==================================
==================================
- Corrected the ExpandFunctionDefinitionConverter when the body of a FunctionDefinition is equal to one of the argument(s), like 'f(x)= x'.
==================================
==================================
-
Added a new interface that can be used to write converters to transform SBML documents.
-
Implemented a converter that expands the user-defined function definitions.
-
Implemented a converter from SBML Level 2 with COBRA annotation to SBML Level 3, encoding the COBRA annotations using the FBC version 1 package. The converter from FBC Version 1 to Version 2 will arrived in a later release.
-
Started to set up validation so that we can customize the SBMLError messages returned to the user. The system allow internationalizing of the error messages if needed. Added a default post message for all SBMLError, to try at least to give the id or metaid of the problematic element.
-
Modified the FormulaParserLL3 to allow to have different behavior when encountering the 'log' operator. Modified as well FormulaCompiler to output 'ln' for the FUNCTION_LN type.
-
Added a new method in Species: getDerivedSubtanceUnitDefinition, which allows getting the derived substance units only, never dividing by the compartment units.
-
Added a new constant and method to allow invalid SBML in other situation than just reading a file, like validation.
-
Included nested CVTerms in the tree structure of JSBML and made them reachable through all the filters.
-
Added some helper methods from ModelPolisher to convert gene association string into FBC version 2 objects.
-
Added a way for the users to set their tidy instance if they want to change the default options.
-
Added a JSBML tutorial for total beginners in SBML.
-
Implementation of validation started for the FBC and Multi packages.
-
Implementation of validation substantially complete for the Qual and Layout packages.
-
Corrected the way we read SBML Level 1 ParameterRule, the 'name' XML attribute in this case need to be set to the 'variable' class attribute.
-
Modified the offline validation API so that we have access to the ValidationContext in all method, in case we need extra information, like all the enabled check categories or the package version.
-
Corrected the offline validationnot to validate SBasePlugin recursively as the TreeNode.children() method already validates their children.
-
Corrected the SBMLErrorConverter class so that any missing elements from the XML returned by the sbml.org validator REST API would be fine. Improved the error messages displayed to the user when we cannot parse properly the String returned by the sbml.org validator API.
-
Corrected the method ASTNode.setName not to revert the type to FUNCTION for FUNCTION_RATE_OF.
-
Corrected the date written to XML as a comment when the SBMLWriter has an application name set (issue #91). The minutes were written instead of the month in the date.
-
Fixed the parsing of the modulo operator. Thanks to Chris Myers who reported the problem.
-
Fixed Reaction.equals so that it works all the time. An exception was thrown in L3V2 when using the method getFast(). Thanks to Chris Myers for reporting the problem.
-
Simplified the way XMLResourceBundleControl load resources.
-
Fixed the method Species.getDerivedUnitDefinition so that the default value for hasOnlySubstanceUnits is used for SBML level and version below L3.
-
Corrected an inconsistency between getDerivedUnitDefinition() and getDerivedUnits() in Species. Thanks to Matthias König who reported the problem.
-
Made sure to convert the units to SI before comparing them. Corrected the method UnitDefinition.convertsToSIUnits as it was provoking a lot of unnecessary warnings.
-
Added a new constant to store the invalid XML differently from the unknown XML so that we can validate them separately.
-
Fixed FormulaParserLL3 so that it transforms FUNCTION_LN into FUNCTION_LOG if it has more than one child.
-
Corrected the way we output 'power' ('^') in infix formula. Added more test to put parenthesis when the arguments of 'power' are logical or relational operators.
-
Fixed the brackets used for denominator when writing MathML to infix formula.
-
The method AbstractSBase.checkAttribute(String) cannot be used when cloning an object as the parent is not set when cloning and many checks depend on the parent to be set to work properly. Thanks to Mihai Glont for reporting this problem.
-
Corrected an indexing problem in the RDF annotation writer.
-
Change the way we set the max element depth as it does fail inside Matlab.
-
Removed the line wrapping by default in the tidy XML writer as it could cause problem by cutting into several lines the old COBRA-style SBML.
-
Small fixes to be able to compile the code with Java 9.
-
Write the id and name of comp related SBaseRef elements into the comp namespace.
-
Checking that the model and the LayoutPlugin are defined before using them in the RenderParser.processEndDocument method. Fixing GitHub issue #132, thanks to Mihai Glont who reported this error.
-
Fixed the clone methods in the render package. Removed some potential NullPointerExceptions.
-
Relaxed a bit the check on the FBC chemical formula when reading an SBML file but the test should probably be relaxed to prevent losing information.
==================================
==================================
-
Updated the implementation of the SBML Level 3 Multi package to the Version 1.0.7 specification (August 2016).
-
JSBML is now deployed to Maven central to make it easier for people using maven to integrate it.
-
Several corrections to the Spatial package, when writing to XML. Thanks to Kaito Ii for reporting the errors.
-
Updated the implementation of the SBML Level 3 Render package to the draft specification (April 2015). Plenty of XML elements where not written with the correct names. Any files created by JSBML before version 1.2 should still be readable by JSBML 1.2 so that you can just read them and write them to get the correct XML.
-
Returning the correct namespace for core elements when calling getNamespace() or getURI() (instead of 'null' previously).
-
Updated the XML attribute name of Samplefield.interpolation. Changed it from 'interpolation' to 'interplationType'. The change should be transparent for the user as the API is unchanged and when reading a file, we are reading both attribute names.
- Miscellaneous updates:
- We have made numerous updates to the JSBML User Guide.
