Placebo allows you to mock boto3 calls that look just like normal calls but actually have no effect at all. It does this by allowing you to record a set of calls and save them to a data file and then replay those calls later (e.g. in a unit test) without ever hitting the AWS endpoints.

$ pip install placebo

Placebo uses the event mechanism in botocore to do most of its work. To start with, you need a boto3 Session object.

import boto3
import placebo

session = boto3.Session()

Once you have a Session object, you can tell placebo about the Session like this:

pill = placebo.attach(session, data_path='/path/to/response/directory')

The data_path is a path to a directory where you want responses to be stored or that contains previously recorded responses you want to playback.

The attach function returns an instance of a Pill object. This object will be used to control all recording and playback of requests for all clients created by this session object.

The first thing you will probably want to do is record some requests. To do this, simply:

pill.record()

By default, the record method will cause all responses from all services to be recorded to the data_path. If you are only interested in responses from one certain services, you can limit the recording by passing in a list of service names.

pill.record(services='ec2,iam')

This would limit to recording to responses from the ec2 service and the iam service. If you want to restrict recording to only certain operations in a single service, you can do this:

pill.record(services='ec2', operations='DescribeInstances,DescribeKeyPairs')

From this point on, any clients that match the recording specification and are created from the session will be placebo-aware. To record responses, just create the client and use it as you normally would.

lambda = session.client('lambda')
lambda.list_functions()
... more lambda calls ...

Each response will be saved as an individual JSON data file in the data_path path you specified when you attached the session. Multiple responses from the same service and operation are stored as separate files and will be replayed in the same order on playback.

Later, to replay saved requests:

import boto3
import placebo

session = boto3.Session()
pill = placebo.attach(session, data_path='/path/to/response/directory')
pill.playback()
lambda = session.client('lambda')
lambda.list_functions()
... mocked response will be returned

Sometimes, Placebo needs to be attached to the Boto3 default session object.

To attach Placebo to the default session, it is necessary to explicitly set up the default session by making a call to boto3.setup_default_session(). The default session is then accessible at boto3.DEFAULT_SESSION.

For example:

import boto3
import placebo

# Explicity set up the default session and attach Placebo to it.
boto3.setup_default_session()
session = boto3.DEFAULT_SESSION
pill = placebo.attach(session, data_path='/path/to/response/directory')
pill.record()

# Now make Boto3 calls using the default session.
client = boto3.client('ec2')
client.describe_images(DryRun=False)

This is particularly useful if you are writing tests for legacy code that makes use of the Boto3 default session.

You can also add mocked responses manually:

list_functions_response = [
    {
        "Version": "$LATEST",
        "CodeSha256": "I8Scq2g6ZKcPIvhKzvZqCiV4pDysxq4gZ+jLcMmDy5Y=",
        "FunctionName": "foobar",
        "MemorySize": 128,
        "CodeSize": 876521,
        "FunctionArn": "arn:aws:lambda:us-west-2:123456789012:function:foobar",
        "Handler": "foobar.handler",
        "Role": "arn:aws:iam::123456789012:role/foobar-role",
        "Timeout": 30,
        "LastModified": "2015-11-06T22:30:32.164+0000",
        "Runtime": "python2.7",
        "Description": "Foos all of the bars"
    }]

pill.save_response(service='lambda', operation='ListFunctions',
                   response_data=list_functions_response, http_response=200)

You can add additional responses to a particular operation and the responses will be returned in order. The final parameter is the HTTP response code which is optional. The default value is 200.

Placebo also provides a decorator for easier usage.

First, you'll want to decorate your test method with placebo_session and include the session kwarg in your method, ex:

@placebo_session
def test_your_function(self, session):
    foo = Foo()
    arn = foo.create_iam_roles(session)
    self.assertEqual(arn, "arn:aws:iam::123:role/{}".format(foo.role_name))

Now, you'll be able to record the AWS interactions with an environment variable:

$ PLACEBO_MODE=record nosetests tests.tests:TestFoo.test_create_iam_roles

You can optionally pass an AWS profile to use:

$ PLACEBO_PROFILE=foo PLACEBO_MODE=record nosetests tests.tests:TestFoo.test_create_iam_roles

In this example, it has created the following JSON blobs:

tests/placebo/TestFoo.test_create_iam_roles
tests/placebo/TestFoo.test_create_iam_roles/iam.CreateRole_1.json
tests/placebo/TestFoo.test_create_iam_roles/iam.GetRole_1.json
tests/placebo/TestFoo.test_create_iam_roles/iam.GetRolePolicy_1.json
tests/placebo/TestFoo.test_create_iam_roles/iam.PutRolePolicy_1.json

After the JSON has been created, simply drop the environment variables and re-run your test:

$ nosetests tests.tests:TestFoo.test_create_iam_roles