python-dependency-injector/docs/catalogs.rst
2015-08-04 17:05:34 +03:00

59 lines
1.9 KiB
ReStructuredText

Catalogs
========
Catalogs are collections of providers. Main purpose of catalogs is to group
providers.
There are, actually, several popular use cases of catalogs:
- Grouping of providers from same architectural layer (for example,
``Services``, ``Models`` and ``Forms`` catalogs).
- Grouping of providers from a same functional components (for example,
catalog ``Users``, that contains all functional parts of ``Users``
component).
Writing catalogs
----------------
Catalogs have to be created by extending base catalog class
``objects.catalog.AbstractCatalog``.
Providers have to be defined like catalog's attributes. Every provider in
catalog has name. This name should follow ``some_provider`` manner, that is
standard naming convention for names of attributes in Python.
.. note::
It might be useful to add such
``""":type: (objects.Provider) -> Object1"""`` documentation blocks one
line after provider definition for every provider. It will help code
analysis tools and IDE's to understand that variable above contains some
callable object, that returns particular instance as a result of call.
Example:
.. literalinclude:: ../examples/catalogs/simple.py
:language: python
Overriding of catalogs
----------------------
Catalogs can be overridden by other catalogs. This, actually, means that
all of the providers from overriding catalog will override providers with the
same names in overridden catalog.
There are two ways to override catalog by another catalog:
- Use ``Catalog.override(Catalog)`` method.
- Use ``@override(Catalog)`` class decorator.
Example of overriding catalog using ``Catalog.override()`` method:
.. literalinclude:: ../examples/catalogs/override.py
:language: python
Example of overriding catalog using ``@override()`` decorator:
.. literalinclude:: ../examples/catalogs/override_decorator.py
:language: python