We are pleased to announce new version 1.2.0. This is probably one of the most significant OpenCGA release in terms of new features and improvements. Here you can find an overview of the most notable improvements.
General
QA and testing: We have now more than 1,000 acceptance tests with Fitnesse! This release has added about 200 new tests.
Catalog
Allow to define data releases: A new release field has been added to most data models. This allows admin to define and create internal releases automatically, Catalog and Variant data can now be associated with a particular release.
JSON Web Tokens (JWT): Session tokens have been migrated to JWT (https://jwt.io/) standard. JWT allows to implement federated systems by authenticating users across different applications. JWT session tokens are not stored in Catalog any more, this has reduced the number queries needed during login. Tokens are now sent in the HTTP header instead of old sid query parameter, note that sid query parameter will still work during next releases.
Overhaul of ACLs physical implementation: ACLs are now checked as part of the query to the database, improving the performance and reducing the number of calls made.
Multi-group support: Users can now belong to more than one group. Permissions defined in different groups of one user are resolved as the union of them.
Improve synchronisation from LDAP: If an LDAP user logs in and it was not registered in Catalog, it will be automatically imported. Besides, synchronised groups with LDAP where the user belongs or belonged to will be resynced.
Private Variable Sets and Annotations: Variable sets can now be defined as private. In such a case, only users with the new proper permission will be able to see or edit those private annotations.
Propagate permission between Samples and Individuals: A new propagate parameter has been added to both sample and individual acl webservices to also propagate permissions to the other related entries if set.
Data model fixes: Some data models such as file contained an array of sample ids instead of an array of samples. This kind of change has been applied to all missing data models to satisfy that all contain a list of other data models instead of ids.
New members group: A new internal members group has been added. This group will keep track of all the users with some kind of permission in the study. Plus, actions such as completely remove access to one user to anywhere in the study is now as easy as removing that user from the members group.
New REST web services
As a result of some of these changes Catalog performance has been improved significantly and the number of queries executed have dropped to 50%.
