Design for easy way to update content

Problem Statement

Currently to update the content implementation team needs to make multiple calls, which makes it a difficult process. Additionally there is no auditing in place to track as content was updated by whom. There should be an easier way to update the content.

SB-10220 - Getting issue details... STATUS

Solution Approach

Solution approach to overcome the problem has below approaches

APPROACH 1

1. Sunbird Platform will expose an API for end user to initiate update content
2. Sunbird Platform will log the request in DB and pass the user a generated process Id
3. Sunbird Platform will send the request with processId to kafka message queue
4. Request will be picked up by Learning platform and would start the processing of update content based on the request
5. Sunbird Platform will use a listener to get the status update and store in DB from the kafka based on processId

Approach 2

1. Sunbird Platform will expose an API for end user to initiate update content
2. Sunbird Platform will internally call Content Service with the request provided by User
3. Content Service will validate the request
4. With valid request, Content Service will initiate the update process through kafka pipeline and will return unique process ID identifying the process.
5. With Invalid request, Content Service will return error message which Sunbird platform will return to User
6. For valid response from Content Service, processId with other details will be saved in Sunbird platform DB

pros and cons of both approaches

Problem Statement

What is the API for update content?

Solution approach

URL:

POST /v1/content/update

Headers:

Authorization

X-authenticated-user-token

Request Params:

Response Params:

Errors:

Sample Request and Response:

Request

{
	request : {
		"script" : "course_rename",
		"version" : "1.2",
		"content" : {
			"oldName" : "AB",
			"newName" : "ab"
		}	
	}
}

Response

{
  "id": "api.content.update",
  "result": {
		"processId" : "01262366359399628812"
	}
}

Problem Statement

How we store the process information in DB

Solution approach

content_update_process table structure should be as below

Problem Statement

What is the API for getting the process status

Solution approach

URL:

GET /v1/content/update/status/{processId}

Headers:

Authorization

Request Params:

None

Response Params:

Errors:

Sample Request and Response:

Request

GET /v1/content/update/status/01262366359399628812

Response

{
  "id": "api.content.update.status",
  "result": {
		"status" : "FAILURE",
		"message" : "Operation to update failed"
	}
}

Problem Statement

What is the API for get all the processes by a user

Solution approach

URL:

GET /v1/content/update/processes

Headers:

Authorization

X-authenticated-user-token

Request Params:

None

Response Params:

returns an array of processes related to user

Errors:

None

Sample Request and Response:

Request

GET /v1/content/update/processes

Response

{
  "id": "api.content.update.processes",
  "result": {
		processes : [{
			"processId" : "01262366359399628812",
			"script" : "course_rename",
			"version" : "1.2"
		}]
	}
}

Deprecated SOLUTION APPROACH

Solution approach is divided into multiple part

• There should be generated key to facilitate the process of update of content in a channel by anyone who has the key, hence bypassing the user authentication process . The details of the master key generation design and approach are here (Master Key APIs).
• Auditing of the update of the content by keys.