waterbutler package

waterbutler.constants module

Constants

waterbutler.settings module

class waterbutler.settings.SettingsDict(*args, parent=None, **kwargs)

Bases: dict

Allow overriding on-disk config via environment variables. Normal config is done with a hierarchical dict:

"SERVER_CONFIG": {
  "HOST": "http://localhost:7777"
}

HOST can be retrieved in the python code with:

config = SettingsDict(json.load('local-config.json'))
server_cfg = config.child('SERVER_CONFIG')
host = server_cfg.get('HOST')

To override a value, join all of the parent keys and the child keys with an underscore:

$ SERVER_CONFIG_HOST='http://foo.bar.com' invoke server

Nested dicts can be handled with the .child() method. Config keys will be all parent keys joined by underscores:

"SERVER_CONFIG": {
  "ANALYTICS": {
    "PROJECT_ID": "foo"
  }
}

The corresponding envvar for PROJECT_ID would be SERVER_CONFIG_ANALYTICS_PROJECT_ID.

get(key, default=None)

Fetch a config value for key from the settings. First checks the env, then the on-disk config. If neither exists, returns default.

get_bool(key, default=None)

Fetch a config value and interpret as a bool. Since envvars are always strings, interpret ‘0’ and the empty string as False and ‘1’ as True. Anything else is probably an acceident, so die screaming.

get_nullable(key, default=None)

Fetch a config value and interpret the empty string as None. Useful for external code that expects an explicit None.

get_object(key, default=None)

Fetch a config value and interpret as a Python object or list. Since envvars are always strings, interpret values of type str as JSON object or array. Otherwise assume the type is already a python object.

full_key(key)

The name of the envvar which corresponds to this key.

child(key)

Fetch a sub-dict of the current dict.

waterbutler.settings.child(key)

waterbutler.sizes module

A utility module for writing legible static numbers >>> 10 * MBs >>> 6 * GBs

waterbutler.version module

Module contents