Added GPU interfaces for shortest paths algorithm

[olegkkruglov]

Description

Contains content from #3482, #3620, #3621, and same CI problems

• GPU kernel — Created backend/gpu/traverse_default_kernel_gpu_dpc.cpp with a label-correcting SSSP algorithm using frontier advance. Includes a weighted_csr_graph_device_wrapper that wraps int64 row offsets + int32 column indices + typed edge weights for the frontier API. Distances are relaxed via sycl::atomic_ref::fetch_min, with optional predecessor tracking.
• GPU kernel headers — Created backend/gpu/shortest_paths.hpp (kernel struct declaration) and backend/gpu/traverse_kernel.hpp (GPU view struct + run_shortest_paths_gpu declaration).
• DPC++ dispatch — Created select_kernel_dpc.cpp that dispatches to CPU (host policy) or GPU (context_gpu) via dispatch_by_device. Extracts topology and edge values from the directed graph, passes to the GPU kernel. Explicit instantiations for both int32 and double edge weight types.
• DPC++ specialization in select_kernel.hpp — Added #ifdef ONEDAL_DATA_PARALLEL block with backend_default<data_parallel_policy, ...> specialization.
• Queue overloads — Added data_parallel_policy dispatcher specialization in traverse_ops.hpp (algo-level), DPC++ overloads in traverse_ops.hpp, and traverse(sycl::queue&, ...) overload in traverse.hpp.
• GPU test — Created test/shortest_paths_gpu_test.cpp with 4 test cases: isolated vertices (double and int32), one-bucket graph, and a 21-vertex graph. Double tests gracefully skip on devices without fp64 support.
• DPC++ example — Created shortest_paths_batch.cpp that runs SSSP on all available SYCL devices, with fp64 detection and skip.
• BUILD updates — Added frontier primitive dependency, split tests into CPU-only and GPU suites, added shortest_paths to DPC++ examples BUILD.

---

Checklist:

Completeness and readability

• I have commented my code, particularly in hard-to-understand areas.
• I have updated the documentation to reflect the changes or created a separate PR with updates and provided its number in the description, if necessary.
• Git commit message contains an appropriate signed-off-by string (see CONTRIBUTING.md for details).
• I have resolved any merge conflicts that might occur with the base branch.

Testing

• I have run it locally and tested the changes extensively.
• All CI jobs are green or I have provided justification why they aren't.
• I have extended testing suite if new functionality was introduced in this PR.

Performance

• I have measured performance for affected algorithms using scikit-learn_bench and provided at least a summary table with measured data, if performance change is expected.
• I have provided justification why performance and/or quality metrics have changed or why changes are not expected.
• I have extended the benchmarking suite and provided a corresponding scikit-learn_bench PR if new measurable functionality was introduced in this PR.