zsl application framework for web based services
Project description
ZSL - z' service layer
ZSL is a Python micro-framework utilizing dependency injection for creating service applications on top of Flask web framework and Gearman job server or Celery task queue.
Motivation
We developed ZSL to modernize our workflow with maintaining our clients' mostly web applications written in various older CMS solutions without the need to rewrite them significantly. With ZSL we can write our new components in Python, with one coherent shared codebase, accessible trough Gearman or JavaScript. Also the same code can be called through various endpoints - web or task queue nowadays.
Disclaimer
At current stage this should be taken as proof of concept. We don't recommend to run in any production except ours. It is too rigid, with minimum test coverage and lots of bad code practices. We open sourced it as way of motivation for us to make it better.
Installation
We recommend to install it trough PyPi and run it in a virtualenv or docker container.
$ pip install zsl
Getting started
For now it is a bit cumbersome to get it running. It has inherited settings trough ENV variables from Flask and has a rigid directory structure like django apps. On top of that, it needs a database and Redis.
The minimum application layout has to contain:
.
├── app # application sources
│ ├── __init__.py
│ └── tasks # public tasks
│ ├── hello.py
│ └── __init__.py
├── settings # settings
│ ├── app_settings.cfg
│ ├── default_settings.py
│ └── __init__.py
└── tests
$ export ZSL_SETTINGS=`pwd`/settings/app_settings.cfg
# settings/app_settings.cfg
TASKS = TaskConfiguration()\
.create_namespace('task')\
.add_packages(['app.tasks'])\
.get_configuration()
RESOURCE_PACKAGE = ()
DATABASE_URI = 'postgresql://postgres:postgres@localhost/postgres'
DATABASE_ENGINE_PROPS = {}
SERVICE_INJECTION = ()
REDIS = {
'host': 'localhost',
'port': 6379,
'db': 0
}
RELOAD = True
# hello.py
class HelloWorldTask:
def perform(self, data):
return "Hello World"
$ python -m zsl web
* Running on http://127.0.0.1:5000/ (Press CTRL+C to quit)
$ curl http://localhost:5000/task/hello/hello_world_task
Hello world!
Deploying
Deploy will happen upon pushing a new tag to Gitlab.
Creating new tag/version
Use bump2version to update version in config files. It will also create commit and new tag.
$ bumpversion --new-version ${VERSION} {major|minor|patch} --tag-name ${VERSION}
Version name uses semver. Starts with number.
Pipeline
Current pipeline tries to copy previous Travis runs. It runs tox target seperately and on a tag push will create deploy.
Tox Docker image
Gitlab pipeline runs inside a docker image which is defined in docker/Dockerfile.tox
. Currently we manually configure, build and push it to gitlab container registry. So to update the container follow this steps.
When pushing for the first time run, you have to create an access token and login to atteq gitlab container registry. Go to https://gitlab.atteq.com/atteq/z-service-layer/zsl/-/settings/access_tokens and create a token to read/write to registry. Then run
docker login registry.gitlab.atteq.com:443
To build/push the image:
-
Build image locally.
docker build -t zsl/tox-env -f docker/Dockerfile.tox .
-
Tag image.
docker tag zsl/tox-env registry.gitlab.atteq.com:443/atteq/z-service-layer/zsl/tox-env:latest
-
Push image.
docker push registry.gitlab.atteq.com:443/atteq/z-service-layer/zsl/tox-env:latest
-
Update image hash in
.gitlab-ci.yml
. (copy from build output ordocker images --digests
).
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.