What's new?
- Enhance Ursa Compass with the ability to perform data- and logic-provenance exercises, tracing back values up through the dependency tree to explain results.
What improvements have been made to existing features?
- Added optional performance profiling, driven by getrusage(), during HCC model runs, by setting the environment variable HCC_PERFORMANCE_DEBUG_MODE to 1.
- Preprocessing scripts will use a deterministic decolliding token (the object ID) instead of a random string when creating the filename for the write-back into the bucket at the end of the preprocessing step. This will improve the reliability of the Prevent Duplicate Imports feature.
What's been fixed?
- Allow the Application Analytics report to report on tables whose row count exceeds the maximum PostgreSQL int value of ~2B.
- Fix the search function from the Ursa Studio help drawer into our knowledge base per updated Document360 querystring conventions.