==================================
==================================
-
The minimum JDK requirement is now 1.7.
-
Added SBML L3V2 support, except for support of VCard4.
-
Added support for the MathML annotations.
-
Added the ASTNodePlugin interface so that ASTNode can have plugins.
-
Added new initDefaults method in most SBase based classes with an additional Boolean flag 'explicit' as parameter. This allows, for example, setting the SBML Level 2 default in an SBML Level 3 class. If the explicit boolean is set to true, the default values will be written to XML.
-
JSBML now tries to retain any unknown XML elements encountered while reading an XML source. This makes it possible to write them back, and users can also access them through a user object XMLNode.
-
Added a static function parseMathML() in ASTNode to make the function easier to find. Thanks to Jason Zwolak who suggested it.
-
Added two methods, hasTerm(int) and hasTerm(String), in SBO to allow users to safely check if a term can be retrieved for a given SBO identifier without having to catch a NoSuchElementException.
-
Updated the implementation of the SBML Level 3 Groups package to the final Version 1, Release 1 specification.
-
Integrated a draft version of an offline SBML validation system that was begun during GSOC 2016 by Roman Schulte. The current version should only be used for testing; it is not complete and should not yet be used in production.
-
The toString() methods of most classes have been replaced by a generic method so you might find some changes. Please don't hesitate to provide us with feedback about this change.
-
We deprecated the Model.getInitialAssignment() and Model.getRule() methods because they have been made ambiguous by the introduction of an id on SBase in SBML L3V2. The JSBML JavaDoc will point you to the replacement methods. The new method names make it explicit what is used to retrieve the element, for example 'Model.getRuleByVariable()'.
-
We introduced a new class, UniqueSId, to help recognize any class that must have a unique SId in the whole Model. This class replaced UniqueNamedSBase which was not change and can help recognize any class that must have a unique SId in the whole Model before SBML L3V2. The same applies to SBase and NamedSBase, every SBase have now an id and a name and the NamedSBase class only apply to elements that had an id and a name before SBML L3V2.
-
A whole SBMLDocument was not serializable because a few classes did not implement the Serializable class. This should be fixed now. Thanks to Jens Einloft who reported the problem.
-
Fixed the SBMLReader class to make sure the UTF-8 encoding was used when reading an SBML model from a String. Thanks to Matthias König who reported this problem.
-
Corrected an encoding problem in the tidy SBML writer that happened when the default JVM encoding was not UTF-8.
-
Several improvements were made to the infix formula reading/writing. Thanks to Miguel de Alba who reported some problem and the libSBML team who provided a large set of test cases.
-
When writing a 'sbml:units' attribute on MathML elements for real numbers with e-notation, the XML tag was closed before the attribute was written. Thanks to Matthias König who reported this problem.
-
The clone constructor for SBaseRef was not cloning properly its child SBaseRef.
-
The equals method of ASTNode was failing when one of the double values was set to NaN.
-
Corrected the 'metaid' attribute value pattern; it did not include the colon character.
-
Corrected the FBCModelPlugin clone constructor so that it properly cloned the 'strict' attribute.
-
Corrected the Boundary and SampledFieldGeometry clone constructor. Thanks to Kaito Ii who reported the problem.
-
Corrected several potential NullPointerExceptions that could occur during the cloning of an SBMLDocument that contain plugins.
-
Corrected a potential NullPointerException in the TreeNodeAdapter(Object, TreeNode) constructor.
-
Corrected the number returned by ASTNode.getReal() in the case of REAL_E, when the mantissa is NaN or infinity.
===========================================================================
===========================================================================
-
Implemented the ability to support multiple version of SBML Level 3 package definitions.
-
Improved the Maven pom files, and changed the Maven group ids and artifact ids. They should be stable now. With these changes in place, we can update the Maven repository much more quickly and easily. Next steps are to make JSBML available in Maven Central.
-
Added some utility classes to improve support for SBML files written by the Cobra toolbox before the SBML Level 3 FBC package was created. The CobraFormulaParser class can parse properly the Cobra gene associations String into an ASTNode. The CobraUtils class contain a method to parse the SBML notes into a Java Properties object.
-
Updated the implementation of the SBML Level 3 Groups package to the Groups version 0.7 specification.
-
The compartment 'units' attribute was not written in SBML L1. This is corrected now.
-
SBML L1 rules were not correctly written: the wrong XML element name was used. Now it uses the correct L1 names and attributes.
-
The clone constructor for the Species class did not copy the speciesType attribute. Thanks to Olivier Martin who reported the problem.
-
The clone constructor for the FunctionDefinition class did not copy the id attribute.
-
Since the merge with the astnode2 branch in 1.1-beta1, the attributes on mathML elements where processed before calling the main processAttributes method and as a consequence, some of them where not processed at all, like 'sbml:units'. Thanks to Matthias König who reported the problem.
===========================================================================
===========================================================================
-
Changed the default JSBML formula compiler. WARNING: this changes the default output for comparison operators. The parsing is now very close to what the libSBML L3 parser is doing.
-
Added support for the nested CVTerms introduced in SBML L2V5 and L3V2. We created some new methods in the CVTerm class to be able to manipulate nested terms.
-
Merged with the ASTNode2 branch (experimental). This is not used by default, but it is starting to work well. We were able to pass all tests from the SBML Test Suite using the ASTNode facade that uses the ASTNode2 classes underneath. We made this work so that no API change is needed -- old code should work without modifications.
-
Changed the printing of ASTNode in SimpleTreeNodeChangeListener to avoid many exceptions being printed when reading an XML file and the log4j DEBUG is activated (SF.net tracker item #88).
-
Improved CellDesigner LayoutConverter.
-
Added a new class, PackageUtil, to check and fix the package versions and namespaces if needed. This method is called after we read a file from XML or before we write the model to XML. In the near future, the method SBase.registerChild will take care of setting everything properly.
-
Added a default load of the parsers for environment not setup properly for it (Eclipse, Matlab dynamic path, OSGi, ...).
-
Modified the way we initialize the XML factories so that it is explicit (to OSGi for example) which classes and packages are used. Then there is no more need to define the System properties about the XML factories which solve an other reported problem with OSGi bundles. Thanks to all users who reported and tested problems related to OSGi.
-
Implemented a new interface CompartmentalizedSBase and changed all elements that can have a compartment so that they extend it.
-
Added a createCurve method to ReactionGlyph.
-
Improvements in the arrays flattening code.
-
Updated spatial to implement version 0.90 of the draft specifications.
-
Implemented FBC version 2.
-
Implemented distrib version 1 draft 0.16.
-
Modified the SBMLReader so that we are able to parse some UncertML Strings as XMLNode, even when they are on their own, via org.sbml.jsbml.xml.XMLNode.convertStringToXMLNode(String).
-
Added a new class DistribModelBuilder to help users to create distrib elements, including uncertML XMLNode.
-
Implemented the multi draft rev 459 (Nov 2014).
-
Methods that accepted Double/Integer/Boolean values or returned those where updated to double/integer/boolean in the spatial and render package. Instead of null values PropertyUndefinedExceptions are now thrown when necessary.
-
Tracker item #83: the equals method from StringBuilder is not implemented so we have to compare the content of the StringBuilder or use other utility method to compare properly the StringBuilders.
-
Tracker item #84: corrected and improved the XMLNode.toString method.
-
Tracker item #85: removing spaces with the String.trim() method in org.sbml.jsbml.Creator.setEmail(String) before checking the validity of the email address. Thanks to Camille Laibe who reported this problem.
-
Tracker item #87: changed HashMap to Map in the SBMLValidator API.
-
When writing math formula, in some cases, parentheses were not generated properly surrounding relational expressions. This is now fixed. Thanks to Miguel de Alba for reporting this problem.
===========================================================================
===========================================================================
-
Added a new module that contain a tidy SBMLWriter that use the jtidy library to correct the formatting of the XML file before writing it out.
-
Created a new class SyntaxChecker that corresponds to the class in libSBML and moved pattern matching for id comparisons and e-mail addresses into this class.
-
Updated JSBML to create identifiers.org URIs instead of URNs for units.
- Support for SBML Level 3 'Dynamic Structures (dyn)' package:
- First draft of JSBML implementation done (experimental).
-
Solved error #82 in Unit.removeMultiplier. Thanks to Christian Thöns for pointing this out.
-
The registerChild method now returns a boolean value to indicate if the operation was successfully completed. This prevents users from adding duplicates to ListOf objects.
-
Improvements made to the java API documentation (javadoc).
-
Added a check to write only the necessary namespaces when writting CVTerm to XML.
-
Registering SBaseRef correctly in the document. Corrected the clone constructor of SBaseRef that did not copy the recursive SBaseRef.
-
Added all the missing equals and hashCode methods in the comp package.
- Several fixes and improvements in the validation code and the flattening code.
===========================================================================
===========================================================================
-
Changed the default behaviour of the parsing of math infix formula. The methods concerned are JSBML#parseFormula and ASTNode#parseFormula. The changes include a different operator precedence, to match the one used in the L3 libSBML parser. The parsing is also case sensitive for mathML elements. Boolean operators are now differently interpreted: '&&' and '||' are used instead of 'and' and 'or'. Please consult the JavaDoc of those methods for more details, it is also explain there what to do if you want to keep the old behaviour.
-
Added two new BioModels.net qualifiers, bqmodel:isIntanceOf and bqmodel:hasInstance.
-
BioJava was updated to the latest version and the dependency is now to biojava3-ontology.
-
Added removeCVTerm methods in Annotation and SBase.
-
An IdManager interface has been introduced to help deal with the different id namespace in SBML core and in the different packages.
-
JSBML now provides a convenient ModelBuilder that can be used to create SBML models with a significantly lower number of lines of code for the end user.
-
Improved Java source code documentation (in particular CVTerm).
-
The SBMLDocument has now a convenient method to access all metaIds within the document in form of an unmodifiable Collection.
-
The SBO class provides more methods to access those terms that are of particular interest for SBGN displays.
-
A general Pair of arbitrary values now generalizes ValuePair.
-
The toString() methods in UnitDefinition and SpeciesReferences have been improved and are now more informative.
-
Internally, JSBML no longer creates Unit objects that lack required attributes when reading models. (However, this does not prevent users from creating new Unit objects without such attributes.)
-
Species and Compartment 'containsUndeclaredUnits' method has been overridden because in SBML L3 Species and Compartments can inherit a default unit from the Model. This is now considered. In this case, containsUndeclaredUnits will return false even if the object itself does not explicitly declare its unit. This is important because the unit of the element is not undefined.
-
Added some missing methods to manipulate EvenAssignment in Event and improved the JavaDoc of the Event class.
-
Added some missing getter in the L3 packages classes.
-
Arrays package implemented, version 05/05/14. Only selector and vector supported for now. Validations, flattening available.
-
Added support for vector and selector infix parsing to support the arrays package.
