Changelog for version 0.02-alpha (03/12/2014)

  • Non-backward-compatible syntax change: Extractors only use the column name instead of \[relation_name\].\[column_name\] for keys to the JSON object. Documentation has been updated accordingly.
  • Non-backward-compatible syntax change: You must explicitly alias id and variable columns in the inference rules. See the updated inference rule documentation for more information.
  • Faster factor graph grounding code using pure SQL
  • Support for high-speed C++ sampler
  • Added documentation for high-speed sampler integration
  • Faster serialization of factor graph
  • Refactored the database-connection code to be JDBC-compatible. This makes it easier to add support for different databases.
  • Automatically assign sequential variable IDs to extractions.