# OOWS
[](https://travis-ci.org/fbidu/oows) [](https://badge.fury.io/py/OOWS)
**Warning — this project is in early development. Contributions and feedbacks are welcome. Usage in production systems is discouraged**
OOWS (_oh-owls_) is an object-oriented friendly client for Amazon Web Services — AWS. It is based on [boto3](https://github.com/boto/boto3) official client and aims to provide a cleaner and more Pythonic interface to handle AWS resources.
Currently, Boto3 already provides [the Resource Model](https://boto3.readthedocs.io/en/latest/reference/core/resources.html) which is a nicer and more OO-friendly interface over raw JSON, [as explained here](https://github.com/boto/boto3/issues/112#issuecomment-100377492). Although it works, it does not provide an interface for resources I currently need. This project started by a personal need of mine to ease the current state of programmatic access to ECS resources.
## Rationale
Suppose you want to list all of a cluster's services' task definitions. Using pure `boto3`, you'd need to write something like:
```python
session = boto3.Session()
ecs_client = session.client('ecs')
services = ecs_client.list_services(cluster="MyCluster")['serviceArns']
for service in services:
s = ecs_client.describe_services(cluster="MyCluster", services=[service])
print("The task definition is {}".format(s['services'][0]['taskDefinition']))
```
But using `oows`, you can rewrite this code to something like:
```python
cluster = Cluster("MyCluster")
for service in cluster.services:
print("The task definition is {}".format(service.task_definition))
```
## Quick Start
Currently, OOWS supports only a few ECS components and operations.
First, install it with PIP
`pip install boto3 oows`
As of now, you'll need to supply a regular [boto3 Session](https://boto3.readthedocs.io/en/latest/reference/core/session.html)
```python
import boto3
from oows import ecs
s = boto3.Session() # Create a new boto3 Session.
cluster = ecs.Cluster("MyCluster", s) # Initialize a new cluster object
service = ecs.Service("server", cluster, s) # Initialize a new service object
task_definition = service.task_definition # Gets the service's task definition
task_definition.update_env("new_env", "new_value") # Creates a new task definition with updated env
```
