For multi-tenant / multi-project support use: Multi-Tenant Branch
RPC decision service that is implemented using json over http. The API replicates the functionality of the SDK in a centralized location. It also implementes user profiles and datafile management. With that in mind, the application exposes two endpoints. The "rpc" endpoint accepts the different methods as JSON that replicate the SDK methods and the "describe" endpoint which returns the descriptions of the different methods and data types. This is not a Rest based API but a "json over http" API. You will only need to communicate with a single endpoint.
This current version does not support multiple projects. The project settings currently live in the configuration file. I am currently working on the next release that will support changing the project datafile through the RPC method call. It will support multiple Optimizely clients each capable of handling a different SDK project datafile.
The decision service currently supports:
- Activate
- Get Variation
- Set Forced Variation
- Get Forced Variation
- Is Feature Enabled (Feature Tests and Feature Flags / Rollouts)
- Get Enabled Features
- Track Conversion Events
The following instructions will get the project up and running on your local machine.
- Download or clone a copy of the project.
- Install the required packages via "npm" from the package.json.
- Install Redis locally.
- By default the decision service is configured to connect to '127.0.0.1:6379'.
- This setting can be configured in the 'config.js' file located in the 'config directory'.
- Datafile settings, like CDN url and SDK key, are also configured in the 'config.js' file located in the 'config directory'.
- Start the local Redis server.
- Run 'server.js' in the root directory.
11/18/2018
Added the following methods:
- Get Variation
- Set Forced Variation
- Get Forced Variation
- Get Enabled Features
All methods are routed through two endpoints.
-
POST - http:// {your_server} /rpc
- Example: http://localhost:9090/rpc
-
GET - http:// {your_server} /describe
- Example: http://localhost:9090/describe
Note: You can find an exported Postman collections file, "DS RPC.postman_collection.json", in the "postman" directory in this project that contains 10 different requests that you can use for testing all of the methods. These requests will work with the default server configuration that uses an already configured full stack test that is currently running in Optimizely.
Activates an A/B test for a user, deciding whether they qualify for the experiment and bucketing them into a variation if they do.
POST - http:// {your_server} /rpc
JSON Body
Required:
- experiment_key
- user_id
Optional:
- attributes
{
"experiment": {
"experiment_key":"ds_test",
"user_id":"123456789",
"attributes": {"test_user":"true"}
}
}- "variation_key" is appended and returns the assigned variation.
{
"experiment": {
"experiment_key":"ds_test",
"user_id":"123456789",
"attributes": {"test_user":"true"},
"variation_key":"assigned_variation"
}
}Activates an A/B test for a user and returns information about an experiment variation.
This method performs the same logic as Activate, in that it activates an A/B test for a user, deciding whether they qualify for the experiment and bucketing them into a variation if they do. Unlike Activate, this method does not send an impression network request.
POST - http:// {your_server} /rpc
JSON Body
Required:
- experiment_key
- user_id
Optional:
- attributes
{
"get_variation": {
"experiment_key":"ds_test",
"user_id":"123456789",
"attributes": {"test_user":"true"}
}
}- "variation_key" is appended and returns the assigned variation.
{
"get_variation": {
"experiment_key": "ds_test",
"user_id": "123456789",
"attributes": {"test_user":"true"},
"variation_key": "assigned_variation"
}
}Forces a user into a variation for a given experiment for the lifetime of the Optimizely client.
The purpose of this method is to forces a user into a specific variation or personalized experience for a given experiment.
POST - http:// {your_server} /rpc
JSON Body
Required:
- experiment_key
- user_id
- variation_key
{
"set_forced_variation": {
"experiment_key":"ds_test",
"user_id":"123456789",
"variation_key":"variation_1"
}
}- "variation_forced" A boolean value that indicates if the set completed successfully.
{
"set_forced_variation": {
"experiment_key": "ds_test",
"user_id": "123456789",
"variation_key": "variation_1",
"variation_forced": true
}
}Returns the forced variation set by Set Forced Variation, or null if no variation was forced.
A user can be forced into a variation for a given experiment for the lifetime of the Optimizely client. This method gets the variation that the user has been forced into.
POST - http:// {your_server} /rpc
JSON Body
Required:
- experiment_key
- user_id
{
"get_forced_variation": {
"experiment_key":"ds_test",
"user_id":"123456789"
}
}- "variation_key" is appended and returns the assigned variation.
{
"get_forced_variation": {
"experiment_key": "ds_test",
"user_id": "sac123456789",
"variation_key": "assigned_variation"
}
}Determines whether a feature test or rollout is enabled for a given user.
