JesusMan0529 opened a new pull request, #12611: URL: https://github.com/apache/seatunnel/pull/12611
### Purpose of this pull request The RAG guide examples set `chunk = "chunk_vector"` and `description = "description_vector"`, but Embedding expects `output_vector_field = input_field`. These configurations fail with `TRANSFORM_COMMON-02` because the vector fields do not exist in the upstream schema. Correct the two mappings in both the English and Chinese guides. The change is limited to four configuration lines. ### Does this PR introduce any user-facing change? Yes. The examples now read the existing `chunk` / `description` fields and produce `chunk_vector` / `description_vector`. ### How was this patch tested? - Parsed all four guide mapping blocks with the project's `ConfigFactory` / `ReadonlyConfig` and initialized the actual `EmbeddingTransform` against the documented input columns: all four failed before the correction; all four passed afterward with the expected output field names. No model or vector database was contacted. - `mvnw.cmd -pl seatunnel-transforms-v2 -am -Dtest=EmbeddingTransformFactoryTest -Dsurefire.failIfNoSpecifiedTests=false -Dskip.spotless=true verify`: the 11-module reactor passed, including `EmbeddingTransformFactoryTest` (1 test, no failures/errors/skips). - Root `mvnw.cmd spotless:apply` and `git diff --check` passed. `mvnw.cmd -q -DskipTests verify` was attempted twice. The first attempt stopped on Maven Central TLS download failures for PostGIS dependencies in `connector-jdbc`; after fetching those official artifacts with matching checksums, the retry advanced to `connector-redis` and stopped on the same TLS failure for `commons-pool2:2.11.1` / `json:20211205`. The complete root build did not finish. Live model/vector-store pipelines and Docker integration tests were not run. ### Check list Not applicable to this four-line documentation correction. -- 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]
