• Improve automated downloads for distributions using MLDS service - 'Member Licensing and Distribution Service'
• Add GitHub Actions to test MLDS downloads for International, and Argentinian distributions and exclude some UK specific tests (localisation)

• Add support for 'null' values in reference set fields (custom fields used by some distributions)

• More lenient import of null/empty string reference set fields
• Improved logging to help with automated pipelines and distribution files
• Improve search for empty search strings

• Fix #70 with handling unicode characters in ECL parsing
• Fix #72 by forcing ordered choice to reduce ambiguity in parsing member filters
• Fix #73 by updating ECL parser with newer 2.2 updates
• Fix #75 with handling of unicode characters and normalization (folding) for search within ECL description term filter
• Upgrade dependencies including lmdbjava 0.9.1

• Fix #69 by improving import of custom reference sets in which there are missing custom field values
• Improve build process for uberjar to ensure backwards compatibility with JDK>=11
• Update to latest trud library as the NHS TRUD service now supports SHA256 for integrity checks on downloaded files
• Add intersect-ecl-fn to improve performance in data pipelines when an ECL expression may be re-used many times

Updated to be compatible with JDK11 as v1.4.1424 was compiled with JDK21 and therefore required >=JDK21. See #69

• Improvements to ranked-search
• Add support for Lucene 10 for testing and evaluation via a runtime switch (usable when run from source code, not from the released uberjar - a Lucene 10/Java>=21 version can be built or released on request).

• Add ranked search
• Fix #68 and permit both additional custom query AND a constraint
• Upgrade dependencies

• Add support for ECL v2.2, including the operators top and bottom

Full Changelog: v1.4.1388...v1.4.1400