DenseMatrix.hpp

Go to the documentation of this file.

#ifndef TATAMI_TILEDB_DENSE_MATRIX_HPP
#define TATAMI_TILEDB_DENSE_MATRIX_HPP
 
#include "tatami_chunked/tatami_chunked.hpp"
#include <tiledb/tiledb>
 
#include "serialize.hpp"
#include "utils.hpp"
 
#include <string>
#include <memory>
#include <vector>
#include <type_traits>
 
namespace tatami_tiledb {
 
struct DenseMatrixOptions {
    size_t maximum_cache_size = 100000000;
 
    bool require_minimum_cache = true;
};
 
namespace DenseMatrix_internal {
 
typedef ::tatami_tiledb::internal::Components Components;
typedef ::tatami_tiledb::internal::VariablyTypedDimension Dimension;
typedef ::tatami_tiledb::internal::VariablyTypedVector CacheBuffer;
 
inline void execute_query(const Components& tdb_comp, const tiledb::Subarray& subarray, const std::string& attribute, bool row, CacheBuffer& buffer, size_t offset, size_t length) {
    tiledb::Query query(tdb_comp.ctx, tdb_comp.array);
    query.set_subarray(subarray);
    query.set_layout(row ? TILEDB_ROW_MAJOR : TILEDB_COL_MAJOR);
    buffer.set_data_buffer(query, attribute, offset, length);
    if (query.submit() != tiledb::Query::Status::COMPLETE) {
        throw std::runtime_error("failed to read dense data from TileDB");
    }
}
 
/********************
 *** Core classes ***
 ********************/
 
template<typename Index_>
struct CacheParameters {
    Index_ chunk_length;
    size_t slab_size_in_elements;
    size_t max_slabs_in_cache;
};
 
template<typename Index_>
class MyopicCore {
public:
    MyopicCore(
        const Components& tdb_comp,
        const std::string& attribute, 
        bool row,
        Index_ target_dim_extent, 
        const Dimension& tdb_target_dim,
        const Dimension& tdb_non_target_dim,
        tiledb_datatype_t tdb_type,
        Index_ non_target_length,
        [[maybe_unused]] tatami::MaybeOracle<false, Index_> oracle, // for consistency with the oracular version.
        const CacheParameters<Index_>& cache_stats) :
        my_tdb_comp(tdb_comp),
        my_attribute(attribute),
        my_row(row),
        my_target_dim_extent(target_dim_extent),
        my_tdb_target_dim(tdb_target_dim),
        my_tdb_non_target_dim(tdb_non_target_dim),
        my_non_target_length(non_target_length),
        my_target_chunk_length(cache_stats.chunk_length),
        my_slab_size(cache_stats.slab_size_in_elements),
        my_holding(tdb_type, my_slab_size * cache_stats.max_slabs_in_cache),
        my_cache(cache_stats.max_slabs_in_cache)
    {}
 
private:
    const Components& my_tdb_comp;
    const std::string& my_attribute;
 
    bool my_row;
    Index_ my_target_dim_extent;
    const Dimension& my_tdb_target_dim;
    const Dimension& my_tdb_non_target_dim;
 
    Index_ my_non_target_length;
    Index_ my_target_chunk_length;
    size_t my_slab_size;
    CacheBuffer my_holding;
 
    struct Slab {
        size_t offset;
    };
    size_t my_offset = 0;
    tatami_chunked::LruSlabCache<Index_, Slab> my_cache;
 
private:
    template<typename Value_, class Configure_>
    const Value_* fetch_raw(Index_ i, Value_* buffer, Configure_ configure) {
        Index_ chunk = i / my_target_chunk_length;
        Index_ index = i % my_target_chunk_length;
 
        const auto& info = my_cache.find(
            chunk, 
            /* create = */ [&]() -> Slab {
                Slab output;
                output.offset = my_offset;
                my_offset += my_slab_size;
                return output;
            },
            /* populate = */ [&](Index_ id, Slab& contents) -> void {
                Index_ target_start = id * my_target_chunk_length;
                Index_ target_length = std::min(my_target_dim_extent - target_start, my_target_chunk_length);
 
