Skip to content

G
gitlab-ce
Commit a065a997 authored by Jeremy Jackson's avatar Jeremy Jackson Committed by jejacks0n
Browse files
Options

Update gitlab-experiment to version 0.7.0

parent ab695c19
Hide whitespace changes
Inline Side-by-side
Showing with 112 additions and 154 deletions
Gemfile
View file @ a065a997
......@@ -489,7 +489,7 @@ gem 'flipper', '~> 0.21.0'
gem 'flipper-active_record', '~> 0.21.0'
gem 'flipper-active_support_cache_store', '~> 0.21.0'
gem 'unleash', '~> 3.2.2'
gem 'gitlab-experiment', '~> 0.6.5'
gem 'gitlab-experiment', git: 'https://gitlab.com/gitlab-org/ruby/gems/gitlab-experiment', branch: 'release-0.7.0'
# Structured logging
gem 'lograge', '~> 0.5'
......
Gemfile.lock
View file @ a065a997
GIT
remote: https://gitlab.com/gitlab-org/ruby/gems/gitlab-experiment.git
revision: 176d767486217ae5080785247f29dc4549d85a4a
branch: release-0.7.0
specs:
gitlab-experiment (0.7.0)
activesupport (>= 3.0)
request_store (>= 1.0)
PATH
remote: vendor/gems/mail-smtp_pool
specs:
......@@ -460,10 +469,6 @@ GEM
gitlab-dangerfiles (2.8.0)
danger (>= 8.3.1)
danger-gitlab (>= 8.0.0)
gitlab-experiment (0.6.5)
activesupport (>= 3.0)
request_store (>= 1.0)
scientist (~> 1.6, >= 1.6.0)
gitlab-fog-azure-rm (1.2.0)
azure-storage-blob (~> 2.0)
azure-storage-common (~> 2.0)
......@@ -1157,7 +1162,6 @@ GEM
sawyer (0.8.2)
addressable (>= 2.3.5)
faraday (> 0.8, < 2.0)
scientist (1.6.2)
sd_notify (0.1.0)
securecompare (1.0.0)
seed-fu (2.3.7)
......@@ -1470,7 +1474,7 @@ DEPENDENCIES
github-markup (~> 1.7.0)
gitlab-chronic (~> 0.10.5)
gitlab-dangerfiles (~> 2.8.0)
gitlab-experiment (~> 0.6.5)
gitlab-experiment!
gitlab-fog-azure-rm (~> 1.2.0)
gitlab-labkit (~> 0.21.3)
gitlab-license (~> 2.1.0)
......
app/experiments/application_experiment.rb
View file @ a065a997
......@@ -41,10 +41,6 @@ class ApplicationExperiment < Gitlab::Experiment # rubocop:disable Gitlab/Namesp
# define a default nil control behavior so we can omit it when not needed
end
def track(action, **event_args)
super(action, **tracking_context.merge(event_args))
end
# TODO: remove
# This is deprecated logic as of v0.6.0 and should eventually be removed, but
# needs to stay intact for actively running experiments. The new strategy
......@@ -64,12 +60,12 @@ class ApplicationExperiment < Gitlab::Experiment # rubocop:disable Gitlab/Namesp
private
def tracking_context
def tracking_context(event_args)
{
namespace: context.try(:namespace) || context.try(:group),
project: context.try(:project),
user: user_or_actor
}.compact || {}
}.merge(event_args)
end
def user_or_actor
......
config/initializers/gitlab_experiment.rb
View file @ a065a997
......@@ -10,6 +10,13 @@ Gitlab::Experiment.configure do |config|
#
config.base_class = 'ApplicationExperiment'
# Customize the logic of our default rollout, which shouldn't include
# assigning the control yet -- we specifically set it to false for now.
#
config.default_rollout = Gitlab::Experiment::Rollout::Percent.new(
include_control: false
)
# Mount the engine and middleware at a gitlab friendly style path.
#
# The middleware currently focuses only on handling redirection logic, which
......
doc/development/experiment_guide/gitlab_experiment.md
View file @ a065a997
......@@ -4,26 +4,24 @@ group: Adoption
info: To determine the technical writer assigned to the Stage/Group associated with this page, see https://about.gitlab.com/handbook/engineering/ux/technical-writing/#assignments
---
# Implementing an A/B/n experiment using GLEX
# Implementing an A/B/n experiment
## Introduction
`Gitlab::Experiment` (GLEX) is tightly coupled with the concepts provided by
[Feature flags in development of GitLab](../feature_flags/index.md). Here, we refer
to this layer as feature flags, and may also use the term Flipper, because we
built our development and experiment feature flags atop it.