The purpose of this method is to separate the process of developing and deploying features from the decision to turn on a feature. Build your feature and deploy it to your application behind this flag, then turn the feature on or off for specific users by running tests and rollouts.
Feature tests and feature flags or rollouts are called by using the same method and replicates the SDK functionality. If a feature test and a feature rollout are running on a feature, the test is evaluated first.
This methos is the equivalent of IsFeatureEnabled
POST - http:// {your_server} /rpc
JSON Body [ Feature Test ]
Required:
- feature_key
- user_id
Optional:
- attributes
- feature_config
- feature_test_key must be included if you want the "variation_key" to be included in the response for a feature test.
{
"features": {
"feature_key":"feature_2",
"user_id":"123456789",
"feature_test_key": "feature_2_test",
"attributes": {"test_user":"true"},
"feature_config": {"string_2": "string", "boolean_2": "boolean"}
}
}- "variation_key" is appended to the response if the feature is in a test.
- "is_enabled" true or false value is appended to the response.
- "feature_config" has been modified and returns the values assigned to the feature varibles if they exist.
{
"features": {
"feature_key": "feature_2",
"user_id": "123456789",
"attributes": {
"test_user": "true"
},
"feature_test_key": "feature_2_test",
"variation_key": "variation_2",
"is_enabled": true,
"feature_config": {
"string_2": "Text for string variation 2",
"boolean_2": true
}
}
}Retrieving the feature variable values is accomplished by including the variable names and their corresponding data type in the "feature_config" property. You may request all of the variables at once or limit the request to an individual variable or a limited group of variables.
The accepted data types are:
- string
- boolean
- integer
- double
A variable named "screen_width" that has an integer data type would be requested in the following way:
{
"feature_config": {
"screen_width": "integer"
}
}The value for the variable would be returned by replacing the "integer" data type value with the actual variable value. Values are returned in their native JSON supported data type. If you have requested more than one variable they will all be returned in the same "feature_config" property.
{
"feature_config": {
"screen_width": 1280
}
}Retrieves a list of features that are enabled for the user.
The purpose of this method is to get a list of features that are enabled for the user. Invoking this method is equivalent to running Is Feature Enabled for each feature in the datafile sequentially.
POST - http:// {your_server} /rpc
JSON Body
Required:
- user_id
Optional:
- attributes
{
"get_enabled_features": {
"user_id":"123456789",
"attributes": {"test_user":"true"}
}
}
- "features_list" is appended and returns a list of the feature keys that are enabled for the user.
{
"get_enabled_features": {
"user_id": "123456789",
"attributes": {
"test_user": "true"
},
"features_list": [
"feature_1",
"feature_2"
]
}
}You can track conversion events from your code with the Track function. The Track function can be used to track events across multiple experiments. It will be counted for each experiment only if Activate or Is Feature Enabled has previously been called for the current user.
POST - http:// {your_server} /rpc
JSON Body
Required:
- experiment_key
- user_id
Optional:
- attributes
- tags
{
"track": {
"event_key": "test_tracking",
"user_id": "123456789",
"attributes": {
"test_user": "true"
},
"tags": {"revenue": 10000}
}
}- "acknowledgement" is appended and returns "success" or a description of the error.
{
"track": {
"event_key": "test_tracking",
"user_id": "123456789",
"attributes": {
"test_user": "true"
},
"tags": {"revenue": 10000},
"acknowledgement": "success"
}
}GET - http:// {your_server} /describe
The describe endpoint, scans through the descriptions of both the methods and the data types, and returns that information to the caller in the response. This makes it easy for the consumers of the api to always know how to use it, even as it is actively being developed.
