scaleway/python-scaleway

:snake: Python SDK to query Scaleway APIs.

Scaleway SDK

Python SDK to query Scaleway's APIs.

Stable release:

Development:

Installation

The package is available on pip. To install it in a virtualenv:

$ virtualenv my_virtualenv
$ source my_virtualenv/bin/activate
$ pip install scaleway-sdk

General principle

If you're looking to send a GET HTTP request against our APIs, like:

GET <api_url>/foo/bar

you only need to call the following pythonic code:

>>> from scaleway.apis import DummyAPI
>>> DummyAPI().query().foo.bar.get()

The magic here lies in scaleway.apis.*API instances, which all have a query method returning a slumber.API object. The latter handling all the excruciating details of the requests.

Documentation

Even if this SDK is designed to be developer-friendly and aim for self-service discovery, it is still recommended to read the official API documentation.

And because most of the provided helpers takes the form of pre-configured Slumber objects, a good read of Slumber documention is encouraged as well.

The list of available resources per API can be found [on the Scaleway API repository](https://github.com/scaleway/api.scaleway.com/blob/master/README.md#apis)

Examples

• List your organizations:

>>> from scaleway.apis import AccountAPI
>>> api = AccountAPI(auth_token='')  # Set your token here!
>>> print api.query().organizations.get()
{u'organizations': [...]}

• List your organizations, but get a flask.Response object instead of a `dict`:

>>> from scaleway.apis import AccountAPI
>>> api = AccountAPI(auth_token='')  # Set your token here!
>>> resp = api.query(serialize=False).organizations.get()
>>> print type(resp)
<Response [200]>
>>> print resp.headers
{...}  # Response HTTP headers.
>>> print resp.links
{...}  # Parsed "Link" HTTP header, for pagination.
>>> print resp.json()
{u'organizations': [...]}

• List your servers:

>>> from scaleway.apis import ComputeAPI
>>> api = ComputeAPI(auth_token='')  # Set your token here!
>>> print api.query().servers.get()
{u'servers': [...]}
# Or choose your region, as in apis/api_compute.py
>>> api = ComputeAPI(region='ams1', auth_token='')  # Set your token here!
>>> print api.query().servers.get()
{u'servers': [...]}

• Get details of a server:

>>> from scaleway.apis import ComputeAPI
>>> api = ComputeAPI(auth_token='')  # Set your token here!
>>> server_id = ''  # Set a server ID here!
>>> print api.query().servers(server_id).get()
{u'server': {...}}

• Check if your token has the permission servers:read for the service compute for the organization 9a096d36-6bf9-470f-91df-2398aa7361f7:

>>> from scaleway.apis import AccountAPI
>>> api = AccountAPI(auth_token='')  # Set your token here!
>>> print api.has_perm(service='compute', name='servers:read',
...     resource='9a096d36-6bf9-470f-91df-2398aa7361f7')
False

Development

Assuming you are in a virtualenv:

$ pip install -e .
$ python -c 'from scaleway.apis import AccountAPI'
  # it works!

Test

To submit a patch, you'll need to test your code against python2.7 and python3.4. To run tests:

$ pip install nose coverage pycodestyle pylint
$ python setup.py nosetests --with-coverage
  (...)
$ pycodestyle scaleway
  (...)
$ pylint scaleway
  (...)

• Coverage score should never be lower than before your patch.
• PEP8 should never return an error.
• Pylint score should never be lower than before your patch.

Alternatively, to run nosetests on both Python2.7 and Python3.4, you can run tox.

Alternative libraries / clients

We maintain a list of the current library/client implementations on the api.scaleway.com repository.

License

This software is licensed under a BSD 2-Clause License.

python-scaleway's Issues

Rename golang client

I rename the golang client and update API : https://github.com/nlamirault/go-scaleway

Create server raise Client Error 400?

Using following:

compute_api = ComputeAPI(auth_token=secret_key, region='pl-waw-1')

compute_api.query().servers.post({'project': org_id, "name": 'test', "commercial_type": "DEV1-M"})

Got:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/draven/.local/lib/python2.7/site-packages/slumber/__init__.py", line 167, in post
    resp = self._request("POST", data=data, files=files, params=kwargs)
  File "scaleway/scaleway/apis/__init__.py", line 87, in _request
    return super(SlumberResource, self)._request(*args, **kwargs)
  File "/home/draven/.local/lib/python2.7/site-packages/slumber/__init__.py", line 101, in _request
    raise exception_class("Client Error %s: %s" % (resp.status_code, url), response=resp, content=resp.content)
slumber.exceptions.HttpClientError: Client Error 400: https://api-pl-waw.scaleway.com/instance/v1/zones/pl-waw-1/servers/

Since there is no documents about how to use POST / Create so based on other sample such as this I assume the above code is correct.

