Crate matten_ndarray 

matten-ndarray — a conversion bridge between matten::Tensor and ndarray::ArrayD<f64>.

This is a deliberately boring companion crate (RFC-025, RFC-027): it converts between the two owned representations and does nothing else. It adds no dependency to core matten, wraps none of the ndarray API, and exposes no view or lifetime types.

use matten::Tensor;
use matten_ndarray::{from_arrayd, to_arrayd};

let t = Tensor::new(vec![1.0, 2.0, 3.0, 4.0], &[2, 2]);
let arr = to_arrayd(&t).unwrap();
let back = from_arrayd(arr).unwrap();
assert_eq!(back.as_slice(), t.as_slice());

Status

Production-ready candidate. The API is stable for the 0.19 family. Supported ndarray: the 0.16 minor.

Behavior

• Both conversions copy; no zero-copy is claimed.
• from_arrayd preserves logical element order even for non-standard (transposed / sliced) ArrayD inputs.
• from_arrayd rejects shapes with a zero-length axis (matten forbids zero-sized dimensions).
• A dynamic tensor passed to to_arrayd always returns MattenNdarrayError::DynamicTensor — this guard is unconditional and does not depend on the companion dynamic feature (RFC-031).

Feature flags

• dynamic — Compatibility forwarding feature. No longer required for dynamic rejection as of v0.19.1. Dynamic tensors are rejected at companion boundaries regardless of whether this feature is enabled. Reconsider removal no earlier than v0.20.0.

Enums

MattenNdarrayError

Errors produced by the matten ↔ ndarray conversions.

Functions

from_arrayd

Converts an ndarray::ArrayD<f64> into a Tensor.

to_arrayd

Converts a numeric Tensor into an ndarray::ArrayD<f64>.