[FLINK-40223][runtime] Support additional Flink SQL string functions in YAML Transform

[haruki-830]

Summary

This commit expands the string function coverage of Flink CDC YAML Transform by adding commonly used Flink SQL string manipulation, search, predicate, and encoding functions. These functions can now be used in row-level transform projections and filters.

Key Changes

1. Additional String Function Support

• Added LTRIM, RTRIM, BTRIM, OVERLAY, POSITION, LOCATE, and INSTR.
• Added CONCAT_WS, LPAD, RPAD, REPLACE, REPEAT, LEFT, and RIGHT.
• Added STARTSWITH, ENDSWITH, TO_BASE64, and FROM_BASE64.
• Handles null inputs, Unicode code points, and function-specific boundary cases consistently with Flink SQL semantics.

2. Parser and Runtime Integration

• Registered the new functions with their operand and return type definitions in the transform operator table.
• Added parser handling for SQL-specific OVERLAY and POSITION syntax.
• Integrated the functions with Janino expression compilation for runtime evaluation.

3. Test Coverage and Documentation

• Added unit tests for function behavior, null handling, Unicode strings, and boundary cases.
• Added parser and transform runtime tests for valid and invalid expressions.
• Added YAML specification coverage for projection results, including matched, unmatched, and null inputs.
• Documented the new functions in both the English and Chinese transform documentation.

JIRA Reference

https://issues.apache.org/jira/browse/FLINK-40223

[haruki-830]

@lvyanquan @yuxiqian PTAL, thanks!

[yuxiqian]

Thanks Haruki for the nice work, just left some minor comments.

[haruki-830]

@yuxiqian Thanks for the review! I addressed the comments in the lastest commit.

[yuxiqian]

+1 from my side (except one minor comment on test specs). Need another eye from Kunni.