Multiple visual and performance improvements and bug fixes in Process Discoverer plugin
Added plugin for importing event logs in a comma-separated value (CSV) format into Apromore
Version 7.1
01/05/2019 (7.1.190501). Release 7.1.0 on GitHub
Visual improvements and bug fixes in Process Discoverer plugin
Integrated SiMo discoverer to extract process models annotated with simulation parameters from event logs
Fixed layering order of boundary events in bpmn.io editor
Version 7.0
09/04/2019 (7.0.190409). Release 7.0.1 on GitHub
Headless installations like Docker used to fail because of Slickerbox-related code in the UIContext of ProM utilities. Changes in ResearchCode bundle now avoid this.
Opening a model version from the lower-left Process Details panel in the portal now uses bpmn.io rather than Signavio (#188)
The BPMN canoniser is now more lenient about accepting incomplete models. Specifically: saving models with isolated events (neither incoming nor outgoing flows) no longer fails, nor do unconnected data output associations or complex gateways.
15/03/2019 (7.0.190315). Release 7.0 on GitHub
Replacement of Signavio with bpmn.io as the default editor of Apromore. bpmn.io is a lightweight process model editor offering a wide range of plugins supported by a large community of developers and researchers. In addition, most plugins that were previously built on top of the Signavio environment such as Log animation and model-model as well as model-log Compare plugins are also migrated to the bpmn.io environment.
Added ability to directly edit models in compare plugin and recompute differences with the original log
Visual improvements in Log animation plugin
Several User Interface improvements in Predictive monitoring plugin
Support for LDAP authentication
Compliance with licensing requirements of dependency libraries
Support for SSL (encrypted session)
Containerization in Docker
Various bug fixings
Version 6.0
11/09/2018 (6.0.180911)
Various bug fixings, incl. issues with arcs creation and undo/redo in Editor.
31/08/2018 (6.0.180831)
New predictive monitoring plugins to train and predictively monitor event streams (using Apache Kafka). You can train a variety of predictive models for regression (predicting the value of a continuous variable) or classification (predicting the value of a discrete variable), such as decision trees, random forests, gradient tree boosting, using a variety of features encodings and bucketing mechanisms. You can compare the trained models in terms of their relative merits along earliness and accuracy dimensions, and then deploy the selected ones to a run-time environment for real-time predictive monitoring from event streams (currently, a stream can be obtained by replaying a log stored in Apromore). Predictions can then be visualized in a dashboard and exported to CSV
Total reimplementation of the process map plugin discovery. Now this plugin supports switching from process maps to BPMN models and vice-versa (by integrating Split Miner), filters, export of logs, models, pdf and png, process maps/BPMN models log animation and much more
Updates to drift detection/characterization plugin: the event-based configuration now handles noise in event streams, allows one focus on the most significant process changes by decreasing the drift detection sensitivity (this functionality is particularly useful when detecting drifts on real-life event logs whose originating processes incur frequent changes), and when using a log already existing in the repository of Apromore, one can now save the sub-logs between detected drift next to your original log in the repository
Some plugins renamed to better reflect their functions
Various bug fixes.
Version 5.0
13/04/2018 (5.0.180413)
New analysis plugin “Check compliance against CTL rules and guards”
New plugin “Mine process performance on process map” added (invokes process map discovery with performance statistics)
Help button added to several plugins, linking to how-to manual
Some plugins renamed to better reflect their functions
New examples added to public distribution
Various bug fixes.
13/03/2018 (5.0.180313)
New discovery methods added: Split Miner and the latest version of Inductive Miner. Users can now configure the parameters specific to each method.
Improved discovery plugin “Visualize Log”: now shows process maps (directly-follows graphs) of an event log with frequency or performance statistics (e.g. min, max, avg). Activity and path frequency filters have been improved and now it’s possible to replay a log on top of the process map.
New analysis plugin: Check compliance with Declare allows users to check the compliance of an event log against a set of business rules defined as a Declare model
New discovery plugin: “Extract variability specification” allows one to extract a set of Linear Temporal Logic (LTL) formulae from a BPMN process model
New analysis plugin “Verify compliance” allows users to check the compliance of an event log against a set of business rules defined in LTL, which can be provided as input or extracted from a BPMN model for conformance checking using the “Extract variability specification” plugin.
New discovery plugin “Filter out infrequent activities” allows users to remove infrequent activities, as opposed to individual events, which is done by the “Filter out infrequent behavior” plugin.
