astx.tensor

`astx.tensor`

Source: packages/astx/src/astx/tensor.py

title: ASTx Tensor AST nodes.
summary: >-
  Provide internal nodes for the Arrow C++ backed tensor runtime, aligned with
  Apache Arrow's homogeneous N-dimensional tensor model.

Classes

`TensorType(AnyType)`

title: Internal Tensor semantic type.
summary: >-
  Represent the homogeneous N-dimensional tensor abstraction while reusing
  the canonical buffer/view representation during lowering.
attributes:
  element_type:
    type: astx.DataType | None
  shape:
    type: tuple[int, Ellipsis] | None

`TensorLiteral(astx.base.DataType)`

title: Internal Arrow C++ backed Tensor literal node.
summary: >-
  Build one Arrow C++ tensor value from scalar values plus shape and stride
  metadata.
attributes:
  values:
    type: list[astx.AST]
  element_type:
    type: astx.DataType
  shape:
    type: tuple[int, Ellipsis]
  strides:
    type: tuple[int, Ellipsis] | None
  offset_bytes:
    type: int
  type_:
    type: TensorType

Methods

`get_struct(self, simplified: bool=False) -> astx.base.ReprStruct`

title: Return the structured representation of the Tensor literal.
parameters:
  simplified:
    type: bool
returns:
  type: astx.base.ReprStruct

`TensorView(astx.base.DataType)`

title: Internal Tensor view node.
summary: >-
  Build one shallow tensor view by reusing the base storage and replacing
  the logical shape, strides, and offset metadata.
attributes:
  base:
    type: astx.AST
  shape:
    type: tuple[int, Ellipsis]
  strides:
    type: tuple[int, Ellipsis] | None
  offset_bytes:
    type: int
  type_:
    type: TensorType

Methods

`get_struct(self, simplified: bool=False) -> astx.base.ReprStruct`

title: Return the structured representation of the Tensor view.
parameters:
  simplified:
    type: bool
returns:
  type: astx.base.ReprStruct

`TensorIndex(astx.base.DataType)`

title: Internal Tensor indexed read.
attributes:
  base:
    type: astx.AST
  indices:
    type: list[astx.AST]
  type_:
    type: astx.DataType

Methods

`get_struct(self, simplified: bool=False) -> astx.base.ReprStruct`

title: Return the structured representation of the indexed read.
parameters:
  simplified:
    type: bool
returns:
  type: astx.base.ReprStruct

`TensorStore(astx.base.DataType)`

title: Internal Tensor indexed store.
summary: >-
  Stores one scalar through tensor shape and stride metadata. Arrow C++
  backed tensors remain readonly in this phase, but the node keeps the
  surface aligned with future writable views.
attributes:
  base:
    type: astx.AST
  indices:
    type: list[astx.AST]
  value:
    type: astx.AST
  type_:
    type: astx.Int32

Methods

`get_struct(self, simplified: bool=False) -> astx.base.ReprStruct`

title: Return the structured representation of the indexed store.
parameters:
  simplified:
    type: bool
returns:
  type: astx.base.ReprStruct

`TensorNDim(astx.base.DataType)`

title: Internal Tensor rank query.
attributes:
  base:
    type: astx.AST
  type_:
    type: astx.Int32

Methods

`get_struct(self, simplified: bool=False) -> astx.base.ReprStruct`

title: Return the structured representation of the rank query.
parameters:
  simplified:
    type: bool
returns:
  type: astx.base.ReprStruct

`TensorShape(astx.base.DataType)`

title: Internal Tensor shape-entry query.
attributes:
  base:
    type: astx.AST
  axis:
    type: int
  type_:
    type: astx.Int64

Methods

`get_struct(self, simplified: bool=False) -> astx.base.ReprStruct`

title: Return the structured representation of the shape query.
parameters:
  simplified:
    type: bool
returns:
  type: astx.base.ReprStruct

`TensorStride(astx.base.DataType)`

title: Internal Tensor stride-entry query.
attributes:
  base:
    type: astx.AST
  axis:
    type: int
  type_:
    type: astx.Int64

Methods

`get_struct(self, simplified: bool=False) -> astx.base.ReprStruct`

title: Return the structured representation of the stride query.
parameters:
  simplified:
    type: bool
returns:
  type: astx.base.ReprStruct

`TensorElementCount(astx.base.DataType)`

title: Internal Tensor element-count query.
attributes:
  base:
    type: astx.AST
  type_:
    type: astx.Int64

Methods

`get_struct(self, simplified: bool=False) -> astx.base.ReprStruct`

title: Return the structured representation of the element-count query.
parameters:
  simplified:
    type: bool
returns:
  type: astx.base.ReprStruct

`TensorByteOffset(astx.base.DataType)`

title: Internal Tensor byte-offset query for indexed addressing.
attributes:
  base:
    type: astx.AST
  indices:
    type: list[astx.AST]
  type_:
    type: astx.Int64

Methods

`get_struct(self, simplified: bool=False) -> astx.base.ReprStruct`

title: Return the structured representation of the byte-offset query.
parameters:
  simplified:
    type: bool
returns:
  type: astx.base.ReprStruct

`TensorRetain(astx.base.DataType)`

title: Internal explicit retain for Tensor-backed storage.
attributes:
  base:
    type: astx.AST
  type_:
    type: astx.Int32

Methods

`get_struct(self, simplified: bool=False) -> astx.base.ReprStruct`

title: Return the structured representation of the retain helper.
parameters:
  simplified:
    type: bool
returns:
  type: astx.base.ReprStruct

`TensorRelease(astx.base.DataType)`

title: Internal explicit release for Tensor-backed storage.
attributes:
  base:
    type: astx.AST
  type_:
    type: astx.Int32

Methods

`get_struct(self, simplified: bool=False) -> astx.base.ReprStruct`

title: Return the structured representation of the release helper.
parameters:
  simplified:
    type: bool
returns:
  type: astx.base.ReprStruct