-
Added missing methods to remove GeneralGlyph objects from a Layout object.
-
Added further missing remove or unset methods in the layout package.
- The class QualitativeModel in the qual package has been marked deprecated and should no longer be used. Aiming to achieve better compatibility to libSBML, the new class QualModelPlugin has been created to be used instead.
- Updated to implement version 0.88 of the draft specifications.
-
Corrected the method AbstractSBase#getExtensionCount that was always returning 0.
-
Corrected the SBMLWriter so that disabled packages are not written to XML any more (they are still cloned).
-
Corrected the method SBase.appendNotes that contained several problems.
-
Corrected a bug when cloning a text XMLToken.
-
Fixed the AbstractSBase equals method that could throw a NullPointerException in some cases and was not testing all fields.
-
Fixed the CVTerm.removeResource(String) method. Thanks to Camille Laibe who reported the problem (tracker item #80).
-
Corrected the method ASTNode.parseFormula so that they catch any Exception or Error and throw a ParseException instead.
-
Added the support for the 'root' function in the parsing of math formula.
-
The FormulaParserLL3 was not reading properly formula like 'kf * S0 * S1'. Only two arguments were read instead of three or more. Thanks to Chris J. Myers for reporting this problem.
-
Corrected a bug in Event#setListOfEventAssignments method that prevented the listOfEventAssignments to be registered properly in the model (ids and metaids registered).
-
Corrected the removeFromParent method to make it working for SBasePlugin instances.
-
The ListOf.remove(String) will not throw a ClassCastException anymore when used on list that do not contain elements with id.
-
The static method JSBML.read(String) was not calling the right helper method.
-
ASTNode threw a NullPointerException when trying to un-set its variable.
-
The AbstractTreeNode recursive method removeTreeNodeChangeListener(TreeNodeChangeListener, boolean) was not actually recursive.
-
Solved source for potential NullPointerException in LaTeXCompiler.
package:
- Fixed several issues when cloning 'comp' elements and L3 packages elements in general. Thanks to Chris J. Myers and the iBioSim team for their help to find and fix all issues.
- The required flag for the fbc package was set to true instead of false, thanks to Chris J. Myers for reporting this problem.
-
The implementation of the layout package contained unnecessary classes that all represented a Point object (Start, End, Position, BasePoint1, BasePoint2). In order to avoid confusion and to simplify the implementation, those classes have all been deleted.
-
Corrected the method Layout#setListOfAdditionalGraphicalObjects to avoid the warnings when cloning the list and to avoid to create a new list when not needed.
- The SBML qual attribute Output.outputLevel was not read properly from XML.
- Fixed severals bugs in the experimental render package. Renamed the Group class to RenderGroup and the Curve class to RenderCurve. Added the possibility to add children to a RenderGroup and corrected the way it was written to XML.
======================================================================
======================================================================
-
The main addition to JSBML 1.0 is support for the SBML Level 3 packages, as described below.
-
JSBML is no longer compliant with Java 1.5, and now requires Java 1.6 as a minimum. We believe this is reasonable given the widespread availability of Java virtual machines version 1.6 on diverse platforms.
-
Support for the SBML Qualitative Models package (qual) has been implemented, following the Version 1 Release 1 specification (17 May 2013).
-
Support for the SBML Flux Balance Constraints package (fbc) has been implemented, following the Version 1 Release 1 spec. (11 February 2013).
-
Support for the SBML Hierarchical Model Composition package (comp) has been implemented, following the Version 1 Release 3 spec. (14 November 2013).
-
Support for the SBML Layout package (layout) has been implemented, following the Version 1 Release 1 specification (13 August 2013).
-
(Experimental) support for the SBML Required Elements package (req) has been implemented, following the draft Version 1 Release 0.5 spec. (30 May 2013).
-
(Experimental) support for the SBML Spatial Processes package (spatial) has been implemented, following the working specification Version 1 release 0.80 (latest version is release 0.85, 13 December 2013).
-
(Experimental) support for the SBML Groups package (groups) has been implemented, following the draft Version 1 Release 0.4 spec. (17 May 2013).
-
(Experimental) support for the SBML Render package (render) has been implemented, trying to follow the SBML Level 2 specification.
-
Support for arrays, distrib and multi is not yet available but planned for the final JSBML 1.0 release.
-
A ParserManager class has been implemented to register SBML Level 3 package parsers. This manager also allows JSBML to enable automatically the packages into the SBMLDocument so there is no longer a need to set the package namespace or the 'required' attribute manually.
-
Reading more than one RDF Description element in the RDF annotation block (the official SBML annotation scheme) is now supported. The whole annotation is parsed first into an XMLNode object, and then the official SBML annotations are extracted, if any are found.
-
A new convenience method, removeFromParent, has been added to SBase and TreeNodeWithChangeSupport. It offers the ability to remove any element from the JSBML tree from its parent.
-
An alternative class to parse mathematical formula has been implemented. It is named FormulaParserLL3 and it respects the same syntax as the libSBML L3 parser.
-
Some filter methods have been added to TreeNodeWithChangeSupport (and thus to SBase as well); they allow for recursive queries in the entire data structure.
-
A new filter, XMLNodeFilter, is available to facilitate the manipulation of XMLNode. It allow to do a search similar to the DOM Element#getElementsByTagNameNS(String, String) method on XMLNode.
-
There is now a new ProgressListener utility class that enables tracking the progress of computationally intensive tasks.