{
"types": {
"experiment": {
"description": "Activates an A/B test for a user, deciding whether they qualify for the experimen",
"props": {
"experiment_key": [
"string",
"required"
],
"user_id": [
"string",
"required"
],
"attributes": [
"object",
"optional"
],
"variation_key": [
"string",
"returns the variation key assigned"
],
"project_id": [
"string",
"optional [currently not implemented]"
],
"datafile_url": [
"string",
"optional [currently not implemented]"
],
"datafile_key": [
"string",
"optional [currently not implemented]"
]
}
},
"features": {
"description": "Determines whether a feature test or rollout is enabled for a given user",
"props": {
"feature_key": [
"string",
"required"
],
"user_id": [
"string",
"required"
],
"attributes": [
"object",
"optional"
],
"feature_test_key": [
"string",
"optional - returns feature test key if feature is in a test"
],
"variation_key": [
"string",
"returns variation key for the feature test"
],
"is_enabled": [
"boolean",
"returns \"true\" if feature is enabled"
],
"feature_config": [
"object",
"returns feature variable values for feature test or flags (rollouts)"
]
}
},
"get_variation": {
"description": "Returns the experiment or feature test variation that a visitor would qualify for",
"props": {
"experiment_key": [
"string",
"required (experiment or feature test key)"
],
"user_id": [
"string",
"required"
],
"attributes": [
"object",
"optional"
],
"variation_key": [
"string",
"returns the variation key assigned"
],
"project_id": [
"string",
"optional [currently not implemented]"
],
"datafile_url": [
"string",
"optional [currently not implemented]"
],
"datafile_key": [
"string",
"optional [currently not implemented]"
]
}
},
"set_variation": {
"description": "Forces a user into a variation for a given experiment or feature test",
"props": {
"experiment_key": [
"string",
"required"
],
"user_id": [
"string",
"required"
],
"variation_key": [
"string",
"required (variation to force visitor into)"
],
"project_id": [
"string",
"optional [currently not implemented]"
],
"datafile_url": [
"string",
"optional [currently not implemented]"
],
"datafile_key": [
"string",
"optional [currently not implemented]"
]
}
},
"get_forced_variation": {
"description": "Returns the variation that the user has been forced into",
"props": {
"experiment_key": [
"string",
"required"
],
"user_id": [
"string",
"required"
],
"variation_key": [
"string",
"returns the variation key assigned"
],
"project_id": [
"string",
"optional [currently not implemented]"
],
"datafile_url": [
"string",
"optional [currently not implemented]"
],
"datafile_key": [
"string",
"optional [currently not implemented]"
]
}
},
"track": {
"description": "Track conversion events across multiple experiments",
"props": {
"event_key": [
"string",
"required"
],
"user_id": [
"string",
"required"
],
"attributes": [
"object",
"optional"
],
"tags": [
"array of objects",
"optional"
],
"acknowledgement": [
"string",
"returns \"success\" or error message"
]
}
},
"get_enabled_features": {
"description": "Returns an array list of all the features that are enabled for the user",
"props": {
"user_id": [
"string",
"required"
],
"attributes": [
"object",
"optional"
],
"features_list": [
"array",
"returns list of enabled features"
],
"project_id": [
"string",
"optional [currently not implemented]"
],
"datafile_url": [
"string",
"optional [currently not implemented]"
],
"datafile_key": [
"string",
"optional [currently not implemented]"
]
}
},
"task": {
"description": "A task entered by the client to do at a later time",
"props": {
"user_id": [
"number",
"required"
],
"content": [
"string",
"require"
],
"expire": [
"date",
"required"
]
}
}
},
"methods": {
"experiment": {
"description": "activates the experiment, and returns the assigned variation",
"params": [
"expObj: the experiment object"
],
"returns": [
"expObj: object with variation key assigned"
]
},
"track": {
"description": "track a conversion event",
"params": [
"trackObj: the track event object"
],
"returns": [
"trackObj: event object with acknowledgement"
]
},
"features": {
"description": "activates a feature flag or feature test and returns the variable values if any",
"params": [
"featuresObj: the features object"
],
"returns": [
"featuresObj: object with the variable values if requested"
]
},
"task": {
"description": "[Not functional] creates a new task, and returns the details of the new task",
"params": [
"taskObj: the task object"
],
"returns": [
"taskObj"
]
},
"get_variation": {
"description": "returns the variation for an experiment or feature test",
"params": [
"expObj: the get_variation object"
],
"returns": [
"expObj: object with variation key assigned"
]
},
"set_forced_variation": {
"description": "sets an experiment or feature test variation",
"params": [
"expObj: the set_forced_variation object"
],
"returns": [
"expObj: object with \"variation_forced\" boolean result of true or false"
]
},
"get_forced_variation": {
"description": "returns the forced variation set by Set Forced Variation, or null if no variation was forced",
"params": [
"expObj: the get_forced_variation object"
],
"returns": [
"expObj: object with the \"variation_key\" value if a variation was forced"
]
},
"get_enabled_features": {
"description": "retrieves a list of all the features that are enabled for the user",
"params": [
"featuresObj: the get_enabled_features object"
],
"returns": [
"featuresObj: object that contains the property \"features_list\" with a list of keys corresponding to the features that are enabled"
]
}
}
}
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent