This API is called to invoke the encoding process in the recommendation engine. The goal is to input a CSV file containing the asset viewing information of the users and convert it into corresponding implied rating (as per the chosen rating algorithm). The rated data is stored as a CSV file and the access URL is returned on completion. This encoded CSV file can be used as input to the Recommendation Rating Computation Engine.

POST DATA:

Notes:

• RatingAlgorithm:
  “explicit”: Do not apply any pre-processing. Data itself has explicit ratings.
  “aggressive”: Computation algorithm requires a low fraction of view time for a high rating.
  “conservative”: Computation algorithm requires a higher fraction of view time for a high rating.
  “tag_aggressive”: This algorithm is used to compute ratings for individual tags as per a user’s viewing patterns. Aggressive rating assignment is used.
  “tag_conservative”: This algorithm is used to compute ratings for individual tags as per a user’s viewing patterns. Conservative rating assignment is used.

• RatingData:
  This field contains the URL from where the CSV file for the raw data can be retrieved.
  In the CSV File, each row has these fields: userId, itemId, rating
  userId: string
  itemId: string
  rating: float

• Catalog:
  This is the CSV-formatted info about assets in the system. Note that, it is useful when RatingData has media-ids as itemId and RatingAlgorithm involves computation on tags. The tag-information for differentitemId (specified in the Catalog file), is used by the corresponding algorithm.

  Each row in the CSV file has these fields: itemId, tags, type, seriesId, duration
  itemId: string
  tags: pipe-separated list of tags
  type: string //vod, series, episode
  seriesId: string // in case the asset is an episode, this is the itemId of its series. Otherwise “”
  duration: int // runtime in seconds — for series it will be sum of runtime of all its episodes

• On successful completion, a callback is made to specified Callback endpoint and following JSON is POSTed:
  
  { "status": "completed", "id": "<request_id>", "output_url": "<file_url>" }

  1 2 3 4 5

  {     "status": "completed",     "id": "<request_id>",     "output_url": "<file_url>" }

  
  The final file is stored at the provided “output_url” with each row in the CSV file having the fields: userId,itemId,rating
  userId: string
  itemId: string
  rating: float

  In case of failure, a JSON with error message is POST to the callback endpoint:
  

  { "status": "error", "id": "<request_id>", "msg": "<error message>" }

  1 2 3 4 5

  {     "status": "error",     "id": "<request_id>",     "msg": "<error message>" }

Resource URL

Example Request URL

Example Request

Example Return

On success: On error: