tensorstudio 1.3.1__tar.gz → 1.3.2__tar.gz

{tensorstudio-1.3.1 → tensorstudio-1.3.2}/CHANGELOG.md

@@ -2,6 +2,15 @@
 
 ## Unreleased
 
+## 1.3.2 - 2026-07-07
+
+- Completed the second ordered roadmap item with exhaustive binary
+  elementwise broadcasting tests.
+- Added NumPy comparison coverage for broadcasted `add`, `sub`, `mul`, `div`,
+  equality, inequality, and ordered comparisons.
+- Added incompatible-broadcast error tests and operator-overload broadcasting
+  parity tests.
+
 ## 1.3.1 - 2026-07-07
 
 - Started the ordered roadmap release train with the dtype-promotion contract.

{tensorstudio-1.3.1 → tensorstudio-1.3.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.2
 Name: tensorstudio
-Version: 1.3.1
+Version: 1.3.2
 Summary: TensorStudio is a compact C++ tensor and autograd engine with a Python API for learning, experimentation, and lightweight ML workloads.
 Keywords: tensor,autograd,machine-learning,cpp,pybind11
 Author: TensorStudio contributors
@@ -71,7 +71,7 @@ Description-Content-Type: text/markdown
 TensorStudio is a compact C++ tensor and autograd engine with a Python API for
 learning, experimentation, and lightweight ML workloads.
 
-TensorStudio `1.3.1` is a CPU-only stable API foundation. It is eager-only,
+TensorStudio `1.3.2` is a CPU-only stable API foundation. It is eager-only,
 intentionally small, and not a replacement for mature ML frameworks.
 
 ## Install

{tensorstudio-1.3.1 → tensorstudio-1.3.2}/README.md

@@ -7,7 +7,7 @@
 TensorStudio is a compact C++ tensor and autograd engine with a Python API for
 learning, experimentation, and lightweight ML workloads.
 
-TensorStudio `1.3.1` is a CPU-only stable API foundation. It is eager-only,
+TensorStudio `1.3.2` is a CPU-only stable API foundation. It is eager-only,
 intentionally small, and not a replacement for mature ML frameworks.
 
 ## Install

{tensorstudio-1.3.1 → tensorstudio-1.3.2}/docs/Autograde/index.md

@@ -40,7 +40,7 @@ Calling `backward()`:
 
 ## Supported Gradient Operations
 
-`1.3.1` supports gradients for:
+`1.3.2` supports gradients for:
 
 - `add`
 - `sub`

{tensorstudio-1.3.1 → tensorstudio-1.3.2}/docs/Benchmarks/index.md

@@ -55,7 +55,7 @@ equivalents where available.
 ## Interpreting Results
 
 Expect NumPy and PyTorch to be faster for many medium and large operations
-because they use highly optimized native kernels. TensorStudio `1.3.1` still
+because they use highly optimized native kernels. TensorStudio `1.3.2` still
 favors clarity, portability, and a compact C++ implementation over full kernel
 library performance.
 

{tensorstudio-1.3.1 → tensorstudio-1.3.2}/docs/Hardware/cpu-backend.md

@@ -1,6 +1,6 @@
 # CPU Backend
 
-TensorStudio `1.3.1` supports CPU tensors only.
+TensorStudio `1.3.2` supports CPU tensors only.
 
 ## Device Abstraction
 

{tensorstudio-1.3.1 → tensorstudio-1.3.2}/docs/Neural Networks/training.md

@@ -80,7 +80,7 @@ w = ts.randn((4, 4), seed=123)
 
 ## Practical Tips
 
-- Keep tensors modest in `1.3.1`; kernels are simple CPU loops.
+- Keep tensors modest in `1.3.2`; kernels are simple CPU loops.
 - Prefer `float32` unless you need `float64`.
 - Compare with NumPy in tests for expected numerical values.
 - Watch shape errors carefully around broadcasting and matrix multiplication.

{tensorstudio-1.3.1 → tensorstudio-1.3.2}/docs/Roadmap/index.md

@@ -1,6 +1,6 @@
 # Roadmap
 
-TensorStudio `1.3.1` is a CPU-first tensor, autograd, neural-network, vision,
+TensorStudio `1.3.2` is a CPU-first tensor, autograd, neural-network, vision,
 project, serialization, and ONNX-export foundation. The long-term direction is
 to become a strong compact ML framework for learning, experimentation, and
 lightweight workloads while staying honest about the scale of mature systems
@@ -18,7 +18,9 @@ tensor semantics.
 - Completed in `1.3.1`: define, expose, test, and document dtype promotion
   rules across binary arithmetic, division, matrix multiplication, and
   comparisons.
-- Add exhaustive broadcasting tests for every elementwise operation.
+- Completed in `1.3.2`: add exhaustive broadcasting tests for every binary
+  elementwise arithmetic and comparison operation, including functional and
+  operator overload paths.
 - Add tuple-axis reductions for `sum`, `mean`, `max`, and `min`.
 - Add arg reductions: `argmax` and `argmin`.
 - Add `where`, `maximum`, `minimum`, `clip`, and richer comparison helpers.

