> An adapter based tool to handle feature toggles, progressive rollouts and A/B Testing services in our products
> An adapter based tool to handle feature toggles, progressive rollouts and A/B Testing services in our products
``sh`
$ npm install @s-ui/pde
This is a tool that abstracts from:
- PDE tools like Optimizely, Houston, Target, Unleash, ...
- Integration with third party analytics tools like Segment
- User id management
- User consents management
- Common audiences and segments
In order to abstract from progressive delivery and experimentation (PDE) tools like Optimizely, sui-pde provides a JS class which requires an adapter that connects directly to one of those tools. We did implement the Optimizely Full Stack adapter.
Setup the tool's own react context in your context factory like this:
`jsx
import {PDE, PdeContext} from '@s-ui/pde'
import OptimizelyAdapter from '@s-ui/pde/lib/adapters/optimizely'
// all options here https://docs.developers.optimizely.com/full-stack/docs/initialize-sdk-javascript-node, but for now only 3 of them are available
const optimizelyInstance = OptimizelyAdapter.createOptimizelyInstance({
sdkKey: MY_API_KEY, // optimizely sdk api key
options, // optional, datafileOptions
datafile // optional
})
const optimizelyAdapter = new OptimizelyAdapter({
optimizely: optimizelyInstance,
userId: // mandatory,
hasUserConsents, // if false, the user won't be part of the A/B test,
applicationAttributes: { // optional, global application attributes that must be send on every experiment activation
site: 'coches.net',
environment: 'development'
}
})
const pde = new PDE({
adapter: optimizelyAdapter,
hasUserConsents: true // Kept because of legacy reasons, pass it by the OptimizelyAdapter constructor
})
// app.js
// children
`
#### SSR considerations
When client-side rendering, sui-pde will load the datafile saved as window.__INITIAL_CONTEXT_VALUE__.pde as initial datafile. Therefore, you'll need to inject the output of the pde.getInitialContextData() function in your html when server side rendering.
Feature tests are similar to A/B/n tests that allow you to control whether for each variation the associated feature is on or off via feature flags (aka feature toggles). It also allows you to control the feature variable values for the various variables associated with the feature.
The useDecision hook retrieves the result for a given decision, based on logic from the decision-making tool you are using (e.g., Optimizely).
`js
import {useDecision} from '@s-ui/pde'
function Component() {
const decision = useDecision('feature_test', {})
return (
<>
{decision.enabled &&
My feature is enabled
}My feature is disabled
}Current Variation
}Better Variation
}You can also use the
component:`js
import {Decision} from '@s-ui/pde'const MyComponent = () => {
return (
{({enabled}) => enabled ? My feature is enabled
: My feature is disabled
}
)
}
`And the function to call elsewhere, like getInitialProps:
js
import {getDecision} from '@s-ui/pde'Page.getInitialProps = async ({context}) => {
const {decide} = getDecision(pde)
const {enabled} = decide('ff_web_my_flag')
}
`#### Attributes
You can pass additional attributes to refine your decision logic:
js
import {useDecision} from '@s-ui/pde'const MyComponent = () => {
const {enabled, variationKey} = useDecision('feature_test', {
attributes: {
isLoggedIn: true
}
})
}
`#### Forcing a decision
You can force specific decision outcomes during testing by adding a query parameter.
-
will enable the example feature test
- http://www.fotocasa.es/es?suipde_example=off will disable the example feature test
- http://www.fotocasa.es/es?suipde_example=variation_a will enable the example feature test and will force the variation variation_a`