Home

Awesome

Gimel

Build Status PyPI

a Scaleable A/B testing backend in ~100 lines of code (and for free*)

What is it?

an A/B testing backend using AWS Lambda/API Gateway + Redis

Key Features:

Looking for contributors

click here for more info

What does Gimel mean?

Gimel (גִּימֵל) is the 3rd letter of the Hebrew Alphabet. The letter (ג) also looks visually similar to the greek Lambda (λ).

Installation / Quick Start

You will need a live instance of redis accessible online from AWS. Then run:

$ pip install gimel
$ gimel configure
$ gimel deploy

It will automatically configure your AWS Lambda functions, API gateway and produce a JS snippet ready to use for tracking your experiments.

Architecture

Client

I suggest looking at Alephbet to get more details, but at a high level, the client runs on the end-user browser. It will randomly pick a variant and execute a javascript function to 'activate' it. When a goal is reached -- user performs a certain action, this also include the pseudo-goal of participating in the experiment -- then an event is sent to the backend. An event typically looks something like "experiment ABC, variant red, user participated", or "experiment XYZ, variant blue, check out goal reached".

Alephbet might send duplicate events, but each event should include a uuid to allow the backend to de-duplicate it. More below

Data Store - Redis HyperLogLog

The data store keeps a tally of each event that comes into the system. Being able to count unique events (de-duplication) was important to keep an accurate count. One approach would be to store each event in an entry / database row / document, and then run some kind of a unique count on it. Or we could use a nifty algorithm called HyperLogLog. HyperLogLog allows you to count unique counts without storing each and every item.

In terms of storage space, redis HyperLogLog offers a fixed size of 12k per counter. This gives us ample space for storing experiment data with low memory footprint.

NOTE: there's no free lunch. HLL isn't as accurate, especially with large experiments. See here or check out lamed if you're looking for a more accurate, but more memory-hungry option.

Backend - AWS Lambda / API Gateway

The backend had to take care of a few simple types of requests:

Dashboard

New! access your dashboard with gimel dashboard

How does tracking work?

Check out Alephbet.

Command Reference

Advanced

custom API endpoints

If you want to use different API endpoints, you can add your own extra_wiring into the config.json file (e.g. using gimel configure).

for example, this will add a .../prod/my_tracking_endpoint URL pointing to the gimel-track lambda:

{
    "redis": {
       ...
    },
    "extra_wiring": [
        {
            "lambda": {
                "FunctionName": "gimel-track",
                "Handler": "gimel.track",
                "MemorySize": 128,
                "Timeout": 3
            },
            "api_gateway": {
                "pathPart": "my_tracking_endpoint",
                "method": {
                    "httpMethod": "GET",
                    "apiKeyRequired": false,
                    "requestParameters": {
                        "method.request.querystring.namespace": false,
                        "method.request.querystring.experiment": false,
                        "method.request.querystring.variant": false,
                        "method.request.querystring.event": false,
                        "method.request.querystring.uuid": false
                    }
                }
            }
        }
    ]
}

see WIRING

Privacy, Ad-blockers (GDPR etc)

Gimel provides a backend for A/B test experiment data. This data is aggregated and does not contain any personal information at all. It merely stores the total number of actions with a certain variation against another.

As such, Gimel should meet privacy requirements of GDPR and similar privacy regulations.

Nevertheless, important disclaimers:

License

Gimel is distributed under the MIT license. All 3rd party libraries and components are distributed under their respective license terms.

Copyright (C) 2016 Yoav Aner

Permission is hereby granted, free of charge, to any person obtaining a copy of this software and associated
documentation files (the "Software"), to deal in the Software without restriction, including without limitation the
rights to use, copy, modify, merge, publish, distribute, sublicense, and/or sell copies of the Software, and to permit
persons to whom the Software is furnished to do so, subject to the following conditions:

The above copyright notice and this permission notice shall be included in all copies or substantial portions of the
Software.

THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE
WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR
COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.