Skip to main content

Release Summary - Mar 06, 2024 (24.05)

The following key features and improvements, along with bug fixes, have been released in Algonomy products in the release version 24.05 during Feb 12, 2024 - Mar 06, 2024.

Enterprise Dashboard

Enabled Merchandising Rules to Do Numeric Comparisons

Recommendation Boosting and Recommendation Restriction rules have been significantly enhanced to include numeric comparisons, empowering digital merchandisers to tailor their strategies with greater precision. Merchandisers can now boost products with specific numeric attributes, such as margins greater than 30%, directly targeting high-margin products for recommendation.

To improve clarity and usability within the merchandising rules interface, the Attribute section has been restructured to include intuitive sub-sections: "Comparison with seed product," "Attribute value of recommendation product," and "Apply diversity" for Only Recommend rules. These additions allow for detailed configurations, such as matching or differing attribute values from the seed product and executing numeric comparisons ("Greater than," "Less than," "Between") with ease.

This functionality also ensures products with non-numeric values for the chosen attribute are automatically excluded from the boost or filter criteria. This comprehensive approach is applicable across both Boosting and Rec Restriction rules, enhancing the adaptability and effectiveness of merchandising strategies.

Mar 06 2024.png

Mar 06 2024_1.png

Jira: ENG-27699, ENG-26112

Improved usage of data-rr-placement attribute for glass views in p13n.js

We have introduced an improvement to the p13n.js library that enhances the glass views functionality by allowing front-end developers to use custom ID values for HTML elements while still leveraging glass view events. This update introduces the use of a data-rr-placement attribute, providing a standardized method for developers to pass data without being restricted by specific ID value requirements.

Now, glass view events will trigger when the id or data-rr-placement attribute value matches the placement name, even if the placement name includes a dot (e.g., home_page.recs_1). This dot notation is seamlessly translated to match the underscore notation used in the glass view URL (e.g., home_page_recs_1), ensuring compatibility and ease of implementation.

Jira: ENG-27516

Added Ingredient Cross-sell model to Configurable Strategies

A new feature, the Ingredient Cross-sell model, has been added to Configurable Strategies, catering to the needs of optimization managers at grocery retailers. This model enables the creation of strategies based on ingredients commonly used together in recipes, aiming to recommend products that complement each other.

The Ingredient Cross-sell model includes:

  • Under Personalization Seed, the same options as the "Viewed Together" model.

  • For Personalization Sort Options, it offers "User Affinity" and "Smart Shuffle."

  • Filter Options are streamlined to enable only "Diversify results by category."

  • All options under Additional Options can be customized.

This strategy is particularly designed for use with User Affinity sorting and Diversify results by category filtering to ensure a varied and relevant product recommendation.

Mar 06 2024_2.png

Jira: ENG-27573

Data Engineering, Recommend

Log External Region Id Data in Avro

A new feature enhancement has been implemented to include both the external region ID and currency code information in Avro event logs. This data, crucial for reporting on sales-related metrics like Overall Sales and Attributable Sales, enhances the accuracy and utility of recorded events.

The process involves logging the internal region ID, mapped from the Catalog Regions feed based on the external region ID provided in the API call (rid). In cases where the external region ID is not listed in the catalog feed, it defaults to '-1', and previously, the external region ID information was not retained. Similarly, currency codes are mapped to their respective external region IDs from the Catalog Region feed, defaulting to the 'Default Currency' set in the Portal when not listed, which is often 'D' or 'USD'.

This update addresses issues related to incomplete or outdated catalog feeds, which previously led to default region IDs and currency information being logged, impacting the accuracy of reporting. By logging the external region ID information alongside the internal region ID in Avro events for views, clicks, and purchases, we now have the capability to retrospectively correct or customize reports based on accurate region and currency data.

Jira: ENG-27365

Data Engineering

Added Strategy Type to Site Analytics Report

The Site Analytics Report has been enhanced by adding a "Strategy Type" column, enabling Personalization Users to differentiate between various strategy types. This update allows for a more detailed analysis and presentation of data to clients, highlighting the value of diverse strategies such as configurable strategies.

Key improvements include:

  • Strategy type mappings have been integrated into the strategies via a lookup table.

  • A new column for strategy type has been added to the site analytics table.

