapacz/ntgen

Repository files navigation

ntgen - named tuple generator

CircleCI PyPI - Package Version PyPI - Python Version Coverage Status PyPI - License

Generate NamedTuple definitions with typehints based on your data automatically. If you've ever felt like preparing NamedTuple skeletons for any json data you're dealing with is tedious and could be automated, well, this is the tool that automates the process.

Usage

Let's say you want to prepare a NamedTuple definition for the following json object:

$ cat apartment.json
{
    "id": "1234-1234",
    "type": "living",
    "isAvailable": true,
    "countryCode": "DE",
    "address": {
        "borough": "Dulsberg",
        "city": "Hamburg",
        "houseNumber": "2",
        "latitude": 53.587485,
        "longitude": 10.063215,
        "postalCode": "22049",
        "streetName": "Nordschleswiger Strasse",
        "area": "Hamburg"
    },
    "_attachments": "attachments/",
    "_ts": 15828103462
}%

All you need to do is run the following command:

$ ntgen apartment.json
class Address(NamedTuple):
    borough: str
    city: str
    house_number: str
    latitude: float
    longitude: float
    postal_code: str
    street_name: str
    area: str


class Apartment(NamedTuple):
    id: str
    type: str
    is_available: bool
    country_code: str
    address: Address
    attachments: str
    ts: int

The output will be directed to stdout by default - you may also redirect it to a file to bootstrap a Python module with the class definitions.

Runtime configuration

To find out about all of the runtime configuration options, run:

$ ntgen --help
usage: ntgen [--out OUT] [--name NAME] [-s] [-c] [-f] [-t]
             [--max_level MAX_LEVEL] [-h]
             input

positional arguments:
  input                 (str, default=None) Json file containing an object
                        with the data to analyzed

optional arguments:
  --out OUT             (Union[str, NoneType], default=None) Destination file
                        to write the Python code to
  --name NAME           (str, default=NTGenTuple) Name of the main NamedTuple
  -s, --snake-case      (bool, default=True) Convert the NamedTuple field
                        names to snake_case
  -c, --camel-case      (bool, default=True) Convert the NamedTuple class
                        names to CamelCase
  -f, --constructors    (bool, default=False) Insert generic methods that will
                        allow for parsing of the analyzed data structures
  -t, --as-dict         (bool, default=False) Insert generic methods allowing
                        for dumping the nested NamedTuple hierarchy to a dict
  --max_level MAX_LEVEL
                        (Union[int, NoneType], default=None) Specify the max
                        nesting level of the NamedTuple
  -h, --help            show this help message and exit

Other invocation options

You can also use the library from the Python context:

>>> from ntgen import generate_from_dict
>>> data = {'name': 'John Wick', 'profession': 'assassin', 'age': 34}
>>> print(generate_from_dict(data=data, name="Character"))
class Character(NamedTuple):
    name: str
    profession: str
    age: int

Installation

You'll need to be running Python >= 3.6.

pip install ntgen

Verify that the latest package version was installed correctly:

>>> import ntgen
>>> ntgen.__version__
'0.3.0'

License

This project is licensed under the MIT License - see the LICENSE file for details

Author

Maciej Rapacz