Recommendations

Owned by Vidhya Ananthanarayanan
Jul 01, 2022
1 min read

Prerequisite

Indexing has to be set-up before moving on to Recommendations-related functionalities. For more details, please refer to https://luminoslabs.atlassian.net/wiki/spaces/HC/pages/3717988426

Event tracking is also a prerequisite. Prior to that, an order data export should be done and the resulting files should be uploaded in the Hawksearch Dashboard. This way, the recommendations engine can create meaningful recommendations before any tracking data is gathered from the live website.

Overall Approach

In order to fetch recommendations, the connector provides the Optimizely.Hawksearch.Client package. This package contains classes which will correspond to the requests and responses provided by Hawksearch’s Recommendation API

  • RecommendationWidgetsRequest → Requesting Product Recommendation Widgets

  • RecommendationBulkWidgetsRequest → Requesting Multiple Recommendation Widgets

  • RecommendationResponse


It also provides a client class named RecommendationsClient that will be responsible for calling the mentioned APIs and for deserializing the responses.

For a better experience and easier development, we included in the Optimizely.Hawksearch.Connector package:

The Flow

  1. Using only Optimizely.Hawksearch.Client

  • Create and populate a RecommendationWidgetsRequest or RecommendationBulkWidgetsRequest object (both derive from BaseRecommendationRequest)

  • Use the GetWidgetItems or GetBulkWidgetItems method of the RecommendationsClient class (or their async counterparts)

    public RecommendationResponse GetWidgetItems(RecommendationWidgetsRequest request) public async Task<RecommendationResponse> GetWidgetItemsAsync(RecommendationWidgetsRequest request) public RecommendationResponse GetBulkWidgetItems(RecommendationBulkWidgetsRequest request) public async Task<RecommendationResponse> GetBulkWidgetItemsAsync(RecommendationBulkWidgetsRequest request)

  • Get the RecommendationResponse object then map it to your business models

 

  1. Using both

Optimizely.Hawksearch.Client and Optimizely.Hawksearch.Connector