Refactor: Optimized rule applies to class functionality for large datasets

[SFJohnson24]

• added function get_contents_metadata when rule_applies_to_class is called to prevent whole dataset from being read versus just metadata
• tested validate with new functionality on SDTMIG 3.4
• implemented new unittest for added functionality

[SFJohnson24]

validate with update.xlsx validate without update.xlsx I performed a validate with both the current changes and pre-changes and the reports are identical

[SFJohnson24]

Why can't we just add something like original_path as another dataset property here?

https://github.com/cdisc-org/cdisc-rules-engine/blob/d364ac102b8365dffde5c1551aedacdbd66cc89f/scripts/run_validation.py#L147-L150

this worked with a bit of decency injection to get it datasets to the read_metadata call. I refactored to reflect this change and ran test validations which appear to be working.

[SFJohnson24]

Still failing. New error. Here:

https://github.com/cdisc-org/cdisc-rules-engine/blob/90ea95d26caf2fcfee58ea9b6142e62b9d1cb106/cdisc_rules_engine/services/data_services/local_data_service.py#L187

Instead of file_path, I think it needs to be file_metadata["path"]. file_path is still pointing to the parquet path without metadata. file_metadata["path"] has the xpt path

I looked to change this but it was extracting metadata properly without the change for some reason (I am using JSON) despite the file_path. pointing to the parquet. First image is original, second is with file_metadata["path"] Regardless, I changed it.