But Error 400 ?

Not able to specify region because of the Hyphen

Trying to list the RDB instances but I have difficulties to specify the region. That -Hyphen seems to be the problem, but I am not Python expert enough figure out how to escape is correctly.

api = API(base_url="https://api.scaleway.com", auth_token="yo")
api.make_requests_session()
instances = api.query().rdb.v1.regions.fr-par1.instances.get()

Traceback (most recent call last):
  File "executeOrder.py", line 11, in <module>
    instances = api.query().rdb.v1.regions.fr-par1.instances.get()
NameError: name 'par1' is not defined

Documentation for other request types than GET missing

All examples in the readme only demonstrate the usage for GET requests. There is not a single example for other request types like POST. I needed to dig inside the docs of Slumber to find out how to do them. Including at least one example on how to create a new server or similar would save some time for a lot of people i guess.

Add explicit error message if SNI fails

With Python 2.x, you get the error message "requests.exceptions.SSLError: hostname 'api.scaleway.com' doesn't match '*.cloud.online.net'" if the some packages, required by python-requests, aren't installed. This message is cryptic, and often misunderstood by the users who don't know what SNI is.

The packages required by python-requests to handle SNI are:

• pyasn1
• ndg-httpsclient
• pyOpenSSL

To compile them you need (on Ubuntu) to run apt-get install python-dev libffi-dev libssl-dev prior their installation.

There's a problem if you do the installation in a bad order. If you:

• run pip install scaleway-sdk
• something doesn't work, because you don't have the required packages, so you apt-get install python-dev libffi-dev libssl-dev
• then you try to re-run pip install scaleway-sdk

The packages won't be recompiled (and I don't really know why, I need to investigate).

If you're in this case, you need to pip uninstall -v pyasn1 ndg-httpsclient pyOpenSSL scaleway-sdk after installing the dependencies, and before running pip install scaleway-sdk.

Anyway, we need to:

• add an explicit message if requests.exceptions.SSLError is raised
• see why the Python packages required to handle SNI aren't reinstalled when you're trying to install sdk, install the dependencies and reinstall the sdk.

[SECURITY] Secret Manager - secret content leaks in debug logs

I have a staging server with logging config setup to logging.DEBUG

If I just use the default logging.basicConfig (ie without turning off scaleway logs), my secrets will leak in the logs as b64-encoded bytes:

PyPi doesn't render package's description

Package's long description is not rendered on PyPi:

Any idea why ?

A CLI in Go

https://github.com/nlamirault/go-onlinelabs

Python 3 unittests fails.

Python 3 unittests are failing because of HTTPretty, which is not committed to support Python 3.

Should we replace HTTPretty by another package ?

Missing license

Looks like we forgot to set the licence and add copyrights header in our files.

add user_data parameter in create server API

I think it would be great to have ability to send user_data(usually text that is interpreted as bash script) when creating a server via the API, https://github.com/scaleway/api.scaleway.com/blob/master/contents/server.md#create-a-new-server-post

This way, a scaleway customer can be able to specify things that they would like to run during the server creation.

This already exists in other cloud platforms, eg:

The ovh create server api has a parameter called userData that you can specify:: https://api.ovh.com/console/#/cloud/project/%7BserviceName%7D/instance#POST
The digitalocean API also takes a parameter called user_data, https://developers.digitalocean.com/documentation/v2/#create-a-new-droplet
This is very useful especially when you want to automate creation of servers that are different from each other but are based off the same image.

Exception text not useful

When a request returns an error the exception printed to the console is quite useless because it does not contain the message of the response. Just the error code. It would be more convenient if the message of result would be included in the Exception text.

404 Error on retrieving organizations

The first example of the readme api.query().organizations.get() fails with error 404:
{"message": "The requested URL was not found on the server. If you entered the URL manually please check your spelling and try again.", "type": "404"}

Is this sdk still maintained

I was wondering if this sdk is still maintained by the scaleway team.
It still has support for the non-working account.scaleway.com part of the api and doesn't support the new endpoints (like https://api.scaleway.com/account/v2alpha1/ssh-keys for managing ssh keys) that are supported in the scaleway go sdk.
Not entirely related to this specific sdk, but this part of the api doesn't seem to be documented anywhere.

Unclear case with api_account.perm_matches and None request

Reading the following line:

        self.assertTrue(self.api.perm_matches(None, 'object:*'))

I would assume the opposite,
can you clarify this case and maybe update the docstring?