figs 0.1.1

A towel wrapped ConfigParser API.

Figs

Figs is a library for reading ini like configuration files easily. Figs leverages the ConfigParser module from python’s standard libraries.

I personally don’t like the ConfigParser API very much, so I wrote this. The idea is that the developer should have little overhead in thinking when using this library, i.e., an intuitive API.

Usage

If you are familiar with PyYaml or the standard library’s json modules, the following should be quite familiar to you.

To load a configuration, use the load/loads functions. The following return the same:

>>> # Takes a filename
>>> conf = figs.load('config.ini')

>>> # or a file-like object
>>> conf = figs.load(open('config.ini'))

>>> # Takes a string to be parsed
>>> conf = figs.loads('''\
        [universe]
        answer = 42
        is_active = yes
        status = expanding
        ''')

And, to dump configuration:

>>> # Takes a filename
>>> figs.dump(conf, 'config.ini')

>>> # or a file-like object
>>> figs.dump(conf, open('config.ini'))

>>> # Dump to string
>>> figs.dumps(conf)
[universe]
answer = 42
is_active = yes
status = expanding

>>> # You can also dump just a section
>>> figs.dumps(conf.universe)
answer = 42
is_active = yes
status = expanding

Those are the only functions in the figs module that you should be concerned with.

Once you have the config object, how’d you use it? Surprise surprise! Anyway you feel comfortable :)

Dictification

I know, you just want a dict of properties from the config file and be done with it. Lets see if you can guess how this can be done?:

>>> # Returns a dict like {'section-name': <Section object>}
>>> dict(conf)

>>> # Returns a dict like {'key': <TypeableStr object>}
>>> dict(conf.universe)

You should keep in mind that dict on does not automatically do a dict on its Section objects. The TypeableStr class is a subclass of unicode with a few methods added (as_bool, as_int and as_float).

If you want a dict of dicts, though, you can get that too.:

>>> figs.as_dict(conf)

>>> # or when loading
>>> conf = figs.load('config.ini', as_dict=True)

The loads method also takes the as_dict argument. Note that the as_dict has to be a keyword argument.

Accesses

>>> conf.universe.answer
u'42'
>>> conf.universe.answer.as_int
42
>>> conf.universe.is_active
u'yes'
>>> conf.universe.is_active.as_bool
True
>>> conf.universe.status
u'expanding'
>>> conf.universe['status']
u'expanding'

Similary to as_int as shown above, there are also as_bool (boolean conversion done similar to how ConfigParser.getboolean does) and as_float.

Check for presence

>>> 'universe' in conf
True
>>> 'multiverse' in conf
False
>>> 'answer' in conf.universe
True
>>> 'is_active' in conf.universe
True
>>> 'is-active' in conf.universe
False

Modifying configs

Set new options…:

>>> conf.universe.is_active = False
>>> conf.universe.planet_maker = 'Magrathea'
>>> conf.universe['earth-owners'] = 'mice'
>>> figs.dumps(conf)
[universe]
answer = 42
is_active = false
status = expanding
planet_maker = Magrathea
earth-owners = mice

…on new sections:

>>> conf.multiverse.is_active = True
>>> figs.dumps(conf)
[universe]
answer = 42
is_active = false
status = expanding

[multiverse]
is_active = true

Deleting

The API is very boring isn’t it?:

>>> del conf.universe.answer
>>> del conf.multiverse

Now what?

Well, if you have a life, get on with it. Seriously, there’s nothing else to reading config files here.

Meta

Author

Shrikant Sharat (http://sharats.me). @sharat87 on twitter.

License

MIT License (http://mit.sharats.me).

Contributing

Code is available at the bitbucket repository. Clone. Modify. Send pull request. If the modification is fairly large, I prefer you open a bitbucket issue first to discuss it.