Release date: 18 December 2019
Component versions:
Solr 8.3.1 |
ZooKeeper 3.5.6 |
Spark 2.4.3 |
Upgrade Instructions
To upgrade from Fusion 5.0.x to 5.0.2 using the GKE installation script (setup_f5_gke.sh
), run the following command:
./setup_f5_gke.sh -c <cluster> -r <release> -n <namespace> -p <project> -z <zone> -i <instance_type> --upgrade
To upgrade from Fusion 5.0.x to 5.0.2 using Helm, run the following commands:
helm repo update
helm upgrade ${RELEASE} "lucidworks/fusion" --timeout 180 --namespace "${NAMESPACE}" --values "${RELEASE}_${NAMESPACE}_fusion_values.yaml"
New features
Predictive Merchandiser
Fusion 5.0.2 integrates Predictive Merchandiser as a component of Fusion AI.
Predictive Merchandiser is an AI-powered merchandising tool that provides insights and recommendations for optimizing search results and product placement.
It provides an easy to use interface, allowing you to:
-
Connect to a Fusion Server data source and simulate the shopping experience of your customers.
-
Pin, boost, bury, and block specific products in your catalog.
-
Create business rules to display custom banners, or redirect customers to a different page on your site, for example.
-
Create search rewriting strategies to manage spelling mistakes, synonym matching, phrase matching, and poorly performing queries.
-
Review query rewriting strategies that were generated automatically using machine learning models in Fusion.
-
Deploy new business rules and search rewrites to the Fusion data source where they can be triggered by live queries.
-
View analytics reports that show, for example, the variation in site visits over time, the most clicked on products, and the most searched for terms.
If you have a Fusion AI license, you can access Predictive Merchandiser by navigating to Relevance > Rules and selecting Merchandiser.
To learn more, see Predictive Merchandiser.
Jupyter Integration
Starting with Fusion 5.0.2, we now provide a Jupyter service that can be run from the Fusion Helm chart. Jupyter is an open-source web application that allows you to create and share documents that contain live code, equations, visualizations, and narrative text.
-
Run/Debug Spark code in Scala/Python (replacement for spark-shell)
-
Run SQL queries via Fusion SQL
-
Debug Scala and SQL transforms in PBL jobs
-
Everything else for which Jupyter is designed
To learn more, see Jupyter Support in Fusion.
Improvements
-
The HTML parser now parses the
application/xhtml+xml
media type by default.
-
The HTML parser can filter HTML content using jsoup. New HTML parser configuration options include:
Configuration option
Default value
Description
excludeFilters
N/A
Jsoup-formatted selectors for elements to exclude from the HTML
filterBeforeMapping
False
Apply exclude filters before performing HTML field mapping
filterBeforeExtractingLinks
False
Apply exclude filters before performing link extraction
-
Index pipelines can now receive nested JSON documents through the Parallel Bulk Loader.
-
The Fusion UI now supports undo for deleted rules.
-
A new boolean property,
shareState
, is introduced for Javascript Query Stages. When set totrue
, a single JavaScript engine will be used for all queries on a Javascript query stage. Previously, a unique engine was used for each thread.Under certain conditions, this can improve performance significantly.
-
If a rule creation API is incorrectly formatted in JSON, Fusion now produces a detailed error message.
-
Improved the output from phrase extraction jobs. It is now more useful for non-signals datasets.
-
Improvement made in error propagation between services.
-
The Windows Share SMB 2/3 Connector now parses data into individual Solr fields instead of a large JSON document.
-
The Custom Spark job has been replaced by a Custom Python job.
-
Various UI improvements and fixes.
Bug Fixes
-
Fixed a bug that sometimes resulted in the
<app>_recommender_models
collection to delete the<app>_recommender_models
collection.
-
The
collection_alias
field is now required for time-based partitioning in the Fusion SQL service.
-
The roles selector in the New User panel now correctly displays the available role types.
-
Fixed a bug that prevented the Box Connector from correctly indexing PowerPoint files.
-
Fixed a bug that prevented the Box Connector from indexing Japanese text with UTF-8 character encoding.
-
Fixed a bug that prevented V2 connector jobs from succeeding if a fetcher was defined without a phase name.
-
Fixed a bug that produced an Out of Memory (OOM) error when indexing large JSON files with a V2 connector.
-
V1 connector stateful jobs are no longer dependent on server IPs. Re-crawling after restarting V1 connectors now works as expected.
-
Fixed a bug that removed CJK characters from file names when uploading a file.
-
Fixed a bug the prevented the Jobs API from scheduling a recently created job.
-
Fixed a bug that sometimes caused
DefaultHostInfoMetricReporter
to fail on Java 11.
-
Fixed a bug that caused head/tail analysis jobs to overwrite published rules.
-
Boost lists will now utilize the correct term,
orig_score
, when rewriting a query.
-
Corrections made to the default synonym detection job settings.
-
Fixed a bug that sometimes caused pytests to fail when using the
fusion
user.
-
Logstash is now enabled by default.
Known issues
-
If High performance mode is selected on the Web connector configuration, the crawl does not start.
NoteThis bug is fixed in Fusion 5.1.0.
-
Deploying an application into the webapps service requires a non-empty, non-root context path.
NoteThis bug is fixed in Fusion 5.1.0.