[Design]Experiment Definition

Purpose :

Design for Experiment API's to populate,list,delete the experiment definition
Design the experiment definition schema
Design Data product to map the users/devices to experiments

Design Flow :

Experiment API :

CREATE API :

URL : POST /experiment/create

Input :

User Criteria Input Request:

UserType Request Body

  "request": {
    "expId": "U1234",                  // Unique Id for Experiment
    "name": "USER_ORG",                // Name for Experiment
    "createdBy": "User1",              // Name of the user who created this experiment
    "description": "Experiment to get users with orgname = sunbird",  // Short Description about the experiment
    "criteria": {                      // Criteria for the Experiment
      "type": "user",                  // Type of the Criteria
      "filters": {                     // Array of Filters Criteria with User Filter field Name and List of Values
        "organisations.orgName": ["sunbird"] // List of Values
      }
    },
    "data" : {                         // Experiment Data 
      "startDate" : "2019-08-01",      // Start date of the Experiment
      "endDate" : "2019-08-02",        // End date of the Experiment
      "key" : "/org/profile"           // Experiment Key
    }
  }

Device Criteria Input Request:

DeviceType Criteria Request Body

  "request": {
    "expId": "D1234",                  // Unique Id for Experiment
    "name": "Device_STATE_AP",         // Name for Experiment
    "createdBy": "User2",              // Name of the user who created this experiment
    "description": "Experiment to get devices with state = AndhraPradesh",  // Short Description about the experiment
    "criteria": {                      // Criteria for the Experiment
      "type": "device",                // Type of the Criteria 
      "filters": {                     // Array of Filters
        "name": "state",               // Name of the field
        "operator": "IN",              // Filter Operator 
        "value": ["Karnataka"]         // List of Values
      }
    },
    "data" : {    // Experiment Data 
      "startDate" : "2019-08-01",      // Start date of the Experiment
      "endDate" : "2019-08-02",        // End date of the Experiment
      "key" : "3fksjfksdlfj",          // Experiment Key
      "client" :  "app"                // T
    }
  }

Output :

Process the request and save the request parameters to Experiment-Definition Cassandra Table with status "SUBMITTED"
Response : "Experiment Submitted Sucessfully"

GET API :

URL : GET /experiment/:id

Request Params : ExperimentId

Repsonse : Experiment Details for the respective experiment Id

Sample Response :

Sample Get API ReSposne

{
  "expId": "U1234",
  "expName": "USER_ORG",
  "expDescription": "Experiment to get users with orgname = sunbird",
  "expData": {
    "startDate": "2019-08-01",
    "endDate": "2019-08-04",
    "key": "/org/profile"
  },
  "createdBy": "User1",
  "updatedBy": "ExperimentDataProduct",
  "udpatedOn": "2019-08-02 19:00:00",
  "createdOn": "2019-08-01 10:00:00",
  "criteria": {
    "type": "user",
    "filter": {
      "organisations.orgName": [
        "sunbird"
      ]
    }
  },
  "status": "active",
  "status_msg": "Experiment is actively mapped",
  "stats": {
    "usersMatched": 234,
    "devicesMatched": 0
  }
}

GET LIST API :

URL : GET /experiment/list

Output : List of all Experiments

Sample Response :

Sample Reponse Experiment List

{
  "experiments": [
    {
      "expId": "U1234",
      "expName": "USER_ORG",
      "expDescription": "Experiment to get users with orgname = sunbird",
      "expData": {
        "startDate": "2019-08-01",
        "endDate": "2019-08-04",
        "key": "/org/profile"
      },
      "createdBy": "User1",
      "updatedBy": "ExperimentDataProduct",
      "udpatedOn": "2019-08-02 19:00:00",
      "createdOn": "2019-08-01 10:00:00",
      "criteria": {
        "type": "user",
        "filter": {
          "organisations.orgName": [
            "sunbird"
          ]
        }
      },
      "status": "active",
      "status_msg": "Experiment is actively mapped",
      "stats": {
        "usersMatched": 234,
        "devicesMatched": 0
      }
    },
    {
      "expId": "D1234",
      "expName": "DEVICE_STATE_ANDHRAPRADESH",
      "expDescription": "Experiment to get deviceids with first_access Sep1st to Sep3rd",
      "expData": {
        "startDate": "2019-08-01",
        "endDate": "2019-08-04",
        "key": "/org/profile"
      },
      "createdBy": "User6",
      "updatedBy": "ExperimentDataProduct",
      "udpatedOn": "2019-08-02 19:00:00",
      "createdOn": "2019-07-04 10:00:00",
      "criteria": {
        "type": "device",
        "filter": {
          "name": "first_access",
          "operator": "RANGE",
          "value": {
            "start": "2019-02-10",
            "end": "2019-02-20"
          }
        }
      },
      "status": "FAILED",
      "status_msg": "Devices Not Found matching the Criteria",
      "stats": {
        "usersMatched": 0,
        "devicesMatched": 0
      }
    }
  ]
}

Experiment Definition Schema:

	Field	type	Description
1	expId	String	unique id for experiment
2	expName	String	name of the experiment
3	expDescription	String	descirption of the experiment
4	expData	Map<String,String>	Experiment Data startDate - start date of the experiment endDate - end date of the experiment key/url - key or url of the experiment modulus - mod number of the experiment
5	createdBy	String	name of the user who created the experiment
6	updatedBy	String	name of product that last updated the experiment
7	udpatedOn	Timestamp	last udpated date by the source
8	createdOn	Timestamp	created date for the experiment
9	criteria	Map<String,String>	criteria of the experiment type - Type of Criteria (user/device) filters - Array of Filters
10	status	String	status of the experiment(SUBMITTED/PROCESSING/ACTIVE/FAILED/STOPPED/INACTIVE)
11	status_message	String	status description of the experiment
11	stats	Map<String,Long>	Statistic Data usersMatched devicesMatched

Experiment Data Product :

Purpose

To compute the user/device experiment mapping based on the experiment criteria and save it to elastic search

Input:

Experiment_Definition Cassandra Table

Output

User experiment mapping index to elastic search

Experiment Index

User Experiment :
{
          "name" : "USER_ORG",
          "expType" : "user",
          "lastUpdatedOn" : "2019-07-30T00:00:00",
          "endDate" : "2019-07-28T00:00:00",
          "url" : null,
          "deviceIdMod" : 0,
          "key" : "2392438420",
          "id" : "A1234",
          "userIdMod" : 0,
          "userId" : "d5fd92df-2683-4357-8620-45bc0683f38c",
          "platform" : "portal",
          "deviceId" : null,
          "startDate" : "2019-07-25T00:00:00"
        }

Device Experiment :

{
          "name" : "DEVICE_LOCATION",
          "expType" : "device",
          "lastUpdatedOn" : "2019-07-30T00:00:00",
          "endDate" : "2019-08-30T00:00:00",
          "url" : null,
          "deviceIdMod" : 0,
          "key" : "4584583045489",
          "id" : "B1534",
          "userIdMod" : 0,
          "userId" : null,
          "platform" : "app",
          "deviceId" : "3f3da025f07f2d6e4ead9042710b8b78",
          "startDate" : "2019-07-29T00:00:00"
        }