Skip to content

josephp27/Zeno

Repository files navigation

Zeno | An Object Config Mapper (OCM)

Build PyPI version Downloads License

I heard you like ORMs, so I got you an OCM.

The idea for this comes from ORMs like SqlAlchemy and how Spring Boot uses @ConfigurationProperties

Library Version Python
1.x.x 2.7, 3.5+
0.x (Beta) 2.7, 3.5+

Zeno maps your configs to objects for you.

Spring:
  Data:
    MongoDb:
      database: TESTDB
      encryption: true
      encryptionKey: "FakePassWord!"
      password: "!54353Ffesf34"
      replicaSet: FAKE-DB-531
    second: 1
    myList:
      - first
      - second
      - third

MyServer:
  host: my.server.com
  port: 8080

Looks like

class Spring(Configuration): #Inheriting from Configuration lets the parser know which class to modify
    """
    loads in from data.yml. accessing nested sections can be done via nested classes
    """

    class Data:
        class MongoDb:
            database = String()
            encryption = Boolean()  # conversion automatically happens when specifying the type
            encryptionKey = String()
            password = String()
            replicaSet = String()

        second = Integer()
        myList = List()

And can now be accessed like:

spring = Spring()
spring.Data.myList  # ['first', 'second', 'third']
spring.Data.MongoDb.encryption is True  # True

Autocompletion

Classes give you autocompletion in your favorite IDE for all your config files

class MyServer(Configuration):
    host = String()
    port = Integer()

Autocomplete

Don't Like Classes?

Classes are a powerful way for autocompletion, type hints and auto conversion to those types. However, if it isn't for you, calling Zeno directly can be done. The parameter to the constructor is the path within the dictionary. If no parameter is set, it will map the whole dictionary. More examples can be found here

from ZenoMapper.zeno import Zeno

zeno = Zeno('Spring.Data.MongoDb')
zeno.database  # TESTDB

Install

pip install ZenoMapper

Import

from ZenoMapper.Configuration import ConfigParser, Configuration
from ZenoMapper.Types import String, Boolean, Integer, List

Using __section__ to map super nested dictionaries/ignore nesting classes

class MyConfig(Configuration):
    __section__ = 'Spring.Data.Mongodb'
    
    database = String()
    encryption = Boolean()
    encryptionKey = String()
    password = String()
    replicaSet = String()

Using your custom configs with Zeno

Using your own configs is easy with Zeno, simply inherit from ConfigParser and instantiate the get_config() function. Load in your file and Zeno will do the rest.

class MyConfig(ConfigParser):
    """
    loading your own config is done via subclassing the ConfigParser class and implementing the
    get_config function.
    
    each time an object is instantiated, this is called, so cache the results to increase performance with the @lru_cache decorator
    """

    @staticmethod
    @lru_cache(maxsize=None)  # @cache if you're on python3.9 or later
    def get_config():
        with open("data.yml", 'r') as stream:
            return yaml.safe_load(stream)

Types

Zeno currently has 4 types, where auto conversion will happen based on the specified type. It also brings a little static typing to Python. The plan is to add more, but currently Zeno supports: Integer, String, Boolean, List. Supported types can be found here

If you have another type you'd like but it isn't supported, etc. None can be used to tell Zeno to not handle conversion

Custom Types

Users can specify their own types by inheriting from ConfigTypes

class AppConf(ConfigTypes):
    """
    AppConf class to be used with Configuration class
    """

    def __init__(self, name, to_type):
        self.name = name
        self.convert_to_type = to_type

    def convert(self, obj):
        """Method called to convert obj to its specified convert type
        NOTE: if the environment is production, it will first get it from appconf
              and override the one in the config file

        Required Args:
            obj (obj) - The object to be converted

        returned:
            obj - The object as the specified type
        """
        if os.environ.get('ENVIRONMENT') == 'production':
            # we are in production and using appconf to get our values
            obj = parse_appconf()[self.name]

        # now convert it to the specified to_type
        return self.convert_to_type().convert(obj)

and then can be called like the other types

class Database(Configuration):
    """section for database"""
    __project__ = 'my_project'

    dbname = String()
    # you can even send in another type to do a later conversion within your new type
    host = AppConf('host', String)
    port = AppConf('port', Integer)

Choosing what to map

Zeno is powerful. Only specify what you'd like to map and Zeno ignores the rest

class Spring(Configuration):
    class Data:
        class MongoDb:
            database = String()

Hold up, that's nice but I still like using dictionary methods

Well then person reading this, Zeno is for you. All Classes are dictionary class hybrids, so you can access them like plain old dictionaries when necessary.

spring.Data.myList # ['first', 'second', 'third']
spring['Data']['myList'] # ['first', 'second', 'third']
spring['Data'].myList # ['first', 'second', 'third']
spring.Data['myList'] # ['first', 'second', 'third']
spring # {'Data': {'MongoDb': {'database': 'TESTDB', 'encryption': True, 'encryptionKey': 'FakePassWord!', 'password': '!54353Ffesf34', 'replicaSet': 'FAKE-DB-531'}, 'second': 1, 'myList': ['first', 'second', 'third']}}

Don't Break Python Naming Conventions!

Classes can fuzzy match while not breaking Python's class naming conventions

lower:
  case_section: true
class Lower(Configuration):
    CaseSection = Boolean()

Supports

  • Any parser
    • YAML
    • INI
    • etc, as long as it parses into a dictionary
  • Automatic type conversion
  • Custom conversion classes
  • All Python 3 versions and 2.7