Optimize Node Processing in ROM Basis Setting

[SADPR]

Description

This PR introduces performance optimizations to the node processing logic in rom_analysis.py, specifically for setting the Reduced Order Model (ROM) basis. The original implementation, which iterated over all nodes using np.where within a loop, was suboptimal for large datasets.

Changes

File Modified

• rom_analysis.py

Key Modifications

• Implemented a preprocessing step to create a dictionary (node_id_to_index), mapping each node ID to its index in the node_ids array. This optimizes the retrieval of node indices during the main processing loop.
• Updated the loop over nodes to leverage the node_id_to_index mapping for efficient index lookup, thereby reducing computational overhead.

Benefits

• Performance Improvement: Significant reduction in computation time, especially beneficial for large models with numerous nodes.

Additional Notes

• The changes maintain the existing functionality while greatly enhancing performance, particularly in large-scale applications.
• Care was taken to ensure compatibility with the rest of the rom_analysis.py module.

[Rbravo555]

Fair enough. You use a python dictionary as a mapping, instead of using np.where