  • The strategy type is determined by the method of strategy creation, with types including Basic, Configurable Strategies, Manual, Compound, Parameterized, DSW, Advanced Merchandising, Recommendation Boosting, and Recommendation Restrictions.

Jira: ENG-26229


Enhanced option to use all user segments for InSegment model builds:

We improved our model building capabilities with a significant update that allowed for the automatic inclusion of all User Segments into InSegment strategies. This new feature simplified the process for digital optimization managers, eliminating the need to manually manage the list of segments for inclusion. With this update, activating the feature meant that every segment stored in the User Segments datastore was instantly considered during the InSegment model builds, streamlining the creation of personalized strategies.

This streamlined approach ensured that the most comprehensive and current data informed strategy development, removing the administrative burden of manually updating segment whitelists. It was designed to enhance the precision and effectiveness of personalized recommendations, making the overall strategy more reflective of the latest user behavior and preferences.

Jira: ENG-26778


Enhanced Search Accuracy for Part Numbers

To enhance search accuracy for customers with specific part number queries, we've introduced "trimSearchable" in FIND's search settings. This feature removes special characters from searchable fields, allowing for accurate matches between part numbers and their variations, like "P2B-SC-100" matching both "P2BSC100" and "P2B SC 100."

The "trimSearchable" attribute can be added to any field in findIndexSettings, and search relevance can be adjusted through the portal's configuration. Exclusive to streaming clients, this development ensures precise search results for complex part numbers, and is now fully functional, improving search accuracy as intended.

Jira: ENG-27413

Other Feature Enhancements

The following feature enhancements and upgrades have been made in the release version 24.05 during Feb 12, 2024 - Mar 06, 2024.

Jira #



General Availability



Enterprise Dashboard:

Glass views code to use the rr_recs object to get the list of placements

Glass views will now use rr_recs object to get the list of placements instead of rr_placement_place_holders array.




Enterprise Dashboard:

Enabled ‘end date’ as optional in put in visual merchandising

Visual merchandising is introduced with ‘End time’ as an optional input in visual merchandising. The UI is similar to the rest of the portal UI with similar behavior.





The tagRefinement logic now works with the advanced tagFilter options

After tagFilter has been applied the tagRefinements can now be applied on the resulting contents. It comes with the following criteria.

The tagRefinement is an optional filter.

If there are contents that have the tags that are passed in tagRefinement, then those are returned first.

If there are no contents that match all the tags from tagRefinement, then progressively drops a tag until a matching content is found.

Once you have dropped all tagRefinement tags, returns those contents that matched on the tagFilter.

If there were no contents after the tagFilter logic is applied, then skips the tagRefinement and returns nothing.




Enterprise Dashboard:

Co-occurrence Report – Corrected the label and link for the report

The link for co-occurrence report had incorrect spellings:

Corrected the spelling to “co-occurrence” in URL & Visualizations both.






Find - Search Test Drive | Add "NumFound" value

The Find Search Test Drive has been enhanced with the addition of the "NumFound" value, offering users critical insight into the total number of results found for a given search query. This update is particularly valuable for evaluating the leniency or restrictiveness of search configurations and for benchmarking our search results against those from clients' previous search engines. With this enhancement, users can now see a concrete "NumFound" value (e.g., NumFound: 83) within the search test drive results, facilitating more informed comparisons and analyses.




Enterprise Dashboard:

Improved Handling of Blank IDs in Social Proof API Calls

We've updated the Track Experience API call functionality to accommodate blank values for session ID and user ID, ensuring the call does not fail under these conditions. This change facilitates uninterrupted tracking, treating blank IDs similarly to null or unspecified values, thus aligning with the seamless tracking experience when IDs are provided.

This enhancement addresses potential disruptions for Digital Optimization Managers by eliminating error responses previously encountered with blank ID inputs.




Update Feed Attribute Flush batch size

To address memory capacity issues experienced by specific websites overwhelmed by extensive product details, we've decreased the volume of information processed simultaneously. This adjustment, aimed specifically at the affected websites, significantly reduces the memory required for importing product catalog feeds, ensuring a smooth and successful import process.





