Semaphore 5.4.0

Semaphore 5.4.0 is a major update to the software and an update to the Semaphore 5.2.2 release.

It includes remediation for the Semaphore Security Advisory - 2021-12-13.

Installation Prerequisites

Prior to installing or upgrading to this version of Semaphore please be sure to review the Third-party software requirements (Linux) or Third-party software requirements (Windows) information as well as the general Planning your Plan your Semaphore Installation (Linux) or Plan your Semaphore Installation (Windows) information for installation server requirements.

Product Notes

This version of Semaphore consists of the following components:

Platform Independent Components

Windows Components

Linux Applications

Integrations

(*) Unchanged from previous release.

(**) Unchanged from previous release but version number updated.

Studio (5.4.0)

Upgrade of all external dependencies

Semaphore Administration

Addition of Semantic Enhancement Server Manager component and SES security improved with adoption process.

New Functionality

• 7945 Testing against a wider range of JDKs
• 9106 Decrease minimum memory requirement for all services other than KMM
• 10104 Make it easier to get CS and SES tokens

Issues Resolved

• 3605 (Expected for release) Tie down /system/console
• 7771 Fixing licence uploading issues
• 9023 Ensure that permissions persist through backup restore to new Studio instance
• 9107 Correct service names in Windows Task Manager

Semaphore Integration

Complete rewrite of the CMIS application.

New Functionality

• 3428 (Expected for release) Visualiser: Include alt labels information in SAYT response
• 7818 Extend visualizer standalone app to allow customer onclick events.
• 8350 Publishing from an on-premise KMM to Semaphore Cloud CS and SES

Issues Resolved

• 8235 It is now possible to publish using the KMM Client and a token

Model Management

Complete rewrite of the KRT application using latest versions of application architectures and Reconciliation Server (Open Refine compliant) now available as Web Service.

New Functionality

• 2532 Rollback changes within task
• 2632 Display of entire model linkage
• 2668 Make easier reuse of Dublin Core properties within a model
• 3148 Selective committal of task
• 4932 Make explicit exclusion of master from reviews
• 5574 Ability to clone model structure alone
• 6269 Allow status of “include alt labels in search” to be remembered across sessions
• 6067 Publishing will only take place for languages supported by Classification Server instances.
• 7567 Allow mapping against external resources, either using OpenRefine API or within KMM using the new Semaphore Reconciliation Service.
• 7988 Improve workflow of adding multiple tags.
• 8002 Arrangement of relationships in picker are now hierarchical.
• 8044 Improve layout of model settings dialogue
• 8397 Additional standard Publisher Configurations now supplied - including new default templates.
• 8552 Restore no longer requires workspace files
• 8801 Related terms can now be founded by alt label
• 8893 Allow the user to select the destination URI for cloned concepts.
• 8907 Allow Mac created files to be part of zip uploads
• 8915 Pagination of alt labels
• 9017 Clarify model name in KMM header
• 9019 Put instantiated data above suggestions
• 9195 Use asynch mode for commits to prevent session timeouts.
• 9196 Apply timeout to all bespoke Sparql and Reports
• 9291 Allow automatic removal of history details on task committal, or after a selectable elapsed time.
• 9382 Warn on upload of large import to task
• 9624 Allow users to define bespoke triples to be included in new models
• 10086 Velocity Writer configuration errors better handled
• 10168 Add select/deselect all to language filter
• 10641 Density template for rulebase generation improved
• 11414 Space Missions expanded for demonstration purposes

