@mutable.Vector

API baseline for @mutable.Vector in the current 0.4.7 repository state.

Overview

• @mutable.Vector is the repository's mutation-oriented vector type.
• Storage is a wrapped Array[T], so indexed writes update the underlying mutable buffer directly.
• Even in this package, many algebraic helpers still return fresh vectors so callers can choose between mutation and value-producing transforms.

Core API

• Vector::from_array(arr) / Vector::make(n, elem) / Vector::makei(n, f) Construct vectors from existing data, a repeated value, or an index function.
• length() Returns the vector length.
• v[i] / v[i] = x Read and write one element. Bounds follow Array[T] behavior.
• copy() Returns a deep copy of the vector.
• iter() Exposes an iterator over the current elements.

Value-Producing Helpers

• map(f) / zip_with(other, f) Return a transformed vector without mutating self.
• add_constant(cst) Adds the same scalar to each element.
• left_scale(scalar) / right_scale(scalar) Return scaled vectors.
• lerp(other, alpha) Computes (1 - alpha) * self + alpha * other.
• +, *, unary - Element-wise addition, Hadamard multiplication, and negation.

In-Place Helpers

• map_inplace(f) Rewrites every element in place.
• left_scale_inplace(scalar) / right_scale_inplace(scalar) Apply scalar multiplication in place.

Scalar And Matrix Helpers

• dot(other) Computes the dot product. Length mismatch aborts.
• lin_comb(weights, vectors) Top-level helper that builds one vector from weighted input vectors. Empty inputs, mismatched counts, or mismatched vector lengths abort.
• to_col_matrix() / to_row_matrix() Convert the vector into matrix form.
• scaled_matrix() Builds a diagonal matrix with the vector on the main diagonal.
• tensor_product(other) Computes the outer product and returns a matrix.

Guidance

• Use direct indexing or *_inplace helpers when the workload is truly mutation-heavy.
• Use the non-inplace helpers when you need a fresh vector or want code that mirrors @immut.Vector more closely.
• backends/default.DenseVector is a wrapper around this concrete implementation. If you want the trait-oriented default backend entry point, see the backends/default API.