Reshape

Version: v0.2.0

Status: Complete

Header: include/stratax/core/algorithms/Reshape.hpp

Overview

Reshape.hpp provides shape-changing copy helpers that preserve flat storage order.

reshape returns a tensor with a requested shape, while flatten returns a rank-1 vector.

Responsibilities

The reshape module is responsible for:

Validating element-count compatibility for reshape operations
Copying values in flat storage order
Producing owning containers for reshaped/flattened output

The reshape module is not responsible for:

View-based reshape semantics
Inferred-dimension syntax (-1 style placeholders)
In-place shape mutation

Relationships

reshape(arr, shape)
├── validation::require_equal_size(arr.size(), shape.elements(), ...)
└── flat copy into Tensor
 
flatten(arr)
└── flat copy into Vector

Depends on:

Invariants

The following conditions are always true:

Reshape and flatten preserve flat element order.
reshape returns Tensor<value_type>.
flatten returns Vector<value_type>.
Output size equals source size.
Both operations return owning containers.

Public Interface

reshape

template<Array A>
stratax::container::Tensor<typename A::value_type>
reshape(const A& arr, const stratax::core::Shape& shape);

Throws

Exceptions::ShapeError when target shape element count differs from source size
Possible propagated shape/overflow exceptions from shape.elements()

Complexity

O(n + r)

flatten

template<Array A>
stratax::container::Vector<typename A::value_type>
flatten(const A& arr);

Complexity

O(n)

Complexity Summary

Operation	Complexity
reshape	O(n + r)
flatten	O(n)

n is element count and r is rank.

Examples

const auto t = reshape(vec, stratax::core::Shape(2, 3));

const auto v = flatten(t);

Design Notes

The module keeps behavior explicit and safe: reshape validates total size first and always performs a full copy.

Future Improvements

Add non-owning view reshape when view types exist
Add optional strict reshaping helpers with additional layout checks