# Step 18: Load Hierarchy

## 담당 역할과 필수 스킬

- 담당 역할: `implementation-agent`
- 필수 스킬: `harness`, `fesa-cpp-msvc-tdd`
- 이 Step만 `RED -> observed failure -> minimal GREEN -> focused/full VERIFY`로 수행한다.

## 읽어야 할 파일

- `/AGENTS.md`
- `/docs/CODINGSTYLE.md`
- `/docs/ARCHITECTURE.md`
- `/docs/ADR.md`
- `/docs/cpp-object-oriented-modular-refactoring/implementation-plan.md`
- Step 11 source-target resolver files
- Step 13 Domain/AnalysisModel files
- Step 15 DofManager files
- `/include/fesa/assembly/load_assembler.h`, `/src/fesa/assembly/load_assembler.cpp`
- `/tests/unit/assembly/load_assembler_test.cpp`
- `/tests/unit/analysis/analysis_model_test.cpp`
- `/src/fesa/io/abaqus/domain_mapper.cpp`
- `/src/fesa/CMakeLists.txt`, `/tests/CMakeLists.txt`
- `/phases/cpp-object-oriented-modular-refactoring/index.json`
- `/phases/cpp-object-oriented-modular-refactoring/step18.md`

## 작업

Requirement `R-LOAD-001`을 구현한다.

1. `/tests/unit/loads/load_test.cpp`와 LoadAssembler generic test를 먼저 추가한다.
   `C-LOAD-001` uses a fake Load to emit ordered contributions, verifies signed duplicate target
   accumulation and rejects nonfinite/index errors before global mutation.
2. Missing Load base/contribution API or concrete-vector-only assembler gives RED.
3. Create `/include/fesa/loads/load.h`,
   `/include/fesa/loads/concentrated_nodal_load.h`, and
   `/src/fesa/loads/concentrated_nodal_load.cpp`; register the source and load test in CMake.
4. Candidate interfaces:

```cpp
struct LoadContribution {
  std::size_t source_order;
  std::size_t full_dof_index;
  double value;
};

struct LoadContext {
  const Domain& domain;
  const DofManager& dof_manager;
  const SourceTargetResolver& target_resolver;
};

class Load {
 public:
  virtual ~Load() = default;
  virtual Result<std::vector<LoadContribution>> ComputeContributions(
      const LoadContext& context) const = 0;
};

class ConcentratedNodalLoad final : public Load {
 public:
  ConcentratedNodalLoad(SourceTargetQuery target,
                        std::array<double, 6> global_components,
                        std::size_t source_order);
  Result<std::vector<LoadContribution>> ComputeContributions(
      const LoadContext& context) const override;

 private:
  SourceTargetQuery target_;
  std::array<double, 6> global_components_;
  std::size_t source_order_;
};
```

5. Domain StepDefinition owns loads through `unique_ptr<Load>` and exposes const active order.
6. `LoadAssembler` requests contributions in active source order, validates them, then accumulates
   into a candidate full vector in the existing deterministic order.
7. Current drilling-moment validation, exact-zero aggregate handling, nonzero prescribed effective
   RHS behavior and diagnostics remain unchanged.

## Acceptance Criteria

```powershell
cmake --build .harness/build --config Debug --target fesa_unit_tests
ctest --test-dir .harness/build -C Debug `
  -R "Load|LoadAssembly|AnalysisModel|InpDomainMapping" --output-on-failure
& "C:/Program Files/LLVM/bin/clang-format.exe" --dry-run --Werror `
  (rg --files include/fesa/loads -g "*.h") `
  (rg --files src/fesa/loads -g "*.cpp") `
  include/fesa/assembly/load_assembler.h src/fesa/assembly/load_assembler.cpp `
  tests/unit/loads/load_test.cpp tests/unit/assembly/load_assembler_test.cpp
cmake --build .harness/build --config Debug
ctest --test-dir .harness/build -C Debug --show-only=json-v1
ctest --test-dir .harness/build -C Debug --output-on-failure
```

## 검증 및 상태 갱신

- RED fake load, ordered/validation focused tests and full CTest를 summary에 기록한다.
- Reduction/order/diagnostic regression is `error`.
- 성공 시 현재 Step만 `completed`로 갱신한다.
- timestamp, retry, commit, advancement는 Executor 소유다.

## 금지사항

- Load object가 global Vector를 직접 mutate하게 하지 마라.
- Distributed load, body force or new keyword를 구현하지 마라.
- Parallel accumulation or order optimization을 추가하지 마라.
- 직접 commit하거나 hook script를 수동 실행하지 마라.