Support OpenSearch alias field type

[penghuo]

Description

Summary

This PR introduces support for alias fields in FlintDataType, allowing fields to be referenced by alternative names. The change enables schema parsing and JSON processing to recognize alias fields and map them to their corresponding primary fields.

Changes

1. FlintDataType.scala

   • Updated deserializeJValue to identify and process alias fields.
   • Modified deserializeField to support alias metadata assignment.

2. FlintJacksonParser.scala

   • Implemented fieldMapping to associate JSON keys with schema field indices, allowing alias fields to be correctly resolved during parsing.
   • Updated JSON parsing logic to handle multiple schema fields mapped to a single JSON key.

3. Add OpenSearch Table datatype in docs

   • https://github.com/penghuo/penghuo-opensearch-spark/blob/dataTypeAlias/docs/opensearch-table.md#data-types

Testing

• Added test cases for alias field deserialization to validate schema correctness.
• Integration tests confirm alias field support in OpenSearch queries.

Related Issues

#1033

Check List

• Updated documentation (docs/ppl-lang/README.md)
• Implemented unit tests
• Implemented tests for combination with other commands
• New added source code should include a copyright header
• Commits are signed per the DCO using --signoff

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache 2.0 license.
For more information on following Developer Certificate of Origin and signing off your commits, please check here.

[penghuo]

@ykmr1224 @dai-chen please help review.

[dai-chen]

Thanks for the changes!

[qianheng-aws]

LGTM!

One minor concern from me is that the implementation of this PR is more like the Option3 we mentioned in: opensearch-project/sql#3246 (comment)

So there are some minor behavioral differences between OpenSearch index PPL and Flint PPL. Since we populate the InternalRow with alias columns filled with real data during scanning, the alias columns won't reflect updates made to the original columns in the following execution.

Nevertheless, since Spark doesn't support alias types, implementing like Option1 is challenging in Flint. Besides, it's a very edge case where users would query using alias columns while updating the original columns in this query.

[penghuo]

LGTM!

One minor concern from me is that the implementation of this PR is more like the Option3 we mentioned in: opensearch-project/sql#3246 (comment)

So there are some minor behavioral differences between OpenSearch index PPL and Flint PPL. Since we populate the InternalRow with alias columns filled with real data during scanning, the alias columns won't reflect updates made to the original columns in the following execution.

Nevertheless, since Spark doesn't support alias types, implementing like Option1 is challenging in Flint. Besides, it's a very edge case where users would query using alias columns while updating the original columns in this query.

Yes, it is option3.
This is not an issue for the OS since it operates in near-real-time. Consequently, the latest document updates may not be immediately reflected during a search.