What's New in RapidMiner Studio 10.2.0?
Released: August 16, 2023
The following sections describe what's new in RapidMiner Studio 10.2.0
Features
- Added user interaction after a project was cleaned up on AI Hub
- Ignore it and keep project disconnected
- Overwrite local version by clean check out from AI Hub
- Archive local changes and then overwrite local version as above
- Added Delete Amazon S3 Resource operator
Enhancements
- Migrated Generate ID and Split Data operators to the new Belt data core, future-proofing them and improving their speed.
- Added new setting in the preferences to control whether RapidMiner Studio should favour speed over memory footprint or vice-versa. It can be changed to reduce memory footprint while trading runtime if memory is critical. The setting can be found under
System
and is calledMemory Management
. - Further reduced start-up time of RapidMiner Studio:
- Introduced lazy loading of operators
- Improved utilization of operator signature cache
- Introduced shallow plugin initialization
- Added repository web action to go to deployment endpoints
- Improved error recovery and error messages for date_parse_str function of the expression parser.
- Trailing white spaces are no longer treated as errors in the expression parser.
- Improved opening URL experience on certain Linux distributions which do not support triggering browsing programmatically
- The Correlation Matrix operator now uses the new and improved subset selector
Bugfixes
- Fixed broken error messages in the Edit Expressions dialog that operators like Generate Attributes use to display the expression parser
- Removed deprecated Stream Database operator (deprecated since version 7.5, six years ago)
- Fixed bug in data splitting code that prevented empty partitions in some cases.
- Fixed Synchronize Meta Data with Real Data not working even though it was selected. The selection is now remembered after restart.
- When Synchronize Meta Data with Real Data is activated and the process has been run, Read operators like Read Excel and Read CSV remember the real metadata even if another operator is added to the process.
- Fixed parameters stay above value and stay below value of Prescriptive Analytics operator
- Fixed a possible concurrency issue when writing json IOObjects in parallel
- Fixed potential access denied error for Read Azure Data Lake Storage Gen2 operator when reading larger files
Development
- Added com.rapidminer.repository.recent.RecentDataManager to allow global access to the recently used data sets. It comes with a listener mechanism and currently keeps track of data opened in the Results view, as well as used in the Interactive Decision Tree wizard.
- Removed deprecated classes and methods pertaining to the old concept of managing Perspectives (including MainFrame#getPerspectives())
- Added DeveloperTools#shouldDeveloperToolsBeShown() to allow for an easy way to check whether you want to offer developer tools of some capacity when appropriate
- Fixed bug that caused TableMetaData#columns() to return a meta data sub-table with random column order
- Fixed bug when registering IOObjects from operator signature
- Plugins now properly also look up resources like icons from the default com/rapidminer/extension/resources path. The old additional lookup for com/rapidminer/resources is kept for compatibility reasons.
- Deprecated: SwingTools#addIconStoragePath(String), it never worked