Lambda Python Bucketer

Using Optimizely Python SDK v1.4

The Python bucketer is a lambda function exposed via an https AWS API Gateway that takes an array of parameters and returns the variation_key that the visitor was bucketed into.

input

https://aws-api-gateway-address.amazonaws.com/prod/bucketer?  
  e=checkout_flow_experiment
  &u=cooper34523569
  &a=tier:"platinum",member_since_days:56
  &debug=1

response

{
  "variation": "one_step_checkout",
  "user": "cooper34523569",
  "experiment": "checkout_flow_experiment",
  "debug": {
    "project_id": "10000000000",
    "account_id": "80000000000",
    "time": "2018-04-09 17:06:02.698754",
    "datafile": {
      "url": "https://cdn.optimizely.com/PATH/TO/DATAFILE.json",
      "revision": "22"
    },
    "user_attributes": {
      "tier": "platinum",
      "member_since_days": 56
    }
  }
}

Bucketing Parameters

parameter	name	details	required	default
e	experiment_key	Pass the experiment_key in as a string	yes	null
u	user_id	Pass the userId in as a string. If no userId is provided, the userId will default to the visitors IP address	no	IP address
a	attributes	A list of user attributes used for activate. These will be passed into the activate call, and can be used to qualify users for Audience and for Results segmentation.	no	{}
debug	debug mode	Enabled debug mode. The response will show additional data, including accountId, projectId, and verbose datafile characteristics	no	False

Datafile Management

The Optimizely Full SDKs require the consumption of the project's JSON datafile in order to be informed of the experiment settings configured within Optimizely.

Current state

Running the webhook.py script will fetch the datafile contents and store them within datafile.py. The core Lambda function file, app.py imports datafile.py and uses its contents to instantiate the Optimizely Client. After you've updated the datafile contents using the webhook, you'll need to manually push the code package up to Lambda.

In order to inform webhook.py where the datafile lives, you must set your datafile url as an environment variable:

$ export DATAFILE_URL="https://cdn.optimizely.com/PATH/TO/FULLSTACK/DATAFILE.json"

Future state

Future versions of this tool will do one of the following

• Include an additional Lambda function that will serve as a webhook that will fetch a fresh copy of the datafile, and update the code package of the Bucketer lambda function.
• Include an additional Lambda function that will serve as a webhook that will fetch a fresh copy of the datafile and push it into an S3 bucket. The Bucketer lambda function would then have to access the datafile contents via a Boto S3 connection, which may not be performant.

Updating your Lambda function code

Flatten everything into a lambda-friendly zip

./build.sh

Push the code to your lambda function

aws lambda update-function-code --function-name BUCKETERFNC --zip-file fileb://package.zip --profile AWSPROFILENAME