1.27.18

Update release version to 1.27.18

1.27.17

Highlights

Gemma CLI extensions

Gemma CLI will add JARs from the contrib folder to its classpath. This is mainly used to scan and include externally defined tools. The tools are loaded from the ubic.gemma.contrib.apps package at runtime.

1.27.16

• update baseCode to 1.1.3
• various dependency update
• handle missing values in ebayes implementation (see https://github.com/PavlidisLab/baseCode/releases/tag/1.1.3)
• secure contexts for runnables and callables
• fix missing security context for batched operations in the CLI
• various fixes for gene updates
• better exceptions handling for ontology search
• fix --logger option for the CLI which needed an update to work with Log4j 2
• add support for multiple --logger options
• revert some of the Future-based rewrite for the NCBI gene loader
• add disease models category for annotating datasets

1.27.15

This patch release introduces a new endpoint for downloading raw expression data.

GET /rest/v2/datasets/{datasetId}/data/raw HTTP/1.1
Accept: text/tab-separated-values; charset=UTF-8

1.27.14

Changeset

• faster ranking of term suggestions
• use FileUtils.forceMkdir and FileUtils.forceMkdirParent from Apache Commons to create directories recursively (fix #350)
• fix deletion of EE in some cases (fix #346)
• add support for RAE230B PavlidisLab/GemmaCuration#238
• ignore missing gene when loading coexpression search result (fix #358)
• skip tests when running Maven GitFlow plugin

Highlights

Faster ranking of term suggestions

Term suggestions for annotating datasets are now ranked entirely inside the SQL database engine. This avoids numerous roundtrips to retrieve and count the frequency of terms and drastically improve the responsiveness of the application when searching for frequently used terms. There are still challenges regarding the retrieval of matches from the Lucene search indices for certain queries.

1.27.12

Changeset

• more efficient search results accretion algorithm
• truncate Slack messages to 300 characters
• add more details about the request when logging exceptions on the RESTful API
• fix #330
• fix experiment splitting when factor values are reused (see #319)

1.27.11

This release backports numerous fixes from the development branch that were scheduled for the 1.28.

Changeset

• update Log4j to 2.17.2 and fix the Slack appender
• fix inefficient query for generating the daily datasets and platforms report, as well as other usages in the code, this was likely the cause for #234
• fix silent Hibernate error when removing raw and processed expression data vectors
• use batch when removing in AbstractDao (which caught the aforementioned Hibernate error)
• new local thread pool to keep thread creation under control for short-lived jobs
• make operations in GeoServiceImpl transactional
• require Maven 3.1.1 for building via the Maven Enforcer plugin
• remove CoexpressionQueryQueue usages and comment out its implementation
• fix numerous deprecated external API usages
• remove Ehcache shutdown listener in web.xml since it's already handled by Spring

Highlights

Local Thread Pool

This release brings a new local thread pool to keep thread creation in Gemma under control. The taskExecutor bean can be injected where needed to launch short-lived jobs:

@Autowired
private TaskExecutor taskExecutor;

Or the async flavour with the Future API:

@Autowired
private AsyncTaskExecutor taskExecutor;

Colour in GUI

This is something we can do with Log4j 2! An image is worth a thousand words.

1.27.10

This release adds gene names to the TSV output and fixes an issue with the version not displaying in the CLI or the Web frontend.

1.27.9

This release updates Log4j to 2.17.1 to address CVE-2021-44832.

1.27.7

Changeset

• fix Log4j 1.2 compatibility API not respecting the appender's threshold
• fix a few typos in the log4j.properties files
• add batching support in AbstractDao.update
• fix permissions issue for GeneSet search
• fix non-zero return code for fillBatchInfo when it does not need to run