feat: refactor pillar project calibration loader for use against simplified json

[bencap]

This pull request significantly refactors and expands the load_pp_style_calibration.py script to support batch loading of Zeiberg-style calibrations from an archive directory into Score Sets. The script is now designed for automated ingestion of multiple calibration files, with improved error handling, flexible file mapping, and detailed output statistics.

The most important changes include:

Major script redesign and batch processing:

• The script now loads all calibration JSON files from a specified archive directory, rather than a single file, and applies them to Score Sets based on a dataset mapping file. This enables efficient bulk calibration loading.
• The script handles both regular and ClinVar 2018 variant calibration files, automatically detecting file types and applying appropriate calibration names.

Flexible dataset-to-URN mapping and error handling:

• Introduced support for mapping dataset names to one or more Score Set URNs via a JSON mapping file, including handling of comma-separated URN lists and skipping files with missing or invalid mappings.
• The script now gracefully skips unmapped datasets, missing Score Sets, and invalid JSON files, while reporting detailed statistics on files processed, calibrations created/updated, unmapped files, and missing Score Sets.

Calibration creation logic improvements:

• Functional ranges are now constructed directly from the point_ranges structure in each