Create a new endpoint in play service

We've successfully developed a new endpoint in the play service, specifically designed to manage boost and bury rules through a comprehensive CRUD API. This enhancement introduces a robust framework for manipulating these rules directly via the Play Service API, ensuring smooth and efficient rule management.





Enhanced Sorting Functionality in Streaming Find

To improve sorting capabilities in Streaming Find, properties are now required to be facetable/filterable, sortable, and singularly valued. We've introduced a "sort_*" field in the streaming Solr schema, enabling this field's effective use in streaming snapshots for more accurate sorting.

Additionally, we've implemented "caseInsensitiveSort": true in both the portal site configuration and Find Index Settings Json, facilitating case-insensitive sorting. This update ensures a more flexible and precise sorting process.





Enhanced Sorting with ASCII Folding Filter Factory

We've enhanced our sorting functionality by integrating the ASCII Folding Filter Factory in both Batch and Streaming processes. This update applies an ASCII folding filter to the sort field, improving the accuracy of sorting across diverse character sets.

For Batch, this enhancement is achieved with a specific filter configuration, and for Streaming, we've aligned the Solr schema to include similar capabilities. These improvements ensure consistent, precise sorting by effectively handling a wide range of ASCII characters.





Streaming Find - case insensitive sort

A case-insensitive sorting feature, previously implemented for Batch Find, has now been extended to Streaming Find. This enhancement ensures consistency in sorting behaviors across both batch and streaming operations, improving the usability and accuracy of Find's sorting functionality.


Bug and Support Fixes

The following issues have been fixed in the release version 24.05 during Feb 12, 2024 - Mar 06, 2024.




General Availability



Enterprise Dashboard:

Catalog Enrichment- Previously selected filters were getting unselected

In Catalog Enrichment previously selected filters were getting cleared (unselected)

For example, when you searched for the category using filters and then selected a category and applied the data displayed with the category attribute, and then when you searched for color using filters and then applied color, it was not showing the selected category, but it was displaying only color attribute.

The issue has been resolved now.




Recommend (Email):

In PROD, the engage content recs API call returned PAGEAREA as NULL

In PROD, the engage content recs API call was returning PAGEAREA as NULL, the issue has been resolved now.




Advanced Merchandising:

MCMP Adv. Merch. Rules not Functioning All The Time

The issue with the inconsistent functionality of MCMP Advanced Merchandising Rules has been addressed. By implementing a custom MCMP attribute within Advanced Merchandising, we've ensured compatibility for seed products, even when they're marked as non-recommendable.





Related searches order issue in final response

The issue with the order of related searches in the final response from the rr server, which was not aligning with the Solr response, has been resolved. Adjustments have been made to ensure that the rr server now correctly mirrors the order provided by Solr, maintaining consistency in the display of related search results.




Engage, rrportal:

Invalid XSS HTML elements present in request body

The issue where an "Invalid XSS HTML elements present in request body" error was encountered while saving manual campaign rules has been successfully resolved.




Data Engineering:

Handle null values in scorecard reports

The scorecard reports have been updated to address an issue where certain columns displayed null values due to a join operation being performed on the product ID rather than the productExtID. The necessary adjustments have been made to ensure accurate data representation, resolving the bug and enhancing the reliability of the scorecard reports.




Enterprise Dashboard:

Social Proof - Report is showing RPV in terms of % value

The display issue in the Social Proof Report, where Revenue Per Visitor (RPV) was incorrectly shown as a percentage instead of a currency value, has been corrected. Misrepresentations due to decimal placements in summary RPV values have also been fixed. These adjustments ensure accurate representation of RPV in currency terms and proper handling of decimals, improving the report's clarity and accuracy.





SFI stalled in EU dcs

The stall in the Streaming Find Indexer (SFI) within EU data centers, triggered by a PartitionConsumerException associated with a NullPointerException, has been resolved. Following the correction of this issue, we have successfully processed the latest feed, confirming that the system is now functioning as expected and the problem has been addressed.





Implement edgeNgram(instant search) as property definition level - streaming only

The instant search feature in streaming has been improved by introducing edgeNgramSearchable within findAttributeSettings, removing the need for the searchauto_ copy field in the schema. Now, the searchauto_ field is indexed in the Streaming Find Indexer (SFI) only when edgeNgramSearchable is set to true, allowing for site-specific customization.


  • Was this article helpful?