A wrapper for combining yaml configuration files and command line argument parsing. It builds upon PyYAML for parsing yaml config files and argparse for handling command line arguments and adds the clue produce one unified program configuration.

• Python 2.7+ or 3.2+
• PyYAML >= 3.1

Because configuration often differs depending on the environment, e.g. production and testing settings, yconf allows you to define multiple environments in a single configuration. The current supported environemts are production, staging and development. For convenience, yconf is able to merge environment configuration to reduce redundancy, meaning that if the environment is set to staging, the configuration will inherit all settings available in a production environment and override them with staging values where neseccary.

config.yml

production:
  rundir: /var/run/example
  database: /var/lib/example/db.sqlite
  loglevel: warning

staging:
  database: :memory:

development:
  loglevel: debug

Configuration

# production environment
config.rundir: /var/run/example
config.database: /var/lib/example/db.sqlite
config.loglevel: warning

# staging environment
config.rundir: /var/run/example
config.database: :memory:
config.loglevel: warning

# development environment
config.rundir: /var/run/example
config.database: :memory:
config.loglevel: debug

Both staging and development environments are optional.

If a directory is specified instead of a configuration file, yconf will look for configuration files by convention of <environment>.yml in that directory and merge them accordingly if possible.

Extending the argument list supported by a configuration will require subclassing yconf.BaseConfiguration and overriding the makeParser method.

Example

import sys
from yconf import BaseConfiguration

class MyConfig(BaseConfiguration):

  def makeParser(self):
    parser = super(MyConfig, self).makeParser()
    parser.add_argument("-f", "--foo", dest="foo")
    return parser


config = MyConfig()
config.parse(sys.argv[1:])

The following configuration arguments are already preset:

• -c Path to configuration file or directory
• -e The environment used for configuration (default: production)

When present, command line argument always take precedent over configuration settings. To override nested yaml values, a dotted destination path can be set for the argument.

config.yml

logging:
  loglevel: debug

Argument

parser.add_argument("--log-level", dest="logging.loglevel")

Configuration can be accessed either by attribute or by key.

config = MyConfig()
config.parse(args)

config.foo == config['foo']
config.foo.bar == config['foo']['bar']