goutamadwant opened a new pull request, #12829: URL: https://github.com/apache/seatunnel/pull/12829
### Purpose of this pull request Closes #12827. Related to #10443. Add a bounded, source-only reader to the existing Couchbase connector. It reads a named bucket, scope, and collection through a read-only SQL++ query and maps JSON fields to an explicitly configured schema. The existing sink is unchanged. The source uses one reader, streams results with backpressure, and cancels active queries during shutdown. It does not partition reads or checkpoint an offset. Restoring a job rereads from the beginning, so downstream writes should tolerate replay. ### Does this PR introduce _any_ user-facing change? Yes. Couchbase can now be used as a bounded batch source. EN/ZH documentation covers configuration, TLS options, single-reader behavior, and replay semantics. No existing sink configuration changes. ### How was this patch tested? - Java 11 connector suite: 61 tests completed with no failures or errors; 3 fixture-gated tests skipped. - Real Couchbase Community 7.1.1 named-collection source-to-sink E2E: 7 jobs passed across Flink 1.13/1.15/1.18/1.20, Zeta, and Spark 2.4/3.3. - Four adjacent sink E2E jobs passed before that separate run was stopped; Spark sink E2E was not rerun. - Connector build, formatting, correctness review, and security review completed. Private-CA TLS and restricted non-admin credentials were not validated against a live server. ### Check list - [x] Added focused tests and EN/ZH documentation. - [x] Validated against a real Couchbase instance and multiple engines. - [x] Left existing sink behavior unchanged. -- 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]
