JingsongLi opened a new pull request, #8659:
URL: https://github.com/apache/paimon/pull/8659

   ## Summary
   
   Expose primary-key full-text search through Spark SQL, a Flink procedure, 
and the Java search path. Hybrid search can now fuse primary-key Vector and 
Full Text results against one snapshot while preserving physical-position 
semantics.
   
   Follow-up to #8652.
   
   ## Changes
   
   - Pin primary-key Hybrid routes to one snapshot, fuse physical results with 
`rrf`, `weighted_score`, or `mrr`, and reject mixed physical/global address 
spaces.
   - Add the Flink `full_text_search` procedure with projection, score 
exposure, dynamic options, validation, and deterministic result ordering.
   - Add Spark and Flink end-to-end coverage using the fixed `full-text` 
provider identifier.
   - Document primary-key Full Text configuration, compaction-visible coverage, 
native scores, Hybrid search, and current FAST-only behavior.
   
   ## Testing
   
   - [x] `mvn -pl paimon-core -am -DfailIfNoTests=false -DwildcardSuites=none 
-Dtest=PrimaryKeyHybridSearchTest,PrimaryKeyFullTextScanTest,PrimaryKeyFullTextSearchTest
 test`
   - [x] `mvn -pl paimon-flink/paimon-flink-common -am -Pflink1 
-DfailIfNoTests=false -DwildcardSuites=none 
-Dtest=FullTextSearchProcedureITCase test`
   - [x] Flink 2 on JDK 11 with `FullTextSearchProcedureITCase`
   - [x] `mvn -pl paimon-spark/paimon-spark-ut -am -Pspark3 
-DfailIfNoTests=false 
-DwildcardSuites=org.apache.paimon.spark.sql.FullTextSearchTest,org.apache.paimon.spark.sql.HybridSearchTest
 -Dtest=none test`
   - [x] Spark 4 clean package on JDK 17 with `FullTextSearchTest` and 
`HybridSearchTest` (13 tests)
   
   ## Notes
   
   Primary-key Full Text remains FAST-only: uncovered files are ignored until 
compaction publishes eligible data files and persistent archives. This PR does 
not add Python APIs or an `index.type` selector for primary-key Full Text.
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]

Reply via email to