New analysis plugin “Check model guidelines with BEBoP” (available both from the Portal and from the Editor) allows users to check that a BPMN model.
New analysis plugin “Verify correctness with BProVe” (available from the Editor) allows users to verify the syntactic correctness of a BPMN model (incl. multi-pool collaboration diagrams)
New cloud nodes hosted by University of Melbourne (Australia) and University of Camerino (Italy).
Various bug fixes.
6/03/2017
New plugin: Mine Process Performance builds cumulative flow diagrams, time series and detailed stage-based performance statistics from an event log.
New plugin: Staged Miner discovers a stage-based model from an event log, by automatically detecting milestones to separate the log into sub-logs, one per stage.
New plugin: Visualize Log shows the direct-follows graph of an even log, where activity and edge filters can be applied.
New version of Measure plugin now shows structural complexity metrics when a model is selected, as well as log statistics when a log is selected.
New version of ProDrift plugin now characterizes a drift discovered in terms of behavioral relations, and prompts these to the use as English statements.
New version of Compare plugin visualizes model2log differences as changes overlaid on the BPMN model.
New version of BPMN Miner plugin now supports infrequent behavior filtering of the event log.
BIMP process simulation plugin now accessible directly from Analyze menu (as well as from the Apromore Editor).
OSGi Portal plugins can now have icons to show in the menus.
We moved to ZK 8.0.2 and Apache Tomcat 7.0.61 on Eclipse Virgo 3.6.4 (previous versions won’t work).
New Apromore branding (“Advanced Process Analytics Platform”) and logo.
Two distributions: “one-click” is a single file to unzip somewhere and run Apromore from localhost, and “full-fledged” is the fully customizable version.
Various bug fixes.
Version 4.0
1/09/2016
Native support for logs. Logs are now first-class citizens as process models. They can be imported and exported in XES and MXML, and be used as input to plugins, e.g. discovering a BPMN from a log.
Apromore fully supports PQL (process query language) to query the repository based on process behavior. This includes a major reduction of the size of the PQL applet.
New plugin Compare allows the following comperisons based on process behavior: model2model, model2log (conformance checking), log2log (variants mining).
ProDrift 2.0 released: gradual drift and event-stream based drift detection now supported.
BPMN Miner 2.0 released: a-posteriori structuring is now integrated into BPMN Miner.
OSGi architecture has been restructured to support 3 types of OSGi plugins: portal, logic and editor. A large number of existing components have now become OSGi plugins.
Menu items have been rearranged according to the phases of the BPM lifecycle (discovery, analysis, redesign, implementation, monitoring).
We moved to MySQL 5.7 (previous versions won’t work).
Various bug fixes.
Version 3.4
3/03/2016
From the editor, one can now measure a variety of process model complexity metrics (size, CFC, structuredness…).
GED matrix can now be computed on a selection of folders and is multi-threading.
iBPStruct has been released: structuring process models with BPStruct now works again and the quality and extent of structuring has been improved.
Installation of Apromore has been substantially simplified (reduced footprint – Apromore can now run from a USB stick).
Various bug fixes.
Version 3.3
29/06/2015
Apromore is now capable of identifying concept drift in logs using the ProDrift plugin.
New models can now be automatically discovered using process mining through the Automated discovery plugin based on BPMN Miner.
Bug-fixing in the model merging function: now merged models obtained from the merging of more than 2 models are configurable.
Version 3.2
25/03/2015
Log files of process model execution traces in the XES and MXML formats can be animated on top of a BPMN model using the editor’s “Animate logs…” (play) toolbar item.
Process models with variants (e.g. merged models or configurable models in general) can show selected variants highlighted using the editor’s “Make selection…” (rainbow) toolbar item. Highlighting can be performed either on explicit groups of variants, or by frequency.
Configurable process models can be converted between the EPML and BPMN formats without having configurability information stripped. This is because now CPF supports configuration elements natively. Configurability information can still only be displayed and edited as BPMN, however.
The “Design/Define questionnaire:model mapping” menu item launches the C-Mapper applet for a selected model. This can be used to author .cmap files which define how the answers to a questionnaire in the QML format should configure the variation points of a configurable model. The .cmap files can subsequently be used by the pre-existing Quaestio applet to perform the configuration. With the addition of the C-Mapper to Apromore, a Questionnaire Designer is the only missing part of an entirely web-based toolchain to create and configure process models by questionnaire.
Improved handling of BPMN data modeling (Apromore no longer discards BPMN I/O specification elements, and only connects data associations to item-aware elements, rather than directly to flow elements).
