zoomy_core.model.derivation.basis_cache

`zoomy_core.model.derivation.basis_cache`#

Manifest-based on-disk cache for basis-function matrices, plus an in-memory wrapper. Sits under the moment/Galerkin projection: consumers call get_basis_matrices instead of building a SymbolicIntegrator.

Module Contents#

Classes#

BasisMatrixCache

Hash-keyed .npz cache of the basis matrices, indexed by a JSON manifest. Stale entries are detected by comparing the stored basis fingerprint.

Functions#

`get_basis_matrices`	Return {phib, M, A, B, D, Dxi, Dxi2, DD, D1, DT} for (basis, level), memoized in memory and on disk.
`clear_basis_matrix_cache`	Drop the in-memory cache and optionally the on-disk cache.

Data#

`CACHE_ROOT`
`MANIFEST_NAME`
`MATRIX_NAMES`

API#

zoomy_core.model.derivation.basis_cache.CACHE_ROOT#: ‘.cache/basismatrices’

zoomy_core.model.derivation.basis_cache.MANIFEST_NAME#: ‘manifest.json’

zoomy_core.model.derivation.basis_cache.MATRIX_NAMES#: [‘phib’, ‘M’, ‘A’, ‘B’, ‘D’, ‘Dxi’, ‘Dxi2’, ‘DD’, ‘D1’, ‘DT’]

class zoomy_core.model.derivation.basis_cache.BasisMatrixCache(cache_root=CACHE_ROOT)#

Hash-keyed .npz cache of the basis matrices, indexed by a JSON manifest. Stale entries are detected by comparing the stored basis fingerprint.

Initialization

Set the cache directory, relative to the project root.

get_or_compute(basis, level, compute_fn)#

Return the cached matrices of a basis or compute and store them.

Args

basis: Basisfunction instance.
level: Polynomial level.
compute_fn: Called as compute_fn(level), returns {name: ndarray}.

Returns

Dict {name: ndarray} of basis matrices.

list_entries()#: Return the manifest dict.

clear()#: Delete the whole on-disk cache.

zoomy_core.model.derivation.basis_cache.get_basis_matrices(basis, level=None, *, cache=None)#

Return {phib, M, A, B, D, Dxi, Dxi2, DD, D1, DT} for (basis, level), memoized in memory and on disk.