Issues Resolved

• 4562 If there are no results in SAYT, do not show Object [Object]
• 7537 Correct presentation issues within Manage SPARQL access interface
• 7661 Time out URL correction
• 7779 Concept metadata refreshing issue sorted
• 7878 Corrections to Publisher config upload errors
• 7892 Prevent “Reviewers” and “Contributors” seeing create review button
• 8020 Side Panels now available in cloned models
• 8041 Escaping issue in Lexical Resources side panel
• 8224 Metadata rendering issue.
• 8307 Clarify clear history button
• 8376 Issue with concurrent imports
• 8445 Editing of preferred labels in non-defined languages
• 8455 It is now possible to generate reports in Publisher using non-URI parameters
• 8594 Deleting label now correctly clears out its Semaphore settings
• 8614 Excel importer bugs
• 8742 Improve performance of SAYT in tasks
• 8847 SPARQL query performance
• 8853 Correct StoreConnection when rebuilding models.
• 8969 Performance improvement to deletion of concept scheme with many concepts
• 8986 Do not allow CS and SES status checks to keep session alive
• 8995 Hit accept/decline buttons from concept edit pane were not appropriate
• 9017 Improvements to display of model/task name in KMM
• 9033 If no licence present, KMM will redirect to studio
• 9037 More explicit SPARQL permissions display
• 9039 Clarification of report descriptions
• 9140 Correct long model name display on model home page
• 9142 Correct long model name display in visualizer
• 9150 Tidy up class display
• 9152 Delete Publisher log files if there are more than 10
• 9156 Ensure that emoticons and emojis are supported throughout
• 9157 Make scrolling of metadata types easier
• 9160 Fix issue displaying publisher button on models ending with a period
• 9267 Performance of advanced search
• 9274 Cosmetic improvement to model settings
• 9275 Cosmetic improvement to tag display
• 9450 Performance improvement adding to concept with many relationships
• 9495 No longer re-run report when logging back in after session timeout
• 9691 Better error message on restoring which leads to file conflict
• 9740 Associative relationship picking respects range
• 9833 Improve tile ordering on KMM home page
• 9868 Improve model ordering on linking page
• 9910 Better handle special characters in SPARQL page
• 9921 Task settings link corrected
• 9923 Correct commit message when creating new concept scheme
• 9924 Correct relationship type in create new concept dialogue
• 9926 Advanced filter corrections
• 9929 Correct comment ordering
• 9991 Copy maxPathCount across to environment away SOLR instances
• 10002 Fix issue with & in URIs in tree
• 10006 Clarify import selectors in Excel importer
• 10059 Correct handling of missing linked models at KMM startup
• 10066 Correct display of Shacl Target relationships
• 10084 Improve rendering time of models home page
• 10099 Fix role display with long user names
• 10244 Correct behaviour of language codes particularly with regard to case invariance
• 10252 Clicking delete in the colour picker no longer closes the dialogue
• 10293 Correct class information in visualizer
• 10298 Creating multiple concept with duplicate filtering fixed
• 10313 Concept details updating accelerated
• 10419 Missing prefix in semaphore-publisher-configuration.ttl
• 10527 Remove publisher webresources from backup
• 10512 Restoring large files caused memory exception
• 10528 JS mapping files removed from artifacts

Document and Text Analysis

Note: The Document and Text Analysis component will only function correctly, if the Classification and Language Server is installed with the separate language packs.

Addition of proximity analysis for model enrichment and Classification Precision and Recall Tool now available as Web Service.

New Functionality

• 6437 Ability to move from evidence to label
• 7673 Make navigation to concept from inspection panel easier
• 7949 Creation of Proximate Phrases analysis
• 8225 DA now reads default classification parameters from the Classification Server instance
• 8314 Ability to open DA rules directly from URL
• 8464 Creation of document set manager for use in text analysis

Issues Resolved

• 7294 Corrections to article navigation
• 7739 Issues removing rulenets in Rulebase Manager
• 7852 Issues with losing sandbox content on hitting back button
• 8170 Upload panel corruption issues
• 10373 DA failed when CS was configured to return hash as metadata

Semantic Widgets and Administration

New Functionality

• 7543 Creation of administration interface over SES indexes
• 7746 Adoption of SES interface

FACTS (5.4.0)

New Functionality

• 8001 By default FACTS wildcard matching do not ignore punctuation
• 8010 Re-organise variant generation within FACTS.

