[Core] Hotfix for geometries_tensor_adaptor

[RiccardoRossi]

📝 Description

The manual calculation of Cartesian shape function derivatives (computing local gradients, calculating the Jacobian, inverting it, and performing matrix multiplication) has been replaced by a direct call to the geometry's gradient method. This removes significant boilerplate and leverages the internal optimizations of the Geometry class.

Key Changes

• Removed Manual Jacobian Inversion: Eliminated the manual loop that calculated , , and the subsequent product .
• Direct Gradient Retrieval: Updated CollectShapeFunctionsDerivatives to fetch the required gradients (presumably Cartesian, depending on the Method passed) directly into the DN_Dx container.
• Code Cleanup: Reduced the nesting level and complexity of the Gauss-point loop, making the data mapping to pData more direct and readable.

Impact

• Readability: Significant reduction in lines of code within the core integration loop.
• Performance: Potential minor speedup by avoiding redundant manual matrix operations and utilizing optimized geometry methods.
• Maintainability: Reduces the surface area for bugs related to manual linear algebra implementations.

🆕 Changelog

• Hotfix for geometries_tensor_adaptor

[loumalouomega]

LGTM

[loumalouomega]

📝 Description fixes an error

It could be more specific...