Optimize query pipelines and machine learning for IPX

This is for:

Developer System Administrator

When you create an IPX search interface, two query pipelines are also created for your IPX configuration. The query pipelines and models are created using default settings, and are fully functional with your IPX configuration. However, Coveo highly recommends that you optimize the query pipelines to enhance the user experience. Additionally, the minimum recommended Coveo Machine Learning (Coveo ML) models for each query pipeline are also created and associated with the corresponding query pipeline. This helps increase the relevance of search results and enhances the overall search experience.

Optimize your IPX query pipelines

This section details how to optimize the query pipelines that are automatically configured for your IPX search interface. The following query pipelines and Coveo Machine Learning (Coveo ML) models are automatically created for each IPX search interface:

Note

The recommendation query pipeline and the associated IPXRECS model are already fully optimized for use with the IPX search interface. No further optimization is required.

Search query pipeline

The search query pipeline is the main query pipeline for your IPX search interface. It’s used to process user queries from your IPX search interface to provide enhanced relevancy of search results.

Queries routed through this query pipeline are modified and processed according to the query pipeline rules and assigned Coveo Machine Learning (Coveo ML) models. As such, you should modify the default configuration of the search query pipeline to add additional models and to configure relevance tuning rules, such as thesaurus and ranking weight rules.

Optimization recommendations

The search query pipeline created for IPX already includes an ART model and QS model. For the best user experience, you should consider the following modifications:

Note

  • RGA and Smart Snippets aren’t supported in the same IPX interface. You must choose to use one or the other.

Recommendation query pipeline

The recommendation query pipeline is used solely for IPXRECS to provide relevant suggestions in your IPX search interface. An IPXRECS model is automatically created and assigned to this query pipeline. IPXRECS models can’t be associated with pipelines that contain other model types.

Coveo IPX recommendations
Notes

  • The recommendation query pipeline and the associated IPXRECS model are already fully optimized for use with the IPX search interface. No further optimization is required.

  • IPXRECS models can’t be added manually. To create a new IPXRECS model, you must create a new IPX search interface.

  • A newly-created IPXRECS model may display an error status. This is expected because it doesn’t have enough click events to generate recommendations yet. Once 5 click events are available and the IPXRECS model rebuilds, the error disappears and the model will begin to provide recommendations.

    IPXRECS showing error status | Coveo In-Product Experience

Automatic query pipeline configuration

When creating an IPX search interface, the two query pipelines that are recommended for IPX, including their minimum associated Coveo Machine Learning (Coveo ML) models, are also automatically created. The query pipelines are associated to the IPX search interface using the IPX’s search hub value.

Specifically, the automatic configuration creates the following, where <SEARCH_HUB> and <IPX_CONFIGURATION_NAME> is the name of your IPX configuration:

  • The search query pipeline for your IPX configuration uses the following settings:

    • The query pipeline name is Search pipeline - <IPX_CONFIGURATION_NAME>.

    • The query pipeline condition is Search Hub is <SEARCH_HUB> and Context[IPX_Name] is <IPX_CONFIGURATION_NAME> and Recommendation is empty. This condition is used to route the user queries from your IPX search interface to this query pipeline.

    Example

    For example, if your IPX configuration name is IPX Workplace, the created search query pipeline is:

    IPX search query pipeline | Coveo

  • The recommendation query pipeline for your IPX configuration uses the following settings:

    • The query pipeline name is Recommendation pipeline - <IPX_CONFIGURATION_NAME>.

    • The query pipeline condition is Search Hub is <SEARCH_HUB> and Context[IPX_Name] is <IPX_CONFIGURATION_NAME> and Recommendation is Recommendation. This condition ensures that your IPX uses this query pipeline for IPX recommendations.

    Example

    For example, if your IPX configuration name is IPX Workplace, the created recommendation query pipeline is:

    Coveo IPX recommendation query pipeline

  • The following Coveo ML models that are associated with the appropriate query pipeline:

    • An Automatic Relevance Tuning (ART) model named ART model - <IPX_CONFIGURATION_NAME> with default settings that’s associated with the search query pipeline.

    • A Query Suggestion (QS) model named QS model - <IPX_CONFIGURATION_NAME> with default settings that’s associated with the search query pipeline.

    • An IPX Recommendation (IPXRECS) model named IPXRECS model - <IPX_CONFIGURATION_NAME> with default settings that’s associated with the recommendation query pipeline.

Notes

  • In addition to the automatically assigned models, Coveo recommends you configure and assign other Coveo ML models to the search query pipeline to further enhance the search experience.

  • If your Coveo organization license doesn’t allow access to create a specific type of ML model, that model isn’t created automatically.
Important

If you delete an IPX configuration for which query pipelines and ML models were automatically created, the corresponding query pipelines, ML models, and conditions will also be deleted. Make sure they’re not used by other search interfaces before deleting the IPX configuration.

Deleting an IPX search interface configuration is irreversible.

What’s next?

Test and add your IPX interface to your website.