1.1.22

This should probably be a new minor release, but the 1.2.0 series is development is already underway, but not yet quite ready for prime time.

In brief:

• rewritten the ontology indexing from scratch using Lucene directly
• add support for stemming exclusions
• better isolate implementation of Jena-based ontology services
• update Jena to 2.13.0
• dropped jena-larq and jena-arq dependencies
• include subproperties when looking for restrictions to use for inferring parents/children
• simplify baseCode configuration
• update dependencies

1.1.16

• add support for ontology imports
• fix a bug with getChildren() and getParents() returning all terms from additional properties when direct is true
• include scores in search results
• include resources for MEDIC and NIF ontologies
• introduce JSR-305 annotations to mark the ontology API's nullability
• use temporary directory for caching ontologies if ontology.cache.dir is unset
• include OBO alternative labels in search index

Note: These release notes also include changes from 1.1.15.

1.1.13

• introduce OntologyService.getParents and OntologyService.getChildren as super-optimized strategy for gathering the parent and child terms of a set of ontology terms
• consistently exclude owl:Thing and owl:Nothing when listing parents and children terms
• cache all restrictions involving additional properties to make listing of parents and children efficient
• exclude deprecated terms when searching and listing parents/children by default and add an option to keep them
• move Jena-specific code under ubic.basecode.ontology.jena
• standardize how ontology resources are displayed and compared
• use owl:deprecated, OBO's ObsoleteClass and ObsoleteProperty to mark properties as obsolete

1.1.12

• fix a NPE when converting a term with a null URI
• remove cobertura to make our build compatible with Java 11 and forward

1.1.11

• drastically reduce memory usage of OntologyService by using OntModel's capabilities for retrieving resources by URIs instead of caching them
• add a new OntologyService interface to ease extension
• add proper support for part of and fix inference for proper part of
• use OWL_MEM_TRANS_INF consistently and remove parameters on OntologyLoader to configure the ontology spec
• suppress JenaException raised if the initialization thread is interrupted
• exclude transitive icu4j dependency from Jena which is old and breaks H2 database

1.1.5

This is a small patch release that add support for list in configuration values using , as a delimiter.

1.1.3

This release introduces checked exception for the ontology search and handling of missing values in ebayes implementation.

1.1.2

This patch release prevents transitive dependency on Log4j 1.2 through Jena which contains known security vulnerabilities.

You might still have to import Log4j 1.2 in your final project, and we suggest using the Log4j 1.2 compatibility API provided by Log4j 2 to that end:

<dependency>
  <groupId>org.apache.logging.log4j</groupId>
  <artifactId>log4j-1.2-api</artifactId>
  <version>${log4j2.version}</version>
</dependency>

1.1.1

Fix a NullPointerException in ontology term children URI which occurs for blank nodes.

1.1

This release was long overdue. I'm including here some of the work that was done since the 1.0.30.

• use fixed locale for number formatting (helps us a long way in standardizing this in PavlidisLab/Gemma#182)
• new counting strategy for nearly identical values that ignore NAs
• dependencies update and sync with Gemma