Rewrite sparse dmatrix using callbacks. (#7092)

- Reduce dependency on dmlc parsers and provide an interface for users to load data by themselves.
- Remove use of threaded iterator and IO queue.
- Remove `page_size`.
- Make sure the number of pages in memory is bounded.
- Make sure the cache can not be violated.
- Provide an interface for internal algorithms to process data asynchronously.

tests/cpp/common/test_column_matrix.cc

@@ -125,12 +125,10 @@ TEST(DenseColumnWithMissing, Test) {
 }
 
 void TestGHistIndexMatrixCreation(size_t nthreads) {
-  dmlc::TemporaryDirectory tmpdir;
-  std::string filename = tmpdir.path + "/big.libsvm";
   size_t constexpr kPageSize = 1024, kEntriesPerCol = 3;
   size_t constexpr kEntries = kPageSize * kEntriesPerCol * 2;
   /* This should create multiple sparse pages */
-  std::unique_ptr<DMatrix> dmat{ CreateSparsePageDMatrix(kEntries, kPageSize, filename) };
+  std::unique_ptr<DMatrix> dmat{ CreateSparsePageDMatrix(kEntries) };
   omp_set_num_threads(nthreads);
   GHistIndexMatrix gmat(dmat.get(), 256);
 }