Support for multiple comment columns of the same type - values are now stored in a list instead of a single string that was overwritten. Term Accession Number handling improved.
Minor bug fix release.
An empty value in the last column of an SDRF no longer causes an error. Better handling of
Support for parsing merged MAGE-TAB documents (delimited with [IDF] and [SDRF] tags on a new line). Minor bug fixes and additional testing.
Full 1.1 release, includes some minor bug fixes detected in the release candidate and full support for rewriting IDF and SDRF objects that have been edited after parsing has completed, or that have been built from scratch.
Release candidate for the 1.1 release of Limpopo. This release is a fairly significant rewrite of portions of the parser and redefines the writing mechanisms, both in terms of MAGE-TAB rendering (after parsing) and MAGE-TAB adapting (converting to a different object model).
This is the first release candidate which omits the ability to write SDRFs from a programmatically constructed graph, as this feature in 1.0.x versions was unreliable and frequently produced incorrect SDRF files.
Minor bug fix release
More robust fix for the case where Factor Value unit attributes are not correctly set on the appropriate factor value
Minor bug fix release
Fix for the case where Factor Value unit attributes are not correctly set on the appropriate factor value
Bug fix release
Factor values are not assigned correctly if there is a Scan Column between the Factor Value and the Hybridization Column due to a bug in the read-back loop.
Fixes for SDRF graph detection to ensure the parsed graph does not contain cycles. This scenario was causing the SDRF Writer to get stuck in an infinite loop. Also controlling character encoding - all SDRFs are parsed on the assumption that they are encoded as ISO-8859-1, for a few reasons. This attribute can be overridden if desired.
SDRF Writing improvements. Anchored positions of sub-attribute columns instead of allowing them to slide around after e.g. Characteristic attributes
Bug fix release
Improved catching of unexpected column names in SDRF, and ensuring these columns are ignored and the graph is still constructed. Channel assignment for Factor Values. ADF parsing improved
Finalised documentation. Reduced critical fail conditions to a minimum - now, only system errors or syntactic errors (normally caused by bad/inconsistent use of quotes) cause critical fails, everything else becomes a warning.
Scalability enhancements. Uses deployed beta version of ErrorItem module. Fix for occasional "wrong number of elements" bug when there was missing node data. This version has been successfully tested against every experiment in ArrayExpress and is almost production-ready.
Release candidate for 1.0 version of Limpopo. Includes SDRF serialization, fixes for assocation errors in Factor Values of multi-channel experiments, and issues with non-registered node types causing the graph to break.
Bug fix release
Mis-association of attributes associated with factor values fixed. Previously, in the case where there were multiple consecutive factor values, of which only the last had attributes (e.g. Term Source REFs), these atributes would be associated with each factor value instead of just the 'last' one. This bug has now been rectified.
Bug fix release.
Finished the migration of SDRF 'lookupNodes()' methods. These methods now return Collections instead of Lists, and the deprecated versions should not be used. Updated documentation.
Bug fix release.
FactorValueNodeHandler fixes, to ensure factor values in "floating" downstream positions are accurately handled. Prevent out of bounds exception if 'handlesName()' is called early.
Bug fix release.
Improved factor value parsing. Protocol location handling is fixed so that now protocols have trackable locations instead of returning '0,0'
Bug fix release.
Additional checking for cyclic SDRF graphs, and fixes for concurrent modification problems that could occur.
Bug fix release.
Fixes for broken edge creation linking to floating Factor Value nodes, present after Protocols.
Bug fix/feature release.
General move towards abstraction of node functionality and progress tracking, allowing individual node implementations to be more lightweight.
Fixed the bug that meant protocols and their parameters could not be disambiguated.