-
For better localization support and to separate source code from user messages, a new resource manager has been implemented and is being used in some JSBML classes already.
-
The CVTerm API has been improved with new methods that should simplify the code needed to manipulate CVTerm objects. The javadoc documentation indicates which methods could be used to replace the existing ones.
-
The libSBMLio module for reading and writing SBML using libSBML as an input and output module has been revised and completely re-implemented.
-
The CellDesigner plugin module has been completely re-written and has become significantly more efficient. This facilitates the creation of CellDesigner plugins for existing JSBML-based programs with very little overhead.
-
MAVEN support has been improved. Thanks to Aurélien Naldi for providing several POM.xml files for JSBML.
-
Syntax checking for SBML 'metaid' attributes have been added to the setMetaId method.
-
SBasePlugin now implements the TreeNodeWithChangeSupport interface.
-
Validation checks for e-mail addresses have been added to Creator.
-
A quick Hash-based search method to access elements in a model based on their identifier is now available. In the same way, any SBase can be found in the SBMLDocument based on its metaId in constant time.
-
In order to make the API more Java-like, for each getNumXXX method a corresponding getXXXCount method has been created.
-
The userObject in ASTNode has been replaced by a more general map of userObjects that is located in TreeNodeWithChangeSupport. In this way, any JSBML object that is part of the SBML tree can now be attached with user- defined objects. These will not be written to SBML later on.
-
The Unit class has the new methods removeMultiplier: it merges multiplier and scale. This can only be done if the logarithm to the base 10 of the multiplier is nearly an integer number, i.e., if the distance to the next integer falls below the ensured double precision.
-
The multiplyWith and divideBy methods in UnitDefinition have been improved to better combine units by using the merge method.
-
The Species class now ensures that its spatial size units can only be set if its surrounding compartment does not have zero dimensions and if its hasOnlySubstanceUnits attribute is false. This behavior is specified by the SBML documents for Level 2 Versions 1 and 2, but was not yet implemented.
-
When deriving its UnitDefinition, the Species class searches for a suitable existing unit from the model, rather than creating a new instance. To this end, inherited Level 3 units from the Model are now taken into account.
-
A new JUnit test class helps to check if simplification and merging of Units and UnitDefinitions works correctly.
-
When writing SBML to files, the streams created by JSBML were not closed in all cases. Fixed. Thanks to Akira Funahashi for pointing this out.
-
Appending text in an XMLToken did not have an effect if there was already some text. Fixed.
-
When trying to remove a modifier, reactant, or product from a reaction using the id of the element, where no such element was present before, a NullPointerException was thrown. This has been solved. Thanks to Curtis Madsen for detecting this bug.
-
In the Reaction class, the setReversible method accepted Boolean, instead of boolean. This has been changed.
-
Removed old note comment from the SubModel class.
-
JSBML did not write negative infinity into SBML files correctly. Fixed.
-
Detected and fixed a source of possible NullPointerExceptions within the hashCode method in ASTNode.
-
There was a bug in KineticLaw that prevented a correct iteration through all children. Instead of looking at the listOfLocalParameters, it returned the math element twice (if there was one).
-
When cloning entire models or SBMLDocuments, JSBML assigned the identical TreeNodeChangeListeners to the new data structure. This could lead to unexpected behavior because listeners were receiving change events from data structures to which these have not been explicitly added by a user.
-
Cloning caused problems because in some cases the recursive method that registers all local parameters and other entities with an id did not do a full recursion. Hence, some entities could not be found in the model later on when searching for a certain id.
-
When calling the method setSymbol in InitialAssignment with an instance of Variable as argument, the listeners were sometimes not correctly informed about changes.
-
The setParent method caused an incorrect and in many cases duplicate call of the nodeAdded method in TreeNodeChangeListeners. Instead of registering a property change (the parent link) the parent node has been passed to the nodeAdded method.
-
QuantityWithUnit did not fire update events when changing its value.
-
The unsetUnits method in ExplicitRule now correctly notifies the listeners.
-
The Model class lacked an unsetSubstanceUnits method. Thanks to Robert Byrnes for reporting this problem.
-
The Compartment class did not override the method getDerivedUnitDefinition from its superclass. Hence, Level 3 units could not be inherited from the model. Fixed.
-
Merging of Units did not correctly work in cases where the scales of two units could not be evenly merged: just one of both was used and the other was ignored. In the corrected version, both scales removed represented by a double-valued multiplier in the merged unit. In particular, this error occurred in the addition or subtraction of ASTNodes.
-
If during the simplification of UnitDefinitions a dimensionless unit was created by canceling terms, its scale was always set to 0 and multiplier to 1, which could lead to errors. Additional checks now yield correct results.
-
The units of formulae containing Advogadro's number could not be derived. This is now fixed.
-
In some cases, containsUndeclaredUnits() in ASTNode yielded incorrect results. These are now fixed.
-
[Tracker item #6]: a NullPointerException was returned when using the method Model.findLocalParameters(String) unless at least one LocalParameter was registered in the model.
-
It was sometimes possible to unregister elements from a model using cloned elements that did not belong to the model.
-
SBMLValidator fixed according to the changes made on the sbml.org validator to add package information.
- If you don't have the code handling a package in your classpath, jsbml-1.0-beta1 will loose the package elements, they won't be kept in memory or went writing back the model to a file. The final release should be able to handle unknown packages.
======================================================================
======================================================================
- The new specialized EventObject TreeNodeRemovedEvent has been implemented in order to make sure that deleted nodes do no longer have a pointer to their previous parent within the tree. At the same time, it is still possible to identify their previous location using the field in the new EventObject.