You're strongly encouraged to read and understand the
[Feature flags in development of GitLab](../feature_flags/index.md) portion of the
documentation before considering running experiments. Experiments add additional
concepts which may seem confusing or advanced without understanding the underpinnings
of how GitLab uses feature flags in development. One concept: GLEX supports
experiments with multiple variants, which are sometimes referred to as A/B/n tests.
The [`gitlab-experiment` project](https://gitlab.com/gitlab-org/ruby/gems/gitlab-experiment)
exists in a separate repository, so it can be shared across any GitLab property that uses
Ruby. You should feel comfortable reading the documentation on that project as well
if you want to dig into more advanced topics.
Experiments in GitLab are tightly coupled with the concepts provided by
[Feature flags in development of GitLab](../feature_flags/index.md). You're strongly encouraged
to read and understand the [Feature flags in development of GitLab](../feature_flags/index.md)
portion of the documentation before considering running experiments. Experiments add additional
concepts which may seem confusing or advanced without understanding the underpinnings of how GitLab
uses feature flags in development. One concept: experiments can be run with multiple variants,
which are sometimes referred to as A/B/n tests.
We use the [`gitlab-experiment` gem](https://gitlab.com/gitlab-org/ruby/gems/gitlab-experiment),
sometimes referred to as GLEX, to run our experiments. The gem exists in a separate repository
so it can be shared across any GitLab property that uses Ruby. You should feel comfortable reading
the documentation on that project if you want to dig into more advanced topics or open issues. Be
aware that the documentation there reflects what's in the main branch and may not be the same as
the version being used within GitLab.
## Glossary of terms
......@@ -35,41 +33,9 @@ when communicating about experiments:
- `control`: The default, or "original" code path.
- `candidate`: Defines an experiment with only one code path.
- `variant(s)`: Defines an experiment with multiple code paths.
- `behaviors`: Used to reference all possible code paths of an experiment, including the control.
### How it works
Use this decision tree diagram to understand how GLEX works. When an experiment runs,
the following logic is executed to determine what variant should be provided,
given how the experiment has been defined and using the provided context:
```mermaid
graph TD
GP[General Pool/Population] --> Running?
Running? -->|Yes| Cached?[Cached? / Pre-segmented?]
Running? -->|No| Excluded[Control / No Tracking]
Cached? -->|No| Excluded?
Cached? -->|Yes| Cached[Cached Value]
Excluded? -->|Yes| Excluded
Excluded? -->|No| Segmented?
Segmented? -->|Yes / Cached| VariantA
Segmented? -->|No| Included?[Experiment Group?]
Included? -->|Yes| Rollout
Included? -->|No| Control
Rollout -->|Cached| VariantA
Rollout -->|Cached| VariantB
Rollout -->|Cached| VariantC
classDef included fill:#380d75,color:#ffffff,stroke:none
classDef excluded fill:#fca121,stroke:none
classDef cached fill:#2e2e2e,color:#ffffff,stroke:none
classDef default fill:#fff,stroke:#6e49cb
class VariantA,VariantB,VariantC included
class Control,Excluded excluded
class Cached cached
```
## Implement an experiment
## Implementing an experiment
[Examples](https://gitlab.com/gitlab-org/growth/growth/-/wikis/GLEX-Framework-code-examples)
......@@ -87,9 +53,9 @@ experiment in code. An experiment implementation can be as simple as:
```ruby
experiment(:pill_color, actor: current_user) do |e|
e.use { 'control' }
e.try(:red) { 'red' }
e.try(:blue) { 'blue' }
e.control { 'control' }
e.variant(:red) { 'red' }
e.variant(:blue) { 'blue' }
end
```
......@@ -146,11 +112,11 @@ We can also implement this experiment in a HAML file with HTML wrappings:
```haml
#cta-interface
- experiment(:pill_color, actor: current_user) do |e|
- e.use do
- e.control do
.pill-button control
- e.try(:red) do
- e.variant(:red) do
.pill-button.red red
- e.try(:blue) do
- e.variant(:blue) do
.pill-button.blue blue
```
......@@ -212,38 +178,30 @@ wouldn't be resolvable.
