dehydrate 0.1.3

Small lib for representing python objects as a dicts

python-dehydrate

Small lib for representing python objects as a dicts.

Motivation

Why would you need library like this? One of obvious use cases is to prepare data for serializing (into json/yaml/xml/pickle/whatever). You can control dehydration process by describing how to fetch values from object and how to present it in dehydrated structure using simple syntax.

Examples

Simple cases

In simplest of possible cases you just want get object, list wanted attributes and get mapping with keys based on attribute names and values from them. Use dehydrate shortcut for this case:

>>> from dehydrate import dehydrate
>>> from examples import Person
>>> iron_man = Person(first_name='Tony', login='iron_man')
>>> dehydrated = dehydrate(obj=iron_man, fields=('first_name', 'login'))
>>> sorted(dehydrated.items())
[('first_name', 'Tony'), ('login', 'iron_man')]

Some notes:

• I use list representation of dict in examples because it has predictable order of items in it.

• In docs I will refer to examples package, which you can find in repo.

If requested attribute name resolves to method of object, then result of calling it will be set in dehydrated dict. In Person class we have method full_name, so lat’s try to get its return value:

>>> from dehydrate import dehydrate
>>> from examples import Person
>>> iron_man = Person(first_name='Tony', last_name='Stark')
>>> dehydrated = dehydrate(obj=iron_man, fields=('full_name',))
>>> sorted(dehydrated.items())
[('full_name', 'Tony Stark')]

But what if you want put first_name attribute in name key of resulted dict? Just specify both strings in fields. (I will refer to this description of a field as a spec (which can be one object or two-tuple)):

>>> from dehydrate import dehydrate
>>> from examples import Person
>>> iron_man = Person(first_name='Tony', login='iron_man')
>>> dehydrated = dehydrate(obj=iron_man, fields=(
...     ('first_name', 'name'),
...     'login',
... ))
>>> sorted(dehydrated.items())
[('login', 'iron_man'), ('name', 'Tony')]

Second argument always be used as a key if exists in spec.

More complex cases

Sometimes you will want to add some value in dehydrated dict, which is not attribute of dehydrated object. Or you may want not use attribute and add some another handling for this field instead. In our example we creating special class for this called PersonDehydrator (inherited from dehydrate.Dehydrator) and set some methods on it:

>>> from examples import Person, PersonDehydrator
>>> iron_man = Person(password='iRon42', login='iron_man')
>>> dehydrated = PersonDehydrator(fields=(
...     'password',
...     ('superhero_status', 'is_superhero'),
... )).dehydrate(obj=iron_man)
>>> sorted(dehydrated.items())
[('is_superhero', True), ('password', '******')]

In example you can see, that object has password attribute, but PersonDehydrator’s get_password used for password spec. Also you can mention, that result of calling get_superhero_status was set in key is_superhero because of second item in spec was declared. You can declare list of spec (fields) using attribute of dehydrator class or by passing argument into its __init__ method.

Recursive dehydration

The most valuable feature of lib is that you can describe how to recursively dehydrate complex fields on object:

>>> from dehydrate import dehydrate
>>> from examples import Person, PersonDehydrator
>>> octopus = Person(login='octopus')
>>> spider_man = Person(login='spidey', archenemy=octopus)
>>> dehydrated = dehydrate(
...     fields=(
...         'login',
...         {'target': 'archenemy', 'fields': ('login',)}
...     ),
...     obj=spider_man
... )
>>> dehydrated['login']
'spidey'
>>> list(dehydrated['archenemy'].items())
[('login', 'octopus')]

Second spec in fields is so-called ComplexField, it described by mapping with one required key target, which describes how to get value for serialization. Other acceptable keys are:

• dehydrator — class, which can be used for dehydrating of complex target.

• fields — iterable of same structure as described above.

• iterable — flag, which specifies should target be handled as iterable.

Installation

Simple:

pip install dehydrate

must be fine.

Requirements

• six (did I mentioned python 3 support? We have one.)

Philosophy

• Easy things should be done easily.

• Complex things must be possible.

Testing

Test written with use of pytest library and neat pytest pep8 plugin. You should run python setup.py test for running full test suite or coverage run --source=dehydrate setup.py test for tests with coverage. Tests automatically runs at Travis CI. Examples in documentation are also picked by test command.

Contribution

Any contribution is welcome. Use fork/pull request mechanism on github.

If you add some code, you should add some tests, so coverage of master branch should always be 100%. Refer to Testing section for more instructions.

Let me speak from my heart :). I will be very glad, if you correct my clumsy english phrases in docs and docstings or even advise more appropriate names for variables in code.

TODO

• Think about giving opportunity to put results in Ordered dict instead of simple dict.

• Add functionality for converting all values of some type using handlers on dehydrator class.

• Rename fields in specs for minimal confusing in terms.

• Review tests, because now they not very maintainable. Use examples like in readme.

• Add comprehensive about everything.

Changelog