doconf 0.2.0

Configuration specified through documentation, supporting multiple formats.

Configuration specified through documentation.

Installation

From the project root directory:

$ python setup.py install

Or from pip:

$ pip install doconf

API Usage

doconf is relatively simple to use. Simply install it, then write out a config module with a class in a similar format:

from doconf import DoconfConfig


class Config(DoconfConfig):
    '''
    name: echo_server

    {default}

    [server]
    HOST (str:"127.0.0.1"): this is the host passed to the app, by default serves locally
    PORT (int:8080): by default it serves on port 8080.
    DEBUG (bool:true): Debug mode is on by default

    [logger]
    LOG_LEVEL (str): the log level, one of debug/info/warning/error. This is required since it has no default
    LOG_PATH: we didn't specify a type so it is by default a string, and required.
    LOG_FORMAT (str:null): None by default

    {production}

    [server]
    HOST (str:"0.0.0.0"): serve to the world
    PORT (int:8081): serve on port 8081
    DEBUG (bool:false): default to false for production

    [logger]
    LOG_LEVEL (str:"INFO"): default to INFO in production
    LOG_PATH (str:"/var/log/echo_server.log"): put in /var/log/ in production
    LOG_FORMAT (str:null): don't care still
    '''
    pass


config = Config.load()

# And now we can access values from it with dictionary access.
log_level = config['logger']['LOG_LEVEL']

# It's case insensitive on section and variable names.
if 'log_path' in config['logger']:
    print('this is true, case insensitive')

In the above example, we can see that we specified a subclass of DoconfConfig, and it’s a config for an app named “echo_server”. We specify a {default} environment first (required), and a few sections and variables with their types and default values. Then we specify a {production} environment next and make a few changes to defaults.

Using that, we can now load our config wth Config.load() below the class definition. It’ll automatically look for echo_server.{conf,cfg,config} in serveral locations depending on the environment, and validate the config it finds and load it and coerce the values into the types specified. It’ll raise an error if a required value (missing default value) isn’t defined, or if the type is wrong. To figure out which paths it’d look for a config file, check the find command under the CLI Usage section.

We simply run config = Config.load() to discover and load our config, and it’ll preload it with the default values based on the default environment. We can also specify the default environment like:

config = Config.load(env='production')
assert config['server']['HOST'] == '0.0.0.0'

And we can pass in a custom path to a config like so:

config = Config.load(path='/my/custom/path.config')

It provides simple dictionary access, and is case-insensitive when matching against section or variable names.

CLI Usage

You can use the CLI tools to find where it locates its config (doconf find <classpath>), you can generate example documented configs using doconf generate <classpath>, and you can validate configuration files with doconf validate <classpath>.

Use –help/-h to view info on the arguments:

$ doconf --help

usage: doconf [-h] {find,validate,generate} ...

positional arguments:
  {find,validate,generate}
    find                find where the config file would be loaded from
    validate            validate your config files match the format
    generate            generate example config files

optional arguments:
  -h, --help            show this help message and exit

Find will show you where the config would be loaded from in the current environment:

$ doconf find --help
usage: doconf find [-h] class_path

positional arguments:
  class_path  path to the module and class, eg:
              custom_example.config:CustomConfig

optional arguments:
  -h, --help  show this help message and exit

Here we can see where the environment would discover a config specified by the class CustomConfig in the directory examples/my_example_app/config.py:

$ doconf find examples.my_example_app.config:CustomConfig

Validate will find your config and parse it, tell you whether it has all required variables and show you the values:

$ doconf validate --help | sed 's/        /    /g'
usage: doconf validate [-h] [--config-path CONFIG_PATH] [--env ENV] class_path

positional arguments:
  class_path            path to the module and class, eg:
                        custom_example.config:CustomConfig

optional arguments:
  -h, --help            show this help message and exit
  --config-path CONFIG_PATH, -c CONFIG_PATH
                        direct path to config
  --env ENV, -e ENV     the environment to use

This will validate that the config passed via –config-path matches the format, and we will see the values it sets:

$ doconf validate examples.my_example_app.config:CustomConfig --config-path examples/my_example_app/my_example_app.cfg

Generate will dump example configuration files for you to provide as examples:

$ doconf generate --help
usage: doconf generate [-h] [--out OUT] class_path

positional arguments:
  class_path         path to the module and class, eg:
                     custom_example.config:CustomConfig

optional arguments:
  -h, --help         show this help message and exit
  --out OUT, -o OUT  output directory, default to current directory

This will dump out an example documented config for the default environment and the production environment:

$ doconf generate examples.my_example_app.config:CustomConfig --out .

Release Notes

0.2.0:

• Handle multiline descriptions.

• Add simple example in ./examples

• Fix issue with sys.path when running doconf on local python modules.

0.1.2:

• Better example in API usage.

0.1.1:

• Extended README.rst with an example.

0.1.0:

• Implemented main logic, including parser and DoconfConfig class.

• Implemented CLI tools: find, validate, generate.

• Added examples.

0.0.1:

• Project created.