🔍 What is mr-contractor?
mr-contractor is a header-only C++ library that lets you declaratively define task execution graphs with seamless support for:
- Sequential & parallel stages
- Arbitrary nesting (parallel inside sequential, sequential inside parallel, etc.)
- Automatic dependency resolution
- Zero-boilerplate thread pooling
Built on latest C++ features, it’s perfect for scientific computing, game engines, or any domain requiring complex, repeatable workflows.
🎯 Key Features
- Declarative API: Describe what your workflow does, not how to schedule it.
- Work Contracts: Underlying thread pool powered by
work_contractfor minimal overhead. - Type Safety: Compile-time validation of task input/output types.
🚀 Usage Examples
1. Simple Sequential Pipeline
auto task = Sequence{
[](int x) -> float { return x / 2.0f; },
[](float y) -> std::string { return std::to_string(y); }
};
auto result = apply(task, 42)->execute()->result();
// result = "21.0" 2. Parallel Processing
auto task = Parallel{
[](int a) { return a * a; }, // Task 1
[](int b) { return b + 10; } // Task 2 (runs concurrently)
};
auto results = apply(task, std::tuple{2, 3})->execute()->result();
// results = (4, 13) 3. Nested Workflows
auto task = Sequence{
[](int x) { return std::tuple{x, x*2}; },
Parallel{
Sequence{ // Nested sequential steps
[](int a) { return a + 1; },
[](int b) { return std::to_string(b); }
},
[](int c) { return c * 0.5f; }
},
[](auto&& inputs) {
auto&& [str, flt] = inputs;
return str + " @ " + std::to_string(flt);
}
};
auto result = apply(task, 5)->execute()->result();
// result = "6 @ 5.0" - via CPM.cmake in your CMake script (suggested):
CPMAddPackage("gh:4J-company/mr-contractor#master") target_link_libraries(<your-project> PUBLIC mr-contractor-lib)
We’d Love Your Feedback!
- Found a bug? Open an issue!
- Have a feature request? Let’s discuss!
- Built something cool with it? Share your story!