Data lower limb extraction

[Brehana-Naidu]

Pull Request Summary

Closes #219 and #293

This PR cleans up and improves the bone data extraction pipeline by removing the hardcoded pelvis-specific logic and making it flexible enough to work with different anatomical regions. Before this change, the scripts were tightly tied to a single bone set, which made it difficult to reuse or extend. Now, the system is driven by configuration files, so adding support for new datasets like the lower limb, upper limb, spine, or skull doesn’t require changing the core Python code.

To make this work, I introduced JSON-based configuration files and updated the extraction scripts to load bone keywords and fallback names dynamically using a --bone-set-config parameter. I also refactored the core functions to accept these configurable inputs and built a more general pipeline that can handle different datasets. I tested everything by running the pipeline with multiple configurations (like lower limb and spine), and it produced the expected results.

Key Changes
All extraction scripts now accept a --bone-set-config parameter and dynamically load configuration data
Added JSON configuration files for multiple anatomical regions:
bone_set_config.json (Lower Limb: Femur, Patella, Tibia, Fibula, Foot)
upper_limb_config.json (Upper Limb: Humerus, Radius, Ulna, Hand)
spine_config.json (Spine: Atlas, Axis, Cervical, Thoracic, Lumbar, Sacrum)
skull_config.json (Skull: Frontal, Parietal, Temporal, Occipital, Sphenoid)
Refactored core functions:
get_slide_bone_name() now accepts bone_keywords and fallback_name
extract_descriptions_from_slide() uses configurable parameters instead of hardcoded values
Implemented a generic extraction pipeline to handle multiple datasets seamlessly
Fixed Unicode encoding issues that previously caused pipeline failures
Removed hardcoded bone set constants while preserving DataPelvis/ as the output destination
Cleanup
Removed all temporary testing directories (test_, spine_test, etc.)
Retained only essential components:
Configuration files
Core extraction scripts and pipeline
Valid dataset directories (lower_limb/, upper_limb/)
Source data (vertebrae_powerpoint/)
Resulting codebase is clean and production-ready
Testing
Verified individual scripts function correctly with different configurations
Successfully executed the generic pipeline across multiple bone sets
Confirmed correct output labeling (e.g., "atlas" for spine data)

[leandrumartin]

This looks like it resolves only #293, not #219 as you wrote in the summary. Those changes should be set to merge into the data branch, not main.

But when I set this to merge into the data branch, GitHub told me that there were not changes between this branch and data:

20260330-1911-34.9182625.mp4

So it seems to me that you haven't actually added anything new on this branch.

You already have a PR open for adding the lower limb data—#312. So it would probably be simpler to just keep working in that branch.