### Advanced experimentation
GLEX allows for two general implementation styles:
There are two ways to implement an experiment:
1. The simple experiment style described previously.
1. A more advanced style where an experiment class can be provided.
1. A more advanced style where an experiment class is provided.
The advanced style is handled by naming convention, and works similar to what you
would expect in Rails.
To generate a custom experiment class that can override the defaults in
`ApplicationExperiment` (our base GLEX implementation), use the rails generator:
`ApplicationExperiment` use the Rails generator:
```shell
rails generate gitlab:experiment pill_color control red blue
```
This generates an experiment class in `app/experiments/pill_color_experiment.rb`
with the variants (or _behaviors_) we've provided to the generator. Here's an example
of how that class would look after migrating the previous example into it:
with the _behaviors_ we've provided to the generator. Here's an example
of how that class would look after migrating our previous example into it:
```ruby
class PillColorExperiment < ApplicationExperiment
def control_behavior
'control'
end
def red_behavior
'red'
end
def blue_behavior
'blue'
end
control { 'control' }
variant(:red) { 'red' }
variant(:blue) { 'blue' }
end
```
......@@ -254,13 +212,13 @@ providing the block we were initially providing, by explicitly calling `run`:
experiment(:pill_color, actor: current_user).run
```
The _behavior_ methods we defined in our experiment class represent the default
implementation. You can still use the block syntax to override these _behavior_
methods however, so the following would also be valid:
The _behaviors_ we defined in our experiment class represent the default
implementation. You can still use the block syntax to override these _behaviors_
however, so the following would also be valid:
```ruby
experiment(:pill_color, actor: current_user) do |e|
e.use { '<strong>control</strong>' }
e.control { '<strong>control</strong>' }
end
```
......@@ -279,11 +237,11 @@ variant, and any account older than 2 weeks old would be assigned the _blue_ var
```ruby
class PillColorExperiment < ApplicationExperiment
# ...registered behaviors
segment(variant: :red) { context.actor.first_name == 'Richard' }
segment :old_account?, variant: :blue
# ...behaviors
private
def old_account?
......@@ -315,9 +273,9 @@ be nothing - but no events are tracked in these cases as well.
```ruby
class PillColorExperiment < ApplicationExperiment
exclude :old_account?, ->{ context.actor.first_name == 'Richard' }
# ...registered behaviors
# ...behaviors
exclude :old_account?, ->{ context.actor.first_name == 'Richard' }
private
......@@ -327,23 +285,11 @@ class PillColorExperiment < ApplicationExperiment
end
```
We can also do exclusion when we run the experiment. For instance,
if we wanted to prevent the inclusion of non-administrators in an experiment, consider
the following experiment. This type of logic enables us to do complex experiments
while preventing us from passing things into our experiments, because
we want to minimize passing things into our experiments:
```ruby
experiment(:pill_color, actor: current_user) do |e|
e.exclude! unless can?(current_user, :admin_project, project)
end
```
You may also need to check exclusion in custom tracking logic by calling `should_track?`:
```ruby
class PillColorExperiment < ApplicationExperiment
# ...behaviors
# ...registered behaviors
def expensive_tracking_logic
return unless should_track?
......@@ -353,16 +299,11 @@ class PillColorExperiment < ApplicationExperiment
end
```
Exclusion rules aren't the best way to determine if an experiment is active. Override
the `enabled?` method for a high-level way of determining if an experiment should
run and track. Make the `enabled?` check as efficient as possible because it's the
first early opt-out path an experiment can implement.
### Tracking events
One of the most important aspects of experiments is gathering data and reporting on
it. GLEX provides an interface that allows tracking events across an experiment.
You can implement it consistently if you provide the same context between
it. You can use the `track` method to track events across an experimental implementation.
You can track events consistently to an experiment if you provide the same context between
calls to your experiment. If you do not yet understand context, you should read
about contexts now.
