Refactor raster_clip_weighted to use threading

[krosenfeld-IDM]

This pull request refactors the raster_clip_weighted function in rastertoolkit/raster.py to improve performance and maintainability. The changes include introducing parallel processing using ThreadPoolExecutor, encapsulating shape-specific operations into a helper function, and streamlining the code structure.

Performance improvements:

• Refactored the shape-specific clipping and aggregation logic into a new helper function _clip_weighted_single. This allows for cleaner code and better separation of concerns.
• Introduced parallel processing using ThreadPoolExecutor to process shapes concurrently, leveraging available CPU cores for faster execution.

Code structure and maintainability:

• Removed redundant initialization of the output dictionary and moved dictionary updates into the parallel processing loop.
• Simplified the handling of optional parameters (weight_summary_func) by using default values directly within the helper function.

These updates enhance the readability and scalability of the raster_clip_weighted function while maintaining its core functionality.

[krosenfeld-IDM]

closes #19

[Copilot]

Pull Request Overview

This PR refactors raster_clip_weighted to run per‐shape clipping and aggregation in parallel threads by extracting the core logic into a helper and streamlining how the output dictionary is built.

• Introduces _clip_weighted_single for shape‐level processing
• Uses ThreadPoolExecutor to submit tasks and aggregate results concurrently
• Simplifies default weight_summary_func handling and removes pre‐initialized data_dict

Comments suppressed due to low confidence (2)

rastertoolkit/raster.py:159

• Add a docstring to _clip_weighted_single describing its parameters (shp, k1, n_shapes) and return value to improve readability and maintainability.

def _clip_weighted_single(shp, k1, n_shapes):

rastertoolkit/raster.py:159

• [nitpick] The parameter name k1 is ambiguous; consider renaming it to shape_index or idx to clarify that it represents the current shape’s index.

def _clip_weighted_single(shp, k1, n_shapes):

[Copilot]

Since this workload is CPU‐bound (NumPy and interpolation), you may see better scaling with ProcessPoolExecutor instead of ThreadPoolExecutor to bypass the GIL.