Name: di-py
Owner: Telefónica I+D
Description: Dependency injection in python #di #python
Created: 2015-03-22 13:09:22.0
Updated: 2018-04-23 11:53:12.0
Pushed: 2018-03-05 01:28:20.0
Homepage: null
Size: 67
Language: Python
GitHub Committers
| User | Most Recent Commit | # Commits |
|---|
Other Committers
| User | Most Recent Commit | # Commits |
|---|
Dependency injection library for python.
Dependency injection enables us to write better code, regardless of doing so manually or with the help of some library. It helps specially to keep our code inside the Law of Demeter, by programming against a known interface without having to worry about how a concrete implementation is composed.
Python is a highly dynamic language with an “open class” implementation for user types, thus the need for a full blown dependency injection framework is not specially needed. For medium to large applications though there is still the issue of how to actually implement dependency injection in the code using only Python's standard syntax/library.
This library is designed to be very lightweight and flexible as to allow its use in a variety of scenarios, including their use to aid with unit testing. It doesn't form a framework but just a set of utilities to keep the dependency injection needs in a project under control by applying it only where it makes sense, with minimum overhead and a lean learning curve.
http.client import HTTPSConnection
di import injector
eate the decorator setting up the dependencies (at configuration time)
ct = injector({
TPSConnection: HTTPSConnection('localhost', '8080')
ply the decorator to our app logic to inject what we have configured (at runtime)
ect
fetch_it(id, http=HTTPSConnection):
tp.request("GET","/?id={0}".format(id))
turn http.getresponse()
ll the logic without worrying about dependencies :)
t fetch_it(100).status
erride the dependency if we have some specific use case
t fetch_it(100, http=HTTPSConnect('google.com', '80')).status
rt hashlib
di import injector, Key, DependencyMap
tup the dependency map
DependencyMap()
fine a custom Key to map a dependency when it's not a class
Dep = Key('hash')
ild a dependency programatically but only the first time it's used
singleton(HashDep)
hash(deps):
turn lambda x: hashlib.md5(x).hexdigest()
eate the decorator and bind it to the dependency map
ct = injector(dm)
fine our logic defining what should be injected by default
ect
hasher(subject, hash=HashDep)
turn hash(subject)
t hasher('foobarbaz')
Those are the python interpreters being validated via travis.ci upon every change in the repository.
install di-py
install git+ssh://git@github.com/telefonicaid/di-py.git@master
Generate RPM from source code
clone git@github.com/telefonicaid/di-py.git
i-py
on setup.py bdist_rpm
See LICENSE
Use the GitHub's pull request and issue tracker to provide patches or report problems with the library. All new functionality must be covered by unit tests before it can be included in the repository.
The master branch always has the cutting edge version of the code, if you are using it in your project it would be wise to create a fork of the repository or target a specific tag/commit for your dependencies.