Skip to main content

Configuration file reader with some nifty bells and whistles added

Project description

Alviss the all-knowing

Generic Configuration file reader/parser with a bunch of nifty extra features, most notably:

JSON & YAML Configuration file reader/parser with built-in fault tolerance and ENV variable injecting, structure inheritance, internal referencing and secret injection.

2.6 Features

2.1 Features

  • JSON & YAML config file reading
    • Files can extend base file for easy overriding and customization
    • Files can include multiple other files for better organization and structure.
    • Includes/extends can be in a nested structure
  • Simple yet powerful internal variable reference parsing
    • Inject values from elsewhere in the config files anywhere
    • Automatic injection of environment variables for values
  • Nested fault tolerant fetching of non-existing values
  • Quickly nest keys into one with dots
  • Easy default value usage
  • Password and secret masking
  • One line config loading
  • Singleton config model
  • Extendable for multiple singleton configs
    • Removed from the basic class in 2.0. Singleton functionality can be added by simply extending the alviss.structs.BaseConfig on the application level and adding the typeutils.metas.Singleton as a metaclass
    • The alviss.structs.SingletonConfig sill behaves as the old singleton base guy did if people want.

Singleton reloading/update/set

Reloading (or loading, setting or updating different values) anywhere in code will change them globally wherever the same model is referenced.

Update in one place...

# Somewhere in the program we decide that reloading the config is in order
from alviss.structs import SingletonConfig

cfg = SingletonConfig()
cfg.load(foo='Foo Too You!', bar='Salad bar vs. Tequila bar!')

# This will be the same config
cfg2 = SingletonConfig()
assert cfg == cfg2

Updated everywhere...

# Anywhere the config is needed...
from alviss.structs import SingletonConfig

cfg = SingletonConfig()
assert cfg.foo == 'What is a Foo anyway?'

Updating methods

Loading/Reloading

Calling BaseConfig.load(**kwargs) will clear the entire config data map and load up the new values given.

from alviss.structs import BaseConfig

# Initial values...
cfg = BaseConfig()
cfg.load(foo='Foo Too You!', bar='Salad bar vs. Tequila bar!')

assert cfg.foo == 'Foo Too You!'
assert cfg.bar == 'Salad bar vs. Tequila bar!'

# ...now we reload...
cfg.load(foo='Foo is not a real word!')

# ...so now there is no `bar` in the config
assert cfg.foo == 'What is a Foo anyway?'
assert not cfg.bar

Giving the BaseConfig constructor **kwargs does the same as calling BaseConfig.load(**kwargs) and will therefor clear and reload the config values from the supplied keyword arguments.

from alviss.structs import BaseConfig

# Initial values...
cfg = BaseConfig()
cfg.load(foo='Foo Too You!', bar='Salad bar vs. Tequila bar!')

# This does the same thing
cfg = BaseConfig(foo='Foo Too You!', bar='Salad bar vs. Tequila bar!')

Setting values directly

from alviss.structs import BaseConfig

# Initial values...
cfg = BaseConfig()
cfg.load(foo='Foo Too You!', bar='Salad bar vs. Tequila bar!')

# Updated only the `foo` value, but on a singleton level so it changes everywhere
cfg.foo = 'Nyoo Foo!'

Updating the config

Calling update(**kwargs) will update only the given keyword values of the config (globally via singleton magic as before) but preserves the existing values of the config that did not change (while load(**kwargs) wiped the config clean first).

from alviss.structs import BaseConfig

# Initial values...
cfg = BaseConfig()
cfg.load(foo='Foo Too You!', bar='Salad bar vs. Tequila bar!')

# Updated only the given values (leaving `bar` unchanged)
cfg.update(foo='Nyoo Foo!', moo='Shoo!')

Extendable for multiple singleton configs or other magic

You can use the alviss.structs.SingletonConfig as a base for implementing other config classes in order to store different singleton instances of different configs.

The singleton mechanism supports inheritance so the new Classes will have their own singleton instances separate from each other and from SingletonConfig.

from alviss.structs import SingletonConfig


class ClientConfig(SingletonConfig):
    pass


class ServerConfig(SingletonConfig):
    pass


client_cfg = ClientConfig(host='localhost')

server_cfg = ServerConfig(host='127.0.0.1')

assert client_cfg != server_cfg

Quick JSON/YAML Config File Loading

User the alviss.quickloader to load a BaseConfig from the given JSON or YAML file.

The config.json file...

{
    "foo": "Foo Too You!",
    "bar": "Salad bar vs. Tequila bar!"
}

Loading from config.json...

from alviss import quickloader

cfg = quickloader.autoload('config.json')

assert cfg.foo == 'Foo Too You!'

Config file extending/including example

Alviss now supports extending base config files and including files via the __extends__ and __include__ config keys respectively.