-
When cloning ASTNodes the direct pointer to referenced variables (instances of CallableSBase) are set to null in the cloned object. However, we have to keep the identifier of the variable. Otherwise it might happen that the cloned Tree completely looses its meaningfulness. So far, when calling setVariable with an instance of CallableSBase only the pointer to the object was stored, but not the id. This has been corrected now.
-
When creating an ASTNode with a direct pointer to a CallableSBase, no reference to the id of the SBase has been stored.
-
There is now a rounding correction for multipliers when merging units that tries to shift multipliers to the scale of a unit. This can only be done if the logarithm to the base 10 of the multiplier is nearly an integer number. Nearly means that the distance to the next integer is not larger than the ensured double precision.
-
[Tracker item #3511439] : The cloning of a Species was not copying the speciesType attribute. Thanks to Thomas Bernard who noticed and reported the problem.
-
The units attribute on ASTNode was sometime not written out and the sbml namespace was sometime missing from the mathML. It should be fine now.
- Reading of more than one RDF Description elements in the RDF annotation block (the official SBML annotation scheme) is not supported at the moment. Planned for the next major release.
- The RDF block at the top level of the annotation is considered to be the official SBML annotation.
- SBML L3 packages are not supported and not kept.
======================================================================
======================================================================
- Major improvements of the reading and writing speed for large models.
-
It was forgotten to implement an unsetSubstanceUnits method on the Model class. Thanks to Robert Byrnes for reporting this problem.
-
In the Reaction class, the setReversible method only accepted Boolean, no boolean.
-
Several bug fixes and improvements to units derivation, merging and simplifying.
-
There was a bug in KineticLaw that prevented a correct iteration through all child elements.
-
When cloning JSBML wrongly assigned the same TreeNodeChangeListeners to the new data structure. Now, the TreeNodeChangeListeners are not copied when cloning a JSBML object.
-
The clone methods were corrected to register properly all entities with an id.
-
QuantityWithUnit did not fire update events when changing its value because of an incorrect comparison between the previous and the new value.
-
Corrected a bug that made it impossible to set a kinetic law with an existing list of local parameters as the new kinetic law in a different reaction.
-
Meike Aichele detected and fixed the problem that some Listeners were set too soon in the Model class, so that no Event was detected.
-
One rdf namespace declaration was missing when creating an History element from scratch.
-
[Tracker item #3511439] : Model.getSpeciesType(String) was corrected and it's speed improved. Thanks to Thomas Bernard who noticed and reported the problem.
-
Annotation elements that are missing a proper namespace declaration will be correctly read and not considered as improper SBML elements.
-
[Tracker item #3487517] : If one empty rdf Bag element was present in an annotation read by jsbml, an XML element was not closed properly, resulting in an invalid generated SBML. Thanks to Anna Zhukova to report this problem.
-
[Tracker item #3434930] : The parsing of a Message as an XML String was incorrect as the method was expecting only a notes element to enclose the HTML.
- Reading of more than one RDF Description elements in the RDF annotation block (the official SBML annotation scheme) is not supported at the moment. Planned for the next major release.
- The RDF block at the top level of the annotation is considered to be the official SBML annotation.
- SBML L3 packages are not supported and not kept.
======================================================================
======================================================================
-
NamedSBase has now the new method isIdMandatory in order to check whether the identifier of this element must be defined in the SBML representation.
-
The SBMLReader can now be used in connection with a TreeNodeChangeListener. This allows users to keep track about the parsing process of XML files.
-
The new interface TreeNodeWithChangeSupport separates manipulation of change listeners from SBase, allowing multiple implementations.
-
The JavaDoc has been extended: It now also includes a statement when ever a method might throw a PropertyNotAvailableException. This class was called PropertyNotAvailableError before. In order to avoid confusion with Java Errors, this class has been renamed. Also the super type XMLError has been renamed to XMLException in order to come closer to the default Java naming convention in JSBML.
-
ASTNode has now a userObject attribute through which any computation result or other additional information can be stored in the tree.
-
Restructured the extension package: There is now one extra directory including a specialized source folder for each extension. The core does no longer contain the ext package. Besides the source folder, there is also an extra doc folder that contains the description of the type hierarchy of the implementation for this package.
-
Data structures for the spatial extension are implemented.
-
When accessing the i-th child of an AbstractSBase, it is no longer necessary to sort all extension packages as these are now maintained in a sorted data structure.
-
The compile method in ASTNode has become more efficient now by restructuring its case distinctions.
-
Implemented the hashCode method for all elements belonging to the hierarchical SBML data structure (SBase, Annotation, ASTNode etc.).
-
In order to get closer to the Java package structure, the change listener for SBML objects is now located in the package org.sbml.jsbml.util package. For the sake of more similarity to standard Java classes and a greater flexibility, the SBaseChangedListener has been renamed to TreeNodeChangeListener and now extends the PropertyChangeListener in the package java.beans. The same holds for the corresponding SBaseChangedEvent that is now called TreeNodeChangeEvent and also located in the package org.sbml.jsbml.util. Furthermore, all elements of the SBML tree, including annotations, ASTNodes etc. now make use of change events and notify their listeners, which are now organized in a list instead of a set to make sure that they are always called in the same order. The management of these listeners is now performed in AbstractTreeNode, the new super class for all recursive elements in JSBML. Thanks to Sarah Rachel Müller vom Hagen to support this effort.
-
Notes, annotations, and extension packages are now part of the tree representation of the JSBML data structure.
