Changelog for version 0.0.3-alpha.1 (05/25/2014)
- Updated example walkthrough and
- Added python utility
ddlibfor text manipulation (need exporting PYTHONPATH, see its pydoc for usage)
- Added utility script
util/extractor_input_writer.pyto sample extractor inputs
sentence_offset, textual sentence_id)
- Cleaned up unused datastore code
- Update templates
- Bug fixes
Changelog for version 0.0.3-alpha (05/07/2014)
Non-backward-compatible syntax change: Developers must include
idcolumn with type
bigintin any table containing variables, but they MUST NOT use this column anywhere. This column is reserved for learning and inference, and all values will be erased and reassigned during the grounding phase.
Updated dependency requirement: requires JDK 7 or higher.
Supported four new types of extractors. See documentation for details:
Even faster factor graph grounding and serialization using better optimized SQL.
The previous default Java sampler is no longer supported. The C++ sampler is now the default sampler.
New configuration tunable supported:
pipeline.relearn_fromto skip extraction and grounding and only perform learning and inference with a previous version of the grounded graph. Useful for tuning sampler arguments.
Supported custom holdout by a holdout query.
spouse_examplewith implementations of different styles of extractors.
nlp_extractorexample has different table requirements and usage. See here: NLP extractor.
db.defaultconfiguration, users should define
user. If not defined, by default system will use the environmental variables
Fixed all examples.
Print SQL query execution plans for extractor inputs.
Skip grounding, learning and inference if no factors are active.
Greenplum users should add
DISTRIBUTED BYclause in all
CREATE TABLEcommands. Do not use variable id as distribution key. Do not use distribution key that is not initially assigned.