Added apply() operator

[cliffburdick]

Add apply operator for custom element-wise transformations

Introduces a new 'apply' operator that allows users to define custom
element-wise transformations using lambda functions or functors on one
or more input operators.

Key features:

• Accepts variadic number of input operators
• Supports host, device, or both annotations on functions
• Automatically deduces return type from lambda/functor
• Output rank matches first input operator's rank
• Output size matches first input operator's size
• Full capability forwarding from nested operators

Implementation details:

• Created ApplyOp class in include/matx/operators/apply.h

Build system changes:

• Added MATX_EN_EXTENDED_LAMBDA CMake option (default ON)
• Automatically adds --extended-lambda flag when enabled
• Required for host device lambdas in certain contexts

Testing:

• Comprehensive test suite in test/00_operators/apply_test.cu
• Tests single, dual, and triple input operators
• Tests with 1D, 2D, 3D tensors
• Tests nested operators, chaining, and scalars
• Tests with all numeric types and executors
• Uses functors to avoid extended lambda restrictions in test contexts

Documentation:

• Added docs_input/api/manipulation/basic/apply.rst
• Enhanced Doxygen comments with usage examples
• Includes notes about std::apply naming collision
• Added example code markers in test file for doc generation

[copy-pr-bot]

This pull request requires additional validation before any workflows can run on NVIDIA's runners.

Pull request vetters can view their responsibilities here.

Contributors can view more details about this message here.

[greptile-apps]

_{5 files reviewed, no comments}

_{Edit Code Review Agent Settings | Greptile}

[cliffburdick]

/build

[cliffburdick]

/build

[cliffburdick]

/build

[cliffburdick]

/build

[cliffburdick]

/build