Avoid capturing default config in numeric parser UDF

[qwerty44-U]

Summary

• avoid capturing the full StandardizationConfig in the numeric parser UDF closure
• pass the narrower ErrorCodesConfig through UDFResult.fromTry, while preserving the existing StandardizationConfig overload
• add a serialization regression covering DefaultStandardizationConfig

Release Notes

• Fixed numeric parser UDF serialization when using the default standardization config.

Verification

• git diff --check

Scala/Spark tests were not run locally because this environment does not have Java, sbt, or Maven installed (java, sbt, and mvn are not available on PATH).

Fixes #57

[MatloaItumeleng]

Hi @qwerty44-U Thank you for the contribution , seems there's compile errors could you please look at them, I've pulled it and the compile error UDFResult.scala:28:8: in object UDFResult, multiple overloaded alternatives of method fromTry define default arguments. is due to this line and the one above L40-L41 the two alternatives of fromTry declare a default on defaultValue causing a clash

[dk1844]

I like the general idea, but please improve

[dk1844]

I agree with simplifying the UDF, but I would just pass the stdConfig.errorCodes

Suggested change

	vErrorCodes,
	stdConfig.errorCodes,

[dk1844]

This change makes sense, just the ErrorCodesConfig type is enough here.

[dk1844]

I really do not see the point of this change with recreating the error basic config -- seems to me that we end up with the very same result vs using stdConfig.errorCodes directly

[qwerty44-U]

Thanks for the review. I pushed updates addressing the feedback.

What changed:

• Kept numericParserToTyped narrowed to ErrorCodesConfig.
• Removed the BasicErrorCodesConfig(...) reconstruction.
• Capture stdConfig.errorCodes once as vErrorCodes before building the UDF, so the UDF does not capture the full StandardizationConfig while still using the configured error codes directly.
• Fixed the overloaded UDFResult.fromTry default-argument compile issue.
• Added coverage for both ErrorCodesConfig and StandardizationConfig UDFResult.fromTry paths, plus executing the numeric UDF through Spark in the default-config regression.

Verification from the latest run on commit 359010a:

• Scala 2.12.18: passed
• Scala 2.13.11: passed
• test (2.12.20, 2.12, 80, 80, 60): passed
• license-test: passed
• check-release-notes: passed

The remaining Scalafmt Check failure is still the checkout issue: the workflow tries to checkout fix-date-udf-default-config-serialization from AbsaOSS/spark-data-standardization, but that branch exists on the fork, so it exits before running Scalafmt.

[MatloaItumeleng]

Looks good , thank you