......@@ -373,10 +314,10 @@ the arguments you would normally use when
of tracking an event in Ruby would be:
```ruby
experiment(:pill_color, actor: current_user).track(:created)
experiment(:pill_color, actor: current_user).track(:clicked)
```
When you run an experiment with any of these examples, an `:assigned` event
When you run an experiment with any of the examples so far, an `:assigned` event
is tracked automatically by default. All events that are tracked from an
experiment have a special
[experiment context](https://gitlab.com/gitlab-org/iglu/-/blob/master/public/schemas/com.gitlab/gitlab_experiment/jsonschema/1-0-0)
......@@ -390,28 +331,20 @@ event would be tracked at that time for them.
NOTE:
GitLab tries to be sensitive and respectful of our customers regarding tracking,
so GLEX allows us to implement an experiment without ever tracking identifying
so our experimentation library allows us to implement an experiment without ever tracking identifying
IDs. It's not always possible, though, based on experiment reporting requirements.
You may be asked from time to time to track a specific record ID in experiments.
The approach is largely up to the PM and engineer creating the implementation.
No recommendations are provided here at this time.
## Test with RSpec
This gem provides some RSpec helpers and custom matchers. These are in flux as of GitLab 13.10.
First, require the RSpec support file to mix in some of the basics:
## Testing with RSpec
```ruby
require 'gitlab/experiment/rspec'
```
You still need to include matchers and other aspects, which happens
automatically for files in `spec/experiments`, but for other files and specs
you want to include it in, you can specify the `:experiment` type:
In the course of working with experiments, you'll probably want to utilize the RSpec
tooling that's built in. This happens automatically for files in `spec/experiments`, but
for other files and specs you want to include it in, you can specify the `:experiment` type:
```ruby
it "tests", :experiment do
it "tests experiments nicely", :experiment do
end
```
......@@ -421,28 +354,32 @@ You can stub experiments using `stub_experiments`. Pass it a hash using experime
names as the keys, and the variants you want each to resolve to, as the values:
```ruby
# Ensures the experiments named `:example` & `:example2` are both
# "enabled" and that each will resolve to the given variant
# (`:my_variant` & `:control` respectively).
# Ensures the experiments named `:example` & `:example2` are both "enabled" and
# that each will resolve to the given variant (`:my_variant` and `:control`
# respectively).
stub_experiments(example: :my_variant, example2: :control)
experiment(:example) do |e|
e.enabled? # => true
e.variant.name # => 'my_variant'
e.assigned.name # => 'my_variant'
end
experiment(:example2) do |e|
e.enabled? # => true
e.variant.name # => 'control'
e.assigned.name # => 'control'
end
```
### Exclusion and segmentation matchers
### Exclusion, segmentation, and behavior matchers
You can also test the exclusion and segmentation matchers.
You can also test things like the registered behaviors, the exclusions, and
segmentations using the matchers.
```ruby
class ExampleExperiment < ApplicationExperiment
control { }
candidate { '_candidate_' }
exclude { context.actor.first_name == 'Richard' }
segment(variant: :candidate) { context.actor.username == 'jejacks0n' }
end
......@@ -450,6 +387,10 @@ end
excluded = double(username: 'rdiggitty', first_name: 'Richard')
segmented = double(username: 'jejacks0n', first_name: 'Jeremy')
# register_behavior matcher
expect(experiment(:example)).to register_behavior(:control)
expect(experiment(:example)).to register_behavior(:candidate).with('_candidate_')
# exclude matcher
expect(experiment(:example)).to exclude(actor: excluded)
expect(experiment(:example)).not_to exclude(actor: segmented)
......@@ -495,18 +436,11 @@ expect(experiment(:example)).to track(:my_event, value: 1, property: '_property_
experiment(:example, :variant_name, foo: :bar).track(:my_event, value: 1, property: '_property_')
```
### Recording and assignment tracking
To test assignment tracking and the `record!` method, you can use or adopt the following
shared example: [tracks assignment and records the subject](https://gitlab.com/gitlab-org/gitlab/blob/master/spec/support/shared_examples/lib/gitlab/experimentation_shared_examples.rb).
## Experiments in the client layer
This is in flux as of GitLab 13.10, and can't be documented just yet.