Issues Resolved

• 8460 Last published date now correct in KMM.
• 9177 Emojis are now supported
• 9526 Forced preclusions were generating superfluous rules

Classification Server (5.4.0)

Upgrade of all external third party libraries

New Functionality

• 2534 Classification Server will now return idle time statistics
• 7724 Improve separation of publish sets at publish time to speed up publication
• 7873 Correction to configuration file documentation wrt 102 messages
• 8043 Allow access to port 5058 by default because permissions are handled separately
• 8650 CS can now use unlimited licences
• 9283 Allow .zip and .tar.gz as transport forms for rulebases.

Issues Resolved

• 8110 Bug fix in CAT to sort display issue
• 8826 Allow escaping of / in extract_regex
• 9456 Improvement to language guessing capabilities.
• 11931 Better wildcard numeric matching
• 11140 Error in handling character cut off in large spreadsheets

Semantic Enhancement Server (5.4.0)

Important: We have changed the term data returned from SES in term responses. We no longer supply the concept URI as metadata. It is (and has been) provided as a property directly attached to the concept. The “versions” response has been simplified to remove redundant informatiom.

Upgrade all external dependencies

New Functionality

• 3380 Allow searching of multiple indexes in one pass
• 7746 Lock down security of SES administration functionality
• 8329 Increase default maximum memory allocation
• 9169 Leading spaces no longer break concept mapping stage allocation
• 9997 Use wildcarding to reduce the size of index configurations
• 10075 SES versions response is more useful.
• 11134 Use of Snowball stemmers in concept mapping

Issues Resolved

• 7245 Ensure SES returns matching name if AZ request does not match preferred label
• 7957 Corrections to log file locations
• 8930 Correct build of Standalone SES widgets

SharePoint 2016/2019 Integration (5.2 r55344)

Issues Resolved

• 9702 Fix to support latest Semantic Enhancement Server APIs (Semaphore 5.2.2 and later)

Known Issues (Semaphore 5.4.0)

Installation of Studio (Semaphore 5.4.0)

• When installing to Windows you may get a warning about a failure to determine the version of Java in selected JAVA_HOME. Please ignore this message.
• If you do not have Java 11 installed, when you install Semaphore Studio you will be warned that Semaphore Studio requires “Java version 11 or later.” In fact it requires Java 11. See here for details on how to install it.

Classification Server (5.4.0)

• Because of changes to how text is parsed prior to analysis, some inline style changes in PDFs (e.g. emphasised or italic text) may be treated as separate sentences or paragraphs. This could affect evidence finding where that evidence is now spread over multiple zones.
• Because of tokenization issues, Hebrew, Indonesian, Malay, Pashto and Vietnamese are temporarily being removed from the supported language set.
• We have updated our file type detection algorithm. This may change the determined file type for some documents which may affect how the document is parsed. If this is an issue, then remove the “external_parser” entry in the “classify_document” workflow in the file conf/config.xml to revert to the old style.

Stand-alone Linux Publisher (5.4.0)

Note: The following applies to the standalone publisher only, not the plugin.

• During the Linux installation you may see a spurious message about setting ACL permissions for the tomcat user. Please ignore these.

Knowledge Model Management (5.4.0)

• When downloading configuration or backup files, it is possible that the default Windows zip extractor will report the downloaded file as invalid. This is, however, likely to be due to long file paths in the zip file not being handled correctly by that extractor. An alternative (such as 7-Zip) should be able to correct extract these files.

Semaphore for SharePoint 2013/2016 Integration (4.0 r54942)

The following are known issues with this version of Semaphore for SharePoint 2013/2016:

• Use of custom hierarchical relationship types for synchronization of models may result in incorrect model information in the term store (“has narrower/has broader” hierarchical relationship synchronization works correctly)
• DE2394 - Classification Server version 4.2.0 won't work when adding in CS URL. An error “Version's parameters must be greater than or equal to zero. Parameter name: revision.” is being shown.