Version 2.1-beta

This is a beta release with much needed improvements.
Some (non-critical) issues have crept up during testing, that need some care and some smaller features have not been implemented yet that will be in the final release.

Features and Enhancements

• Added a runModelPolisher script to run the dockerized version with BiGG annotation and offline validation. Takes path to a directory containing models to be polished and writes the polished, gzipped models into an out directory inside the input directory as user.
  This runs docker with user id and corresponding group id, so that the output has correct ownership.
  Due to some issues, which are hopefully fixed now, this might hang when trying to find java user prefs.
  Additionally, as both databases are started in detached mode, this will likely crash on first run.
  Start up the DB containers manually before first run or wait a bit after the crash and run the script again
• Updated to use BiGGDB release 1.6
• ModelPolisher now downloads a current version of the identifiers.org registry to validate annotations
• Optimized several bottlenecks, like waiting on database connections and caching results from some queries that only need to be run once instead for every species, reaction, etc.
  ModelPolisher should now run significantly faster compared to former releases
• Changed internal BiGGId handling, so longer and more complex ids should be recognized correctly
• Updated JSONparser to conform with the schema defined by cobrapy
• Validation option now works offline instead of using the online service
• Moving away from purely manual testing for more reliability, however this is an ongoing process
• BiGGAnnotation now has a progress bar
• Added code to convert bounds and objectives from a reactions kinetic law to their FBC representation, if not already set
• Added code to convert gene associations set in reaction or model notes to GeneProductAssociations
• Changed annotations to https://identifiers.org from http

Bugfixes

• Not using BiGGAnnotations or AnnotateDB should not throw NullPointerExceptions anymore
• Fixed several possible NullPointer sources
• Mostly fixed ClassCastException for GPR parsing, might still occur for very large gene reaction rules
• Fixed connections to BiGGDB leaking in a few places
• Changed default ports for BiGGDB and ADB to work correctly
• MatFiles are now closed after reading
• Entities with empty ids are not processed

Caveats

• add-adb-annotations should not be used for now, as some of the annotations added can not be resolved or are wrong. Most likely an issue with the current implementation in ModelPolisher and will be investigated and possibly fixed for the full release
• Some annotations retrieved are not specific for the entity described, e.g. annotations might describe the same entity in different organisms or compartments, not just for the actual model being polished.
  This needs further investigation for the full release