{tensorstudio-1.3.1 → tensorstudio-1.3.2}/docs/Usage/interchange.md

@@ -1,6 +1,6 @@
 # ONNX Interchange
 
-TensorStudio `1.3.1` includes a limited ONNX exporter for supported
+TensorStudio `1.3.2` includes a limited ONNX exporter for supported
 TensorStudio module stacks.
 
 Install the optional dependency:

{tensorstudio-1.3.1 → tensorstudio-1.3.2}/docs/Usage/serialization.md

@@ -1,6 +1,6 @@
 # Serialization
 
-TensorStudio `1.3.1` provides two serialization paths: trusted pickle
+TensorStudio `1.3.2` provides two serialization paths: trusted pickle
 roundtrips for internal objects and non-pickle NPZ files for tensor/state_dict
 interchange.
 

{tensorstudio-1.3.1 → tensorstudio-1.3.2}/docs/index.md

@@ -3,7 +3,7 @@
 TensorStudio is a compact C++ tensor and autograd engine with a Python API for
 learning, experimentation, and lightweight ML workloads.
 
-`1.3.1` is a CPU-only stable API foundation for the tensor, autograd,
+`1.3.2` is a CPU-only stable API foundation for the tensor, autograd,
 neural-network, optimizer, data, project, serialization, ONNX export, vision,
 docs, packaging, and wheel workflows.
 

{tensorstudio-1.3.1 → tensorstudio-1.3.2}/docs/publishing.md

@@ -22,8 +22,8 @@ Windows, Linux, and macOS.
 After the release checklist passes, tag the stable release:
 
 ```bash
-git tag v1.3.1
-git push origin v1.3.1
+git tag v1.3.2
+git push origin v1.3.2
 ```
 
 Stable release notes must avoid unsupported performance claims. Publish measured
@@ -63,7 +63,7 @@ clean environment:
 python -m venv .venv-testpypi
 . .venv-testpypi/bin/activate
 python -m pip install -U pip
-python -m pip install --index-url https://test.pypi.org/simple/ --extra-index-url https://pypi.org/simple/ tensorstudio==1.3.1
+python -m pip install --index-url https://test.pypi.org/simple/ --extra-index-url https://pypi.org/simple/ tensorstudio==1.3.2
 python -c "import tensorstudio as ts; import tensorstudio._C; print(ts.__version__)"
 deactivate
 ```
@@ -74,7 +74,7 @@ On Windows PowerShell:
 python -m venv .venv-testpypi
 .\.venv-testpypi\Scripts\Activate.ps1
 python -m pip install -U pip
-python -m pip install --index-url https://test.pypi.org/simple/ --extra-index-url https://pypi.org/simple/ tensorstudio==1.3.1
+python -m pip install --index-url https://test.pypi.org/simple/ --extra-index-url https://pypi.org/simple/ tensorstudio==1.3.2
 python -c "import tensorstudio as ts; import tensorstudio._C; print(ts.__version__)"
 deactivate
 ```

{tensorstudio-1.3.1 → tensorstudio-1.3.2}/include/tensorstudio/version.hpp

@@ -2,6 +2,6 @@
 
 namespace tensorstudio {
 
-inline constexpr const char* version = "1.3.1";
+inline constexpr const char* version = "1.3.2";
 
 }  // namespace tensorstudio

{tensorstudio-1.3.1 → tensorstudio-1.3.2}/pyproject.toml

@@ -7,7 +7,7 @@ build-backend = "scikit_build_core.build"
 
 [project]
 name = "tensorstudio"
-version = "1.3.1"
+version = "1.3.2"
 description = "TensorStudio is a compact C++ tensor and autograd engine with a Python API for learning, experimentation, and lightweight ML workloads."
 readme = "README.md"
 requires-python = ">=3.10"

{tensorstudio-1.3.1 → tensorstudio-1.3.2}/python/tensorstudio/_version.py

@@ -1,5 +1,5 @@
 """Package version."""
 