-
The new super class AnnotationElement has been introduced to gather all those elements that can be used to annotate instances of SBase.
-
The new Type AbstractTreeNode implements basic recursive functions and is now used as the new super class for ASTNode, AbstractSBase, the new type AnnotationElement, and XMLToken (as the super class of XMLNode). In this way, recursion and iteration over child elements within the SBML data structure now follow a unified structure and can be accessed uniquely within all such elements, including annotations and extensions. Furthermore, the new interface TreeNodeWithChangeSupport that is extended by SBase and AbstractTreeNode gathers methods to add, remove, and notify listeners about changes within the data structure (such as adding, removing or exchanging child nodes, or change of any attributes).
-
A simple implementation of AbstractTreeNode, the class TreeNodeAdapter has been implemented that contains some methods also present in the interface MutableTreeNode. However, it does not fully implement this interface because there are also some methods that are not suitable for JSBML purposes. The TreeNodeAdapter works as a wrapper class for elements that are part of the SBML data hierarchy but that do not implement the TreeNode interface. With this adapter, it is possible to include the list of Creators within a history that is not a ListOf object, or the list of CVTerms in an annotation in the hierarchy.
-
Sebastian Fröhlich provided some contributions to the parser for the SBML level 3 layout extension package. JSBML now supports experimental reading for this package. The full stable API for reading/writing of SBML level 3 packages is planned for the next release of JSBML.
-
The SBMLReader now follows the factory pattern by providing static methods to directly read SBML content from Files, InputStreams, or Strings. Thanks to Martin Gräßlin for this idea in tracker issue #3300433.
-
JSBML now provides a Maven script file. Thanks to Igor Rodchenkov for contributing this file.
-
Where necessary, initDefaults has been complemented with an additional method, for which the level/version combination can be directly specified, i.e., it is now possible for those elements to set their defaults as these would be for the specified level/version combination.
-
A basic check to avoid setting the same meta identifier more than once has been implemented. It maintains a set of metaIds on the SBMLDocument. Thanks to Stefan Hoops for making clear that such a check is mandatory.
-
A check for duplicate identifiers has been implemented. JSBML now registers all elements that might have an identifier in hashes mapping the id to the object itself. Different rules depending on the specifications of the levels and versions of SBML ensure that only allowable identifiers can be accepted. To this end, the new interface UniqueNamedSBase tags all those elements whose id is intended to be unique within one model.
-
When adding an element with undefined Level/Version combination or at least with undefined Version to an element whose L/V configuration has been defined, JSBML now recursively updates this configuration for the child element. In case of differing L/V configurations an exception will be thrown on adding the element.
-
Corrected mistakes in the org.sbml.jsbml.util.Maths class, where some trigonometric functions were incorrectly calculated.
-
When writing SBML Level 1 the name attribute (which has become the id attribute in later SBML versions) was never written because JSBML uses the id only as an internal representation. The check for Level 1 name attributes was not correct.
-
In SBML Level 1 MathML code was written in KineticLaws although in this SBML version formulas are represented in infix notation only.
-
When creating a Parameter object using the constructor that takes a LocalParameter as its only argument, a call to setConstant(true) was always performed, irrespective of the fact that for Level = 1 models this method throws a PropertyNotAvailableError. Thanks to Sarah R. Müller vom Hagen for reporting this problem.
-
[Tracker issue #3323886]: The units of Species were not derived correctly because if the units were not explicitly defined, the method getDerivedUnitsInstance() always returned the predefined unit "substance", or null for models in Level 3. Now it considers the surrounding compartment, the hasOnlySubstanceUnits, and spatialSizeUnits attributes depending on the Level/Version configuration.
-
Cloning of SBMLDocuments was not done correctly: The set of already existing metaIds was initialized after adding the model to the document. In the mean time, there could already be access to this set.
-
When reading SBML content from a file or writing JSBML data structures as SBML to a file, the SBMLReader/Writer did not close streams after finishing. This caused that sometimes files could not be opened by other programs although writing was already done. Unfortunately, by correcting this bug, an IOException needs to be thrown, but previously, a FileNotFoundException was already sufficient. Hence, users need to catch a different exception now when reading/writing SBML.
-
Renamed the getters for Creator and ModifiedDates in History to getListOfCreators() and getListOfModifiedDates (before it was spelled without "Of", for the sake of having equal names for all such getters.
-
If the Level/Version combination was not set, the identifier of NamedSBases was sometimes not written to an SBML file.
-
[Tracker issue #3300430]: In some condition, the JVM was not checking the type of an Object before adding it to a typed variable which later lead to some ClassCastException that prevented jsbml to save some models. Thanks to Martin Graesslin to report the bugs alongside some unit tests and to help resolved it.
-
[Tracker issue #3300490] : The method SBase.getNotesString() was returning null instead of the empty string, as stated in the java documentation. This has been corrected and the empty String is returned now. Thanks to Martin Graesslin to report the bugs alongside some unit tests.
-
The SBML level 1 rules were incorrectly read, now when the 'type' attribute is set to 'rate', a RateRule is properly created.
-
The SubModel utility class was fixed and an example of how to use it added.
-
Corrected the clone method for listOf objects. The typeOfList was not set in the clone so it was not possible to save a model that contain a cloned list.
-
[Tracker item #3306135] : the handling to the number NaN (NotANumber) has been improved and the correct mathML element is created when saving the formula to XML.
-
The attribute denominator on level1v2 SpeciesReferences was never read. Thanks to the numerus models in the SBML Test Suite to help us detecting this problem.
