0.3.2

• Allow to specify order in which entity attributes are serialized and deserialized. This is important for object references, because the standard JSON-LD expansion algorithm (used by JB4JSON-LD in deserialization) orders properties in node lexicographically, which is not always suitable.
• Allow an entity to be reconstructed from multiple occurrences throughout the JSON-LD. I.e., when deserialization encounters an object with an already known ID, it will attempt to reopen the instance and add the discovered values to it.
• Automatically generate blank node identifiers for entities without id, so that references to them can be used.