Tutorials
Every page in this section is generated from a runnable script in scripts/ by Literate.jl, so each one can be executed, edited and re-run locally:
bash
julia scripts/03_walpole.jlThe generated pages come with a pre-run Jupyter notebook and a cleaned standalone script under docs/generated_notebooks/ and docs/generated_scripts/.
Simplest first within each group.
Tensor algebra
| Page | What it shows |
|---|---|
| Bases, variance and the metric | an oblique basis, dual basis and metric, covariant vs contravariant components, and where variance stops mattering |
| Tensor products, contractions and their identities | the ⊠/⊠ˢ algebra checked numerically, the identity that fails, isotropic projection, and why it does not commute with inversion |
| The Walpole basis end to end | the six |
| The eight-dimensional axially invariant space | |
| Projection and symmetry detection | fixed axis vs optimized, determinism, the detection cascade, and the cost of each path |
Differential calculus
| Page | What it shows |
|---|---|
| Differential operators, symbolically | the five operators in polar, cylindrical and spherical coordinates; the hoop strain that comes from the connection alone; the index-placement convention |
| Christoffel symbols | the definition checked term by term, every predefined system tabulated, symbolic and numerical routes compared, spheroidal harmonics |
| Differential operators, numerically | pointwise evaluation by AD, the Lamé hollow-sphere problem with its equilibrium residual, and sensitivities through the operators |
| Surfaces: fundamental forms and curvature | sphere, cylinder, plane and paraboloid; Gauss's Theorema Egregium; the array-comparison trap |
| Classical identities of the differential operators | every identity checked, not quoted, on four charts — including a deliberately non-orthogonal one, the only case that tests the covariant/contravariant bookkeeping |
Applications in mechanics
| Page | What it shows |
|---|---|
| Elastic Green's functions | the Kelvin solution in 2-D and 3-D, |
| A two-point Green tensor between offset frames | combining tensors that live in two different spherical charts |
| The elastic sphere, solved symbolically | hydrostatic and deviatoric problems, the Lamé exponents, and how they assemble into an |
| Acoustic and Hill tensors |
Types, differentiation and performance
| Page | What it shows |
|---|---|
| Differentiation and interoperability | ForwardDiff through the algebra and through a projection, symbolic vs automatic derivatives, Tensors.jl interop |
| Performance of the structured types | measured speed-ups of the closed forms over the dense route, allocations, and where the structure stops paying |
Where to start
If you are new to the library, read Getting started first, then Bases, variance and the metric — variance is the one notion a Cartesian-only tensor library does not have, and everything else rests on it.