Bug-fixing in the model merging function: now merged models are configurable.
Performance of clone detection has been improved.
Apromore can now be built, and Applets can run, under both JDK 7 and 8.
Several libraries have been converted into OSGi.
Version 3.1
16/07/2014
New and improved management of users, this will speed up performance on server side and reduce registration time (in the future user groups will be available we are already working on it).
Configurable Process Models. Under Home/BPMN/Airpots we have a set of configurable processes.
If you select one of them and click on Design/Questionnaire-Driven Configuration you will be able to configure the process based on your requirements. You can also create new configurable process models by clicking on Design/Define model: questionnaire mapping. What you need is a questionnaire and a process model with configurable gateways. A gateway can be turn configurable selecting it and tipping on “configurable” in the right panel). At the moment you can only reuse existing questionnaires or loading them from your local machine. Soon you will be able to create a new questionnaire directly on Apromore (we are already working on it).
Apromore knows you. On the top left corner you will receive a welcome with your name every time you login.
Version 3.0
15/04/2014
Tolerate and fix various EPML errors during import.
Support for EPML Data Objects and Relations.
Changed C-BPMN gateway variants.
Support for reset nets in PNML.
Conversion from BPMN tasks with interrupting events to reset nets in PNML.
Tasks can now be converted to simple transitions or to start transition – place – end transition in PNML.
Bug fixes to the database for persisting PNMLs.
10/02/2014
Added support for editing PNML models.
7/02/2014
Added support for editing PNML models.
Added some extra plugins to change annotation shape sizes for when we edit a model not in it’s native format.
Fixed the import of PNML models.
New Versioning and Branching support in the Portal and Manager.
The Manager now has a scheduler support, currently only programatically.
The Manager now supports metric calculations via plugins (first plugin for model size).
Added the ability to run the Similarity Search from a folder level not the whole repo.
Quaestio has been added to Apromore.
Support for configurable C-BPMN models in the Editor.
Added Print / Export to PDF to the Editor.
Lots of Bug Fixes to the Portal.
Lots of Bug Fixes to the Editor.
Known Issues:
Plugins for converting annotations from/to PNML absent. Auto-layout for PNML absent.
Saving a marked PNML place produces the number of tokens on the marked place. Extra attributes besides tokens aren’t saved.
Even through we have a screen to setup security for users the information isn’t used consistently. The viewing of models uses the info but not the rights to edit or delete.
The filtering in Clone detection doesn’t work correctly (decimals not shown, needs ZK 7).
The code for clustering clones needs lots of refactoring and the HAC algorithm isn’t working.
GED Matrix construction only works against the whole repo not against folders, and so thus the clones clustering that depends upon it.
To create a new scheduled task you have write code. UI to create tasks on the fly needs to be implemented.
Portal and Manager security needs improvement. The Portal and Manager need to be integrated together so full authentication and authorisation can be supported.
Editor has some issues importing and exporting some elements of some languages, e.g. Positions and Objects in EPCs.
BPStruct support has some issues, either the json we send is incorrect or BPStruct has an error.
Similairy search doesn’t use nor shows in the results merge models, i.e. models that don’t have native format.
Double-clicking on a merge models won’t edit it becuase of lack of native format.
Documentation on a process model can be edited but is not shown – needs UI.
Version 2.1
23/09/2013
Lots of portal bug fixes.
Editor changes, don’t show the old header and make the display a little larger.
Added a WebDav service to apromore to store files, will be used for Questio, Native formats and antoher other file needed for apromore.
Simon has added a fair bit of work for Questio.
Running ZK 6.5.3 CE version now as we didn’t use any EE features. yet.
25/07/2013
OSGI-ified the rest of the Apromore project (Portal, Manager-Client and supporting projects).
Bundled the ZK libraries into a new location, they have to uploaded into your local repo before a build will work now.
The start for Spring WS security between the Manager and it’s built client. This is very difficult and could be removed in the future.
Upgraded the ZK libraries to the last version 5 release, 5.0.12.
14/06/2013
Converted the simple graph implementation to use the Canonical Process Format for the GED Matrix Calculations.
moved the simple graph implementation into apromore and stripped code to be clean and fast try and help debug any issues.
added the AML to BPMN Annotation pre processor.
Updated the CFP Importer to not worry about format and handle loading whole folders at a time.
Fixed the Clustering visualisation to disable and enable buttons correctly and stopped exact clones from showing in the d3 graph.