-__version__ = "1.3.1"
+__version__ = "1.3.2"
 
 __all__ = ["__version__"]

tensorstudio-1.3.2/tests/test_broadcasting.py

@@ -0,0 +1,106 @@
+from __future__ import annotations
+
+from collections.abc import Callable
+
+import numpy as np
+import pytest
+import tensorstudio as ts
+from tensorstudio import ops
+
+TensorBinary = Callable[[ts.Tensor, ts.Tensor], ts.Tensor]
+NumpyBinary = Callable[[np.ndarray, np.ndarray], np.ndarray]
+
+
+def _broadcast_cases() -> list[tuple[np.ndarray, np.ndarray]]:
+    return [
+        (
+            np.array(2.0, dtype=np.float64),
+            np.array([[1.0, -2.0, 4.0]], dtype=np.float64),
+        ),
+        (
+            np.array([1.0, -2.0, 4.0], dtype=np.float64),
+            np.array([[0.5, 2.0, -1.0], [3.0, -4.0, 5.0]], dtype=np.float64),
+        ),
+        (
+            np.array([[1.0, -2.0, 4.0]], dtype=np.float64),
+            np.array([[0.5], [2.0]], dtype=np.float64),
+        ),
+        (
+            np.arange(6, dtype=np.float64).reshape(2, 1, 3) - 2.0,
+            np.arange(4, dtype=np.float64).reshape(1, 4, 1) + 1.0,
+        ),
+    ]
+
+
+@pytest.mark.parametrize(
+    ("name", "tensor_op", "numpy_op"),
+    [
+        ("add", ops.add, np.add),
+        ("sub", ops.sub, np.subtract),
+        ("mul", ops.mul, np.multiply),
+        ("div", ops.div, np.divide),
+        ("equal", ops.equal, np.equal),
+        ("not_equal", ops.not_equal, np.not_equal),
+        ("less", ops.less, np.less),
+        ("less_equal", ops.less_equal, np.less_equal),
+        ("greater", ops.greater, np.greater),
+        ("greater_equal", ops.greater_equal, np.greater_equal),
+    ],
+)
+@pytest.mark.parametrize(("left_np", "right_np"), _broadcast_cases())
+def test_binary_elementwise_broadcasting_matches_numpy(
+    name: str,
+    tensor_op: TensorBinary,
+    numpy_op: NumpyBinary,
+    left_np: np.ndarray,
+    right_np: np.ndarray,
+) -> None:
+    left = ts.from_numpy(left_np)
+    right = ts.from_numpy(right_np)
+
+    actual = tensor_op(left, right).numpy()
+    expected = numpy_op(left_np, right_np)
+
+    if expected.dtype == np.bool_:
+        np.testing.assert_array_equal(actual, expected, err_msg=name)
+    else:
+        np.testing.assert_allclose(actual, expected, rtol=1e-7, atol=1e-7, err_msg=name)
+
+
+@pytest.mark.parametrize(
+    ("name", "tensor_op"),
+    [
+        ("add", ops.add),
+        ("sub", ops.sub),
+        ("mul", ops.mul),
+        ("div", ops.div),
+        ("equal", ops.equal),
+        ("not_equal", ops.not_equal),
+        ("less", ops.less),
+        ("less_equal", ops.less_equal),
+        ("greater", ops.greater),
+        ("greater_equal", ops.greater_equal),
+    ],
+)
+def test_binary_elementwise_broadcast_errors_are_consistent(
+    name: str,
+    tensor_op: TensorBinary,
+) -> None:
+    with pytest.raises(Exception, match="broadcast"):
+        tensor_op(ts.ones((2, 3)), ts.ones((4,)))
+
+
+def test_operator_broadcasting_paths_match_functional_ops() -> None:
+    left = ts.tensor([[1.0, 2.0, 3.0]])
+    right = ts.tensor([[10.0], [20.0]])
+
+    np.testing.assert_allclose((left + right).numpy(), ops.add(left, right).numpy())
+    np.testing.assert_allclose((left - right).numpy(), ops.sub(left, right).numpy())
+    np.testing.assert_allclose((left * right).numpy(), ops.mul(left, right).numpy())
+    np.testing.assert_allclose((left / right).numpy(), ops.div(left, right).numpy())
+    np.testing.assert_array_equal((left == right).numpy(), ops.equal(left, right).numpy())
+    np.testing.assert_array_equal((left != right).numpy(), ops.not_equal(left, right).numpy())
+    np.testing.assert_array_equal((left < right).numpy(), ops.less(left, right).numpy())
+    np.testing.assert_array_equal((left <= right).numpy(), ops.less_equal(left, right).numpy())
+    np.testing.assert_array_equal((left > right).numpy(), ops.greater(left, right).numpy())
+    np.testing.assert_array_equal((left >= right).numpy(), ops.greater_equal(left, right).numpy())

{tensorstudio-1.3.1 → tensorstudio-1.3.2}/tests/test_import.py

@@ -4,7 +4,7 @@ from __future__ import annotations
 def test_import_and_version() -> None:
     import tensorstudio as ts
 
-    assert ts.__version__ == "1.3.1"
+    assert ts.__version__ == "1.3.2"
     assert hasattr(ts, "Tensor")
     assert hasattr(ts, "data")
     assert hasattr(ts, "nn")