- Reading of more than one RDF Description elements in the RDF annotation block (the official SBML annotation scheme) is not supported at the moment. Planned for the next major release.
======================================================================
======================================================================
-
The new interface CallableSBase ensures that only instances of this type may be referenced in ASTNodes. In this way, it is no longer possible to refer to Events from ASTNodes.
-
Additional type of error: abstract PropertyError with its two sub- types PropertyNotAvailableError and PropertyUndefinedError. The PropertyNotAvailableError one indicates that a certain property of an SBase is not defined for its current Level/Version combination, whereas the latter one indicates that the value belonging to a mandatory property, for which there is no default value has not been defined. At least in the Trigger class, PropertyUndefinedErrors are now thrown when attempting to access the value of the persistent or initialValue attribute for models with Level >= 3 and an undeclared value. Thanks to Roland Keller and Lucian Smith for pointing out this necessity.
-
The class Maths, a compilation of useful mathematical functions, provided an implementation of the factorial function that took a double as its argument. However, for this purpose, the Gamma function must be implemented. We changed this function to take integer values only.
-
A NullPointerException could occur when deriving the unit of an expression containing a call to the factorial function. Thanks to Roland Keller for reporting this.
-
[Tracker issue #3208887] : When trying to access the previously set value for UseValuesFromTriggerTime it was possible that a NullPointerException could occur due to the attempt to clone this element. This has been fixed. Thanks to Sebastian Fröhlich for reporting this.
-
[Tracker issue #3208903] : Priority elements were properly read and written. Thanks to Sebastian Fröhlich for reporting this.
-
[Tracker issue #3204275] : It was impossible to set the size of a compartment for models in Level 3 if the spatial dimensions was undefined, i.e., NaN. Since it is only prohibited to set the size if the dimensions equal 0, we provided another test in the set method in Compartment to allow for this.
-
Corrected a typo in the method SBO.getAntisenseRNA(), the method was always returning -1. It is now returning the correct value (317), thanks to Anushya Muruganujan for reporting this.
-
[Tracker issue #3175906] : Removed plenty of unused methods in the Annotation class that were dealing with writing the annotation as XML. Everything is handle now in the stax SBMLWriter. The getAnnotationString method was not present in the SBase interface, it is now added and should be the only method used to get the Annotation as XML.
-
[Tracker issue #3175911] : compartment.spatialDimensions type changed to double to be consistent with the SBML level 3 specifications, as a consequence the return type of the getSpatialDimensions as been changed.
-
[Tracker issue #3176548] : id containing underscores were not parsed properly by the infix formula parser, this is now fix and any valid SId should work. Thanks to Goksel Misirli for reporting this.
-
[Tracker issue #3175833] : annotation tags were not properly closed when a mathML element was present on the same element. Thanks to Goksel Misirli for reporting this.
-
A new interface, CallableSBase, was introduced to encompass all the different SBML elements that can be referenced inside a mathML 'ci' element. As a consequence, some methods signature have changed.
-
[Tracker issue #3175833] : in some cases, an empty annotation was written if a metaid was defined. Some methods were added to the Annotation class to test which part of the annotation is defined.
-
[Tracker issue #3196284] : the unit exponent was incorrectly always read as an integer. This has been corrected and if the level of the model is 3, the exponent is read as a double.
-
[Tracker issue #3196638] : the message constraint was not properly read, it was stored in the notes instead of the message ! This is now corrected, thanks to the person who reported that.
-
Corrected several bugs in the SBMLValidator and implemented properly all the classes related to SBMLdocument.checkConsistency() so that is works as expected.
-
[Tracker issue #3199792] : localParameter were not parsed correctly in SBML level 3. Thanks for the persons who reported it.
-
[Tracker issue #3216183] : all the XML elements that were belonging to the RDF namespace were sent to a parser that understood only the specific SBML RDF annotations. This as been fixed and we can now have some RDF elements anywhere in the annotation.
-
SBML level 1 rules when the 'type' attribute is equal to 'rate' are created as AssignmentRule instead of RateRule.
-
The support for the mathML elements notanumber and infinity is not complete.
-
Reading of more than one RDF Description elements in the RDF annotation block (the official SBML annotation scheme) is not supported at the moment.
===========================================================================
===========================================================================
This the first beta release of JSBML 0.8, JSBML is a community-driven project to create a free, open-source, pure Java library for reading, writing, and manipulating SBML files and data streams. It is an alternative to the mixed Java/native code-based interface provided in libSBML.
For more details, please visit http://sbml.org/Software/JSBML
- [Tracker issue #3137967] Namespaces in the sbml element were incorrectly parsed, some known namespaces like html were not kept for examples. This is fixed, thanks to Paulo Maia for reporting this.
-
problem of output indentation for notes, when there is several top level elements.
-
the method SBMLDocument.checkConsistency() is not working. There is a problem in the parsing of the XML return by the sbml.org validator.
-
dependency on the sbml.org validator which might not be robust enough (you cannot submit a file bigger 15MB for example). It would be better to be able to install your own localy or at least have several mirrors. We will have to see how it performs.
-
in a mathML block, if one 'ci'element is suppose to represent a functionDefinition but the id is not a valid id, the type of the ASTNode is not set properly, so there are some errors when trying to use the corresponding ASTNode tree. (cf model generated by COBRA)
The line should not have more than 76 characters to avoid them being put over two lines when sending the release email. Although, since we are using markdown we do not cut the lines any more.