Minor big fixed with the XPDL canoniser to handle the chnages (from a while ago) to the CPF and ANF formats.
04/06/2013
Updated The call to Signavio to not auto select everything.
Fixed the BPMN in Signavio so we can now save a modified model.
03/06/2013
For some languages there are source to target language annotation processors that make sure the nodes are the correct shape and line up correctly.
Create default annotations for Merged models that don’t have a source language.
Removed Routing / Gateway names from being displayed in signarvio.
Clustering can now have the GED Matrix generated by selecting a button in the clustering screen.
Fixed clustering so that the create clustering and show clustering now show the correct information.
various UI fixes.
26/04/2013
Fixed lots of bugs with the construction of BPMN models.
Integrated a new version of the Version Control.
Made the native format of the mode part of the data for each version. We will always save a native format even if the model is incomplete can not be processed by the RPST.
introduced load time weaving for Eclipse Link, this greatly increased the performance of queries.
New algorithm for the Similarity Search.
Clustering GED Matrix Calculation is now a lot faster.
various UI Fixes.
02/04/2012
Logs of bug fixes within the UI and manager.
Both Save and Save As now work for the editor.
Further work on the OSGi framework to make Apromore more compliant to the standard.
Performance fix for the UI. now the UI only requests the details for the models that are displayed, not all of the repository.
Cleaned up the UI menus to not show the menu items that haven’t been implemented.
Some fixes to the canonisers for various BPM Modelling languages.
10/03/2012
Osgi has been implemented but there still a long way left to go. Currently only the plugins are osgi bundles but soon the manager/ portal will be bundles as well.
Updated the CPF format to contain more information. useful to changing from executable YAWL to BPMN.
Now have XPDL, BPMN, YAWL, ARIS, PNML and EPML canonisers.
Lots of bug fixes and speed improvements to version control and clustering.
Fixed a bug with Similarity Search not giving the same results after each run.
Basic security settings and UI improvements.
25/07/2012
Lots of work done to include (Chathura’s) Version Control and Clustering into Apromore.
Lots of updates to the build in maven, now allows for Maven 3, jdk7 and tomcat 7 for running.
Slowly adding more tests to the test suite to try and pick up errors more quickly.
Version 0.2 (AKA 1.0)
23/11/2010
Annotations: for each process version available annotations are displayed.
GUI for editing process enhanced: user can not make incompatible choices.
Users can choose to edit annotation only. Oryx will disable features so users cannot modify any part of the process structure
Similarity search: still ongoing work on Reina’s algo integration
Merge: same.
1/12/2010
Similarity search issues fixed
Merge: ongoing work
27/1/2011
Short message located at the bottom of the main window is re-initialised at each action
Similarity search keeps going even in case of faulty process versions
For similarity search and merging, greedy algorithm is the default choice.
31/1/2011
Search for similar processes: query process displayed
Scores for returned process versions still to be fixed
12/02/2011
Merge works
Similarity search results: scores are displayed
02/03/2011
Results from similarity search are displayed according to their score
12/04/2011
Squeleton for initialising epml process fixed
SimpleSearch? component accessible by anonymous users (public)
Display error message in popup window when similarity search fails
15/04/2012
First major release in a while. Added PNML canoniser for the Woped integration. Also, major structural and architectural changes to the code base have been completed.
Quite a few jUnit test have been included to help ensure the functionality of the application is correct. (A lot more is needed).
Version 0.1
23/07/2010
Release notes accessible from Apromore poral.
GUI import/export/edit improved
21/07/2010
User is given the choice of using annotations or not when exporting or editing a process model.
When editing process model meta data users can choose owner
Ranking can be null (not necessarily 0 or 1, ..)
30/7/2010
When saving modifications with Oryx, users are given the choice to override the current version (only if the current version is a leaf of the derivation tree) or to derive a new version from the current one (the new name must be different from all existing version names for the same process).
4/8/2010
Unselect button behavior is fixed.
10/8/2010
cpf_0.4 and anf_0.2 new releases are integrated.
Process model Id is now a double whose decimal representation is concat(processId, versionName-no”.”, timestamp).
processId remove from table,
Documentation fixed: clic on it to display the full content.
11/8/2010
Ranking number replaced with fancy stars. Still needs to add interactions so users can rank process versions (currently possible via edit meta data feature).
Feature “create model” enhanced: users are no longer requested to give a version name in the create model dialog. The only one process version resulting by “create process” is the one named in Oryx.
