factor out uses of omp_get_num_threads() and omp_get_max_threads() outside of OpenMP wrapper

[jameslamb]

Contributes to #4705.
Contributes to #5987.

LightGBM uses OpenMP for to parallelize many operations via multi-threading.

It's C/C++ codebase uses some wrapper functions defined in a header file openmp_wrapper.h to get and set the number of threads to use in such operations:

LightGBM/include/LightGBM/utils/openmp_wrapper.h

Lines 20 to 35 in f901f47

	inline int OMP_NUM_THREADS() {
	int ret = 1;
	#pragma omp parallel
	#pragma omp master
	{ ret = omp_get_num_threads(); }
	return ret;
	}
	inline void OMP_SET_NUM_THREADS(int num_threads) {
	static const int default_omp_num_threads = OMP_NUM_THREADS();
	if (num_threads > 0) {
	omp_set_num_threads(num_threads);
	} else {
	omp_set_num_threads(default_omp_num_threads);
	}
	}

There are a few places in the code which bypass those helpers and instead directly call omp_get_num_threads() and omp_get_max_threads().

This PR proposes switching such calls to instead use LightGBM's wrappers. That reduces the number of different code paths that multi-threading configuration goes through, therefore reducing the risk of configuration inconsistencies and issues like #4705 and #5124.

[jameslamb]

thanks so much for the quick reviews @guolinke !

[github-actions]

This pull request has been automatically locked since there has not been any recent activity since it was closed. To start a new related discussion, open a new issue at https://github.com/microsoft/LightGBM/issues including a reference to this.