Zrp

**Updates & Enhancements:**

- Optimized Memory Usage: We've made significant modifications to reduce the memory footprint, ensuring smoother and more efficient performance.
- Custom Modeling Flexibility: You can now provide custom modeling parameters for XGBoost model builds, giving you greater control and flexibility in your modeling process.
- Real-Time Progress Updates: Live model building progress messages

**Updates & Enhancements:**

- Updated how `ZGeo` was processing missing values for GEOID_ZIP
- Updated joblib package version

**Updates & Enhancements:**
• Improved accuracy & expanded coverage of the geocoder
• Adapted `ZRP` to run on smaller machines by adding “chunksize” parameter
• Updated handling of non-numeric house numbers
• Updated naming of files saved to the artifacts folder

**Updates & Enhancements:**
• Improved accuracy & expanded coverage of the geocoder
• Added functionalities allowing for model training based on custom target classes
• Extended the set of standard target classes
• Improved handling of user-defined block group and census tract GEOIDs
• Added `GeoLookUpLooper` class which allows for easier building of geolookup tables
• Expanded the `FEtoPredict` class to support non-standard target classes

**Updates & Enhancements**:

- Improved & reduced memory usage
- Updated documentation including more clear docstrings , error messages, and progress messages
- Updated validators by reducing extensive checks, correcting null calculations, and including more flexibility
- Improved & reduced `ZRP` runtime
- Updated documentation including more clear docstrings , error messages, and progress messages
- Provide more robust errors
- Extend race & ethnicity proxying options. Now ZRP name + geo, BISG name + geo, name-only, and geo-only proxies can be generated using `ZRP` and its subclasses
- Updated BISG generation to normalize the data
- Include more robust support file checks to provide early indicator if supporting files have not been downloaded from GitHub
- Additional pre-proxy generation data processing to support customization