Amazon Personalize Search Ranking plugin requirements

Before you start personalizing results from OpenSearch, note the following guidelines and requirements for the Amazon Personalize Search Ranking plugin:

• You must use OpenSearch version 2.9.0 or later. If you use Amazon OpenSearch Service, your domain must use version 2.9 or later.

• If you haven't already, complete the instructions in Setting up permissions to grant your users permission to access Amazon Personalize and give Amazon Personalize permission to access your resources in Amazon Personalize.

• You must be able to access your Amazon Personalize resources from your OpenSearch Service domain or open source OpenSearch cluster.

  • For information about granting access for an OpenSearch Service domain, see Setting up Amazon OpenSearch Service permissions.

  • For information about granting access for an OpenSearch cluster, see Setting up open source OpenSearch permissions.

• You can use only custom Amazon Personalize resources. If you created a Domain dataset group, you can still add custom resources.

• You can only use the custom recipe Personalized-Ranking. For more information about this recipe, see Personalized-Ranking recipe.

• You must create an Item interactions dataset in Amazon Personalize. Items and Users datasets are optional.

• You can't apply Amazon Personalize filters when you're using the Amazon Personalize Search Ranking plugin.

• By default, the plugin assumes that the _id for an indexed document in OpenSearch matches the itemId in your Amazon Personalize data. If your OpenSearch data uses a different field that corresponds with your Amazon Personalize itemIds, you must specify the name of the field when you configure the plugin.

• The userId that you use for a user making a query must match their userId in the data you import into Amazon Personalize.

• The plugin re-ranks only the top 500 search results from OpenSearch. The remaining items are not re-ranked and end up at the bottom of the list.