[PLUGIN-1957] Validate PK Chunking for incremental loads

[harishhk107]

PLUGIN-1957 Autodetect PK Chunking for incremental loads

What

Adds an autonomous pre-flight check to decide whether to enable PK Chunking in SalesforceBatchSource.getSplits().

• If the estimated record count is below the user-configured chunk size (config.getChunkSize()) and the query is selective, PK chunking is skipped and executed as a standard un-chunked query to avoid "empty chunk" overhead on small datasets.
• If the dataset volume exceeds the threshold, or is flagged as non-selective, PK Chunking is enabled.
• This feature is specifically designed for automated Data Transfer Service.

Why

PK chunking is designed for very large datasets. Enabling it on small datasets (like highly restrictive incremental intervals where 0 or very few records changed) causes dozens of empty chunk boundaries to be spooled and executed, wasting daily Bulk API allocations and adding massive pipeline overhead. This change ensures chunking is only applied when operationally justified by the actual matching record volume.

Changes

• SalesforceBatchSource.java: Configured PK chunking threshold dynamically using chunkSize.
• SalesforceSplitUtil.java: Added REST Query Plan explain check and removed unused constants.
• SalesforceQueryUtil.java: Replaced magic HTTP status codes with constants.
• SalesforceSourceConstants.java: Cleaned up the unused threshold constant.
• SalesforceQueryUtilTest.java: Added unit tests covering Query Plan and fallback paths.
• SalesforceBatchSourceETLTest.java: Renamed ETL test case to remove UI naming.

---

[google-cla]

Thanks for your pull request! It looks like this may be your first contribution to a Google open source project. Before we can look at your pull request, you'll need to sign a Contributor License Agreement (CLA).

View this failed invocation of the CLA check for more information.

For the most up to date status, view the checks section at the bottom of the pull request.

[Sunish-Dahiya]

On a high level change looks fine but can you please test all these scenario

1. When chunking is enabled for data less than <100K(old logic)
2. Chunking should be disabled for data less than <100K - compare time/performance of #1 and #2.
3. All other cases , adding few of them for reference: query is selective, query is expensive, query is selective but record count >100K

[Sunish-Dahiya]

Comment regarding unit tests is still not addressed

[Sunish-Dahiya]

Please address the comments before merging.

[harishhk107]

Add all the testcases in the SalesforceQueryUtilTest to reuse the existing PowerMockito class-level configuration for mocking query plan REST calls and

[Sunish-Dahiya]

this can be inlined, we don't need a new variable