Extending a base file

Using the __extends__ key you can embed one or more additional config files wholesale into another. This works from within the nested key hierarchy so extended files within nested keys/sections will inject the values there.

Values of keys duplicated in both the current file and the extended files defined in __extends__ will retain the value from the current file (i.e. the current values override those from the extended base files).

The __extends__ key can take either a single string file name or a list of files.

The base.json file...

{
    "foo": "Foo Too You!",
    "bar": {
        "first": "Beer",
        "second": "Wine"
    }
}

The production.json file...

{
    "__extends__": "base.json",
    "bar": {
        "second": "Tequila"
    }
}

Loading base.json and overriding with production.json...

from alviss import quickloader

# Assume this comes from an environment variable for example
CONFIG_FILE = 'production.json'

# Loading the base config...
cfg = quickloader.autoload(CONFIG_FILE)

assert cfg.foo == 'Foo Too You!'
assert cfg.bar.first == 'Beer'
assert cfg.bar.second == 'Tequila'

Including additional files

Using the __include__ key you can embed one or more additional config files wholesale into another. This works from within the nested key hierarchy so including files within nested keys/sections will inject the values there.

Values of keys duplicated in both the current file and the included files defined in __include__ will retain the value from the last included file (i.e. the included values override those from the current one).

The __include__ key can take either a single string file name or a list of files.

The logging.yaml file...

level: INFO
file: stuff.log

The db.yaml file...

database:
  driver: mysql
  name: mydata

The extra.yaml file...

bar: Foo

The app.yaml file...

foo: Bar
logging:
  # Default level is DEBUG
  level: DEBUG
  __include__: logging.yaml

bar: NotFoo
__include__: 
- db.yaml
- extra.yaml

Loading app.yaml...

from alviss import quickloader

# Loading the config...
cfg = quickloader.autoload('app.yaml')

assert cfg.foo == 'Bar'
assert cfg.logging.level == 'INFO'
assert cfg.database.driver == 'mysql'
assert cfg.bar == 'Foo'

Quickly reference a nested key with a single key string

You can reference a nested key (e.g. for overriding in an include/extension ) by using dots to separate keys in the path.

The config.json file...

{
  "database": {
    "settings": {
      "connection": {
        "host": "1.2.3.4",
        "port": 1234      
      }   
    }    
  }
}

The local.json file...

{
  "__extends__": "config.json",
  "database.settings.connection.host": "localhost"
}

Loading local.json...

from alviss import quickloader

# Loading the config...
cfg = quickloader.autoload('local.json')

assert cfg.database.settings.connection.host == 'localhost'

Internal variable referencing

You can reference other config value keys from within any values by using ${some.other.key} from any string value, using dots to navigate key nestings.

The config.json file...

{
  "basics": {
    "env": "prod"
  },
  "db": {
    "name": "my-data-${basics.env}"    
  }
}

Loading config.json...

from alviss import quickloader

# Loading the config...
cfg = quickloader.autoload('config.json')

assert cfg.db.name == 'my-data-prod'

Automatic injection of environment variables for values

You can automatically inject any environment variables set on the system by using ${__ENV__:SOME_ENV_VAR} from within any string value.

The config.yaml file...

user: Foo
password: ${__ENV__:MY_PASS}

Loading from config.yaml...

import os

os.environ.update('MY_PASS', 'SuperSecret!')

from alviss import quickloader

# Loading the config...
cfg = quickloader.autoload('config.yaml')

assert cfg.password == 'SuperSecret!'

Nested fault tolerant fetching of non-existing values

Alviss config models will return an Empty and/or EmptyDict) from typeutils if asked for a value that was not in the config file read.

This means that you will never get a KeyError or other such exceptions when looking up config values, even if search multiple nested levels.

from alviss.structs import *

# Load config...
cfg = BaseConfig(foo={'alpha': 42})

assert cfg.foo.alpha == 42
assert not cfg.foo.beta
assert not cfg.bar
assert not cfg.bar.alpha
assert not cfg.a.b.c.d.e.f

To check explicitly if a value has not been set (vs. it's value just being an empty string for example) you can check if the value is typeutils.empty.Empty

from alviss.structs import *  # Includes a reference to `typeutils.empty.Empty`

# Load config...
cfg = BaseConfig(foo={'alpha': 42})

assert cfg.foo.alpha == 42
assert cfg.foo.beta is Empty
assert cfg.bar is Empty

Easy default value usage

Empty object evaluate as False in boolean comparisons so you can ensure default values with an or bitwise operator.

from alviss.structs import *

# Load config...
cfg = BaseConfig(foo='Foo Too You!')

# Get value or default
my_bar = cfg.bar or 'There was no bar'

assert my_bar == 'There was no bar'

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

alviss-3.2.2.tar.gz (24.6 kB view hashes)

Uploaded Source

Built Distribution

alviss-3.2.2-py3-none-any.whl (27.9 kB view hashes)

Uploaded Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page