[BUG] Preserve dots in table name when generate index name

[dai-chen]

What is the bug?

Currently Flint replaces all dots in qualified table name with underscore. However, this doesn't align with PPL.

Flint:

https://github.com/opensearch-project/opensearch-spark/blob/main/flint-spark-integration/src/main/scala/org/opensearch/flint/spark/FlintSparkIndex.scala#L91

PPL:

https://github.com/opensearch-project/sql/blob/main/spark/src/test/java/org/opensearch/sql/spark/flint/IndexQueryDetailsTest.java#L67

What is the expected behavior?

Modify the generate index name logic and align with PPL behavior.

[Swiddis]

Also seems to be behind a bug in integrations, we haven't accounted for the underscore replacement which causes issues in finding materialized views for table names with dots.

[dai-chen]

Also seems to be behind a bug in integrations, we haven't accounted for the underscore replacement which causes issues in finding materialized views for table names with dots.

Here is the naming convention for your reference: https://github.com/opensearch-project/opensearch-spark/blob/main/docs/index.md#opensearch