Specific Search Refinement: Tuning Stages

Tip: Tuning stages were formerly referred to as "Pipeline" stages.

Introduction to SmartHub Pipelines Video Walk-through

In this video, Content Strategist Jason Green takes you through BA Insight SmartHub User Experience Tuning stages.

  • User Experience Tuning enables you to enforce consistent search results, map different refiners to the same name, and much more.

  • Note that Tuning stages can be applied to different data sources (search engines), or to all of the data in your site, regardless of source.

  • This demonstration utilizes the SharePoint Online search engine as an example, but any search engine can be used.

User Experience Tuning stages reference

Refer to the table below for information on specific user experience tuning stages that you can select when adding a tuning stage to refine your metadata. For more information on how you can add these tuning stages to your environment, see Add a Tuning Stage.

Stage Description
Property Mapper

This stage translates managed properties names to map them to an existing property on a specific search engine. To get the expected results, this stage must be added to both the Query and Results stages.

See the following example:

OriginalProperty1,MappedProperty1;
OriginalProperty2,MappedProperty2;
PropertyToRemove,;
Url,Path;format,fileextension;fcoid,;
Range Facets

This stage allows you to specify custom ranges for dates and numeric refiners, such as size, LastModifiedTime, etc. This lets your users filter their search results to return items that fall in between the selected interval. SmartHub supports configuration with both overlapping and non-overlapping range intervals, allowing you to specify how flexible or accurate you want the refined search results to be.

To get the expected results, this stage must be added to both the Query and Results tuning stages:

See the following example:

<RangeFacetsConfig> 
    <Facet name="Size" isDateInterval = "false">
        <Intervals minLabel="Less than 200 Kb" maxLabel = "More than 500 Kb"> 
            <Interval startPoint="100" endPoint="200">100-200 Kb</Interval>
            <Interval startPoint="200" endPoint="500">200-500 Kb</Interval> 
        </Intervals>
    </Facet> 
    <Facet name="LastModifiedTime" isDateInterval = "true">
        <Intervals minLabel="+3 years" maxLabel = "Today">
            <Interval startPoint="-1065" endPoint="-7">1 week - 3 years</Interval> 
            <Interval startPoint="-7" endPoint="0">1 week - today</Interval> 
        </Intervals>
    </Facet> 
    <Facet name="Date" isDateInterval = "false">
        <Intervals minLabel="Before 2015" maxLabel = "After 2018"> 
            <Interval startPoint="2015-01-10" endPoint="2017-01-10">2015-2016</Interval> 
            <Interval startPoint="2017-01-10" endPoint="2018-01-10">2016-2017</Interval> 
        </Intervals>
    </Facet> 
</RangeFacetsConfig>
Refinement Mapper

This stage maps a search engines refinement names and values to match the local refinements. To get the expected results, this stage must be added to both the Query and Results stages.

See the following example:

Property#OriginalName1|OriginalValue1|...#MappedName|MappedValue
Format#docx|docx|doc|docf#Word|Microsoft Word
NLQ Processing This stage extracts information from the user's search query using machine learning algorithms to better understand the query text. This stage is already configured in the NLQ Service Settings.
Query Playback This stage overwrites the query with a previously captured query in JSON format.
Results Playback This stage overwrites the results with a previously captured results set in JSON format.
Query Recorder This stage saves the executed search queries in JSON format. You must use local paths to make sure you have write access. The file is created relative to the root level: ".\Folder"
Results Recorder This stage saves the received search results in JSON format. You must use local paths to make sure you have write access. The file is created relative to the root level: ".\Folder"
Query Scripting Processor This stage processes search queries and translates managed property names so that these names map to an existing property on a specific search engine. Use this stage to process the received search query requests by providing complete scripting access to the SearchQuery object.
Results Scripting Processor This stage processes search results and maps the search engines refinement names and values to match the local refinements. This stage post-processes the returned search results by providing complete access to the SearchResults object
Suggestion Query Processor This stage performs suggestion query processing and allows search queries to be processed.
Query Suggestions Results Processor This stage processes query suggestion results by providing complete access to the QuerySuggestions, ResultSuggestions, and PeopleSuggestions members of the SearchResults object. All results stages have access to ExecutedQuery, which is a component of the SearchResults class.
Similar Documents Query Processor This stage allows processing similar documents queries using VB scripts.
Similar Documents Results Processor This stage allows processing similar documents results using VB scripts.
Admin Relevancy This tuning stage can be used to boost keywords found in document properties (metadata) by a specific factor (number). Only full page URLs are supported, for example, https://smarthub.bainsight.com/Results.html.
Synonym Processor This stage expands queries to include synonyms. SmartHub supports One-Way and Two-Way Synonyms.
Graph Boosting Stage This stage boosts results that are more relevant, according to the Microsoft Graph API.
Personalized Query This stage performs personalized query adjustments allow you to you to append information to an actively running user query.
SharePoint OnPremise Refiner Mapper This stage maps a search engine refinement names and values to match the local refinements. Each parameter must be on a new line and must have a semicolon at the end of the line.
SharePoint Online Refiner Mapper This stage maps taxonomies so that the taxonomy labels are displayed and not the GUID. Each parameter must be separated by a semicolon.
Elastic Aggregator Stage This stage alters the elastic search query so that Admins can append aggregations.
Elastic Translator Stage This stage translates the search query into a JSON which is sent to and consumed by the elastic search query.
Longitude Security Trimmer This stage handles the native security trimming feature for the BA Insight Connectors. This stage should be placed before the Property Mapper stage.
Graph Claims Provider This stage enhances the current list of user claims using the properties from the Microsoft Graph API.
Learn to Rank Results Boosting This stage boosts results based on clustering algorithms.
Learn to Rank Results Suggestions This stage suggests results based on clustering algorithms.
Advanced Security Module This stage handles the native security trimming feature for the BA Insight Connectors. This should be placed before the Property Mapper stage.
Vector search This stage enables vector search functionalities for Microsoft Azure search engines.