Any experiment that's been run in the request lifecycle surfaces in and `window.gl.experiments`,
and matches [this schema](https://gitlab.com/gitlab-org/iglu/-/blob/master/public/schemas/com.gitlab/gitlab_experiment/jsonschema/1-0-0)
so you can use it when resolving some concepts around experimentation in the client layer.
and matches [this schema](https://gitlab.com/gitlab-org/iglu/-/blob/master/public/schemas/com.gitlab/gitlab_experiment/jsonschema/1-0-3)
so it can be used when resolving experimentation in the client layer.
### Use experiments in Vue
......@@ -628,7 +562,7 @@ is viewed as being either `on` or `off`, this isn't accurate for experiments.
Generally, `off` means that when we ask if a feature flag is enabled, it will always
return `false`, and `on` means that it will always return `true`. An interim state,
considered `conditional`, also exists. GLEX takes advantage of this trinary state of
considered `conditional`, also exists. We take advantage of this trinary state of
feature flags. To understand this `conditional` aspect: consider that either of these
settings puts a feature flag into this state:
......@@ -638,7 +572,7 @@ settings puts a feature flag into this state:
Conditional means that it returns `true` in some situations, but not all situations.
When a feature flag is disabled (meaning the state is `off`), the experiment is
considered _inactive_. You can visualize this in the [decision tree diagram](#how-it-works)
considered _inactive_. You can visualize this in the [decision tree diagram](https://gitlab.com/gitlab-org/ruby/gems/gitlab-experiment#how-it-works)
as reaching the first `Running?` node, and traversing the negative path.
When a feature flag is rolled out to a `percentage_of_actors` or similar (meaning the
......
spec/experiments/application_experiment_spec.rb
View file @ a065a997
......@@ -117,7 +117,7 @@ RSpec.describe ApplicationExperiment, :experiment do
describe '#publish_to_database' do
using RSpec::Parameterized::TableSyntax
let(:publish_to_database) { application_experiment.publish_to_database }
let(:publish_to_database) { ActiveSupport::Deprecation.silence { application_experiment.publish_to_database } }
shared_examples 'does not record to the database' do
it 'does not create an experiment record' do
......@@ -358,11 +358,11 @@ RSpec.describe ApplicationExperiment, :experiment do
stub_feature_flags(namespaced_stub: true)
end
it "returns the first variant name" do
it "returns an assigned name" do
application_experiment.try(:variant1) {}
application_experiment.try(:variant2) {}
expect(application_experiment.variant.name).to eq('variant1')
expect(application_experiment.variant.name).to eq('variant2')
end
end
end
......
spec/support/gitlab_experiment.rb
View file @ a065a997
......@@ -10,9 +10,24 @@ RSpec.configure do |config|
# Disable all caching for experiments in tests.
config.before do
allow(Gitlab::Experiment::Configuration).to receive(:cache).and_return(nil)
# Disable all deprecation warnings in the test environment, which can be
# resolved one by one and tracked in:
#
# https://gitlab.com/gitlab-org/gitlab/-/issues/350944
allow(Gitlab::Experiment::Configuration).to receive(:deprecator).and_wrap_original do |method, version|
method.call(version).tap do |deprecator|
deprecator.silenced = true
end
end
end
config.before(:each, :experiment) do
stub_snowplow
end
end
# Once you've resolved a given deprecation, you can disallow it here, which
# will raise an exception if it's used anywhere.
ActiveSupport::Deprecation.disallowed_warnings << "`experiment_group?` is deprecated"
spec/support/shared_examples/lib/gitlab/experimentation_shared_examples.rb
View file @ a065a997
# frozen_string_literal: true
RSpec.shared_examples 'tracks assignment and records the subject' do |experiment, subject_type|
before do
stub_experiments(experiment => true)
end
it 'tracks the assignment', :experiment do
expect(experiment(experiment))
.to track(:assignment)
......@@ -11,9 +15,7 @@ RSpec.shared_examples 'tracks assignment and records the subject' do |experiment
end
it 'records the subject' do
stub_experiments(experiment => :candidate)
expect(Experiment).to receive(:add_subject).with(experiment.to_s, variant: :experimental, subject: subject)
expect(Experiment).to receive(:add_subject).with(experiment.to_s, variant: anything, subject: subject)
action
end
......
Markdown is supported
0% or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment
GitLab Nexedi Edition | About GitLab | About Nexedi | 沪ICP备2021021310号-2 | 沪ICP备2021021310号-7