Your first calculation¶

This tutorial constructs a matrix, multiplies it, and reads the result. Complete installation first.

Add two scalars¶

from hmath import Tensor, add

assert add(2, 3) == 5.0

The Python interface calculates with IEEE 754 binary64 values, corresponding to C double. Results are Python floats.

Construct and multiply a matrix¶

HMath stores matrix values in row-major order. The list below represents [[1, 2], [3, 4]].

from hmath import Tensor

with Tensor.matrix([1, 2, 3, 4], 2, 2) as matrix:
    assert matrix.shape == (2, 2)
    assert matrix.sum() == 10.0
    with matrix @ matrix as product:
        assert product.shape == (2, 2)
        assert product.values() == [7.0, 10.0, 15.0, 22.0]

The product follows ordinary matrix multiplication:

[1 2] [1 2] = [ 7 10]
[3 4] [3 4]   [15 22]

Each with block releases its tensor when the block ends. Operators create new tensors; they do not modify their operands.

Handle invalid input¶

The number of values must equal the product of the dimensions:

from hmath import HMathError, Tensor

try:
    Tensor([1.0], shape=(2, 2))
except HMathError as error:
    assert error.status == 2  # HMATH_SHAPE_ERROR
else:
    raise AssertionError("An invalid shape was accepted")

Continue with NumPy interoperability, or consult the complete Python reference.