                serialize([&]() -> void {
                    tiledb::Subarray subarray(my_tdb_comp.ctx, my_tdb_comp.array);
                    int rowdex = my_row;
                    configure(subarray, rowdex);
                    my_tdb_target_dim.add_range(subarray, 1 - rowdex, target_start, target_length);
                    execute_query(my_tdb_comp, subarray, my_attribute, my_row, my_holding, contents.offset, my_slab_size);
                });
            }
        );
 
        size_t final_offset = info.offset + static_cast<size_t>(my_non_target_length) * static_cast<size_t>(index); // cast to size_t to avoid overflow
        my_holding.copy(final_offset, my_non_target_length, buffer);
        return buffer;
    }
 
public:
    template<typename Value_>
    const Value_* fetch_block(Index_ i, Index_ block_start, Value_* buffer) {
        return fetch_raw(
            i,
            buffer,
            [&](tiledb::Subarray& subarray, int rowdex) -> void {
                my_tdb_non_target_dim.add_range(subarray, rowdex, block_start, my_non_target_length);
            }
        );
    }
 
    template<typename Value_>
    const Value_* fetch_indices(Index_ i, const std::vector<Index_>& indices, Value_* buffer) {
        return fetch_raw(
            i,
            buffer,
            [&](tiledb::Subarray& subarray, int rowdex) -> void {
                tatami::process_consecutive_indices<Index_>(
                    indices.data(),
                    indices.size(),
                    [&](Index_ s, Index_ l) -> void {
                        my_tdb_non_target_dim.add_range(subarray, rowdex, s, l);
                    }
                );
            }
        );
    }
};
 
template<typename Index_>
class OracularCore {
public:
    OracularCore(
        const Components& tdb_comp,
        const std::string& attribute, 
        bool row,
        Index_ target_dim_extent,
        const Dimension& tdb_target_dim,
        const Dimension& tdb_non_target_dim,
        tiledb_datatype_t tdb_type,
        Index_ non_target_length,
        tatami::MaybeOracle<true, Index_> oracle, 
        const CacheParameters<Index_>& cache_stats) :
        my_tdb_comp(tdb_comp),
        my_attribute(attribute),
        my_row(row),
        my_target_dim_extent(target_dim_extent),
        my_tdb_target_dim(tdb_target_dim),
        my_tdb_non_target_dim(tdb_non_target_dim),
        my_non_target_length(non_target_length),
        my_target_chunk_length(cache_stats.chunk_length),
        my_slab_size(cache_stats.slab_size_in_elements),
        my_holding(tdb_type, my_slab_size * cache_stats.max_slabs_in_cache),
        my_cache(std::move(oracle), cache_stats.max_slabs_in_cache)
    {}
 
private:
    const Components& my_tdb_comp;
    const std::string& my_attribute;
 
    bool my_row;
    Index_ my_target_dim_extent;
    const Dimension& my_tdb_target_dim;
    const Dimension& my_tdb_non_target_dim;
 
    Index_ my_non_target_length;
    Index_ my_target_chunk_length;
    size_t my_slab_size;
    CacheBuffer my_holding;
 
    struct Slab {
        size_t offset;
    };
    size_t my_offset = 0;
    tatami_chunked::OracularSlabCache<Index_, Index_, Slab, true> my_cache;
 
private:
    template<class Function_>
    static void sort_by_field(std::vector<std::pair<Index_, Slab*> >& indices, Function_ field) {
        auto comp = [&field](const std::pair<Index_, Slab*>& l, const std::pair<Index_, Slab*>& r) -> bool {
            return field(l) < field(r);
        };
        if (!std::is_sorted(indices.begin(), indices.end(), comp)) {
            std::sort(indices.begin(), indices.end(), comp);
        }
    }
 
    template<typename Value_, class Configure_>
    const Value_* fetch_raw([[maybe_unused]] Index_ i, Value_* buffer, Configure_ configure) {
        auto info = my_cache.next(
            /* identify = */ [&](Index_ current) -> std::pair<Index_, Index_> {
                return std::pair<Index_, Index_>(current / my_target_chunk_length, current % my_target_chunk_length);
            }, 
            /* create = */ [&]() -> Slab {
                Slab output;
                output.offset = my_offset;
                my_offset += my_slab_size;
                return output;
            },
            /* populate = */ [&](std::vector<std::pair<Index_, Slab*> >& to_populate, std::vector<std::pair<Index_, Slab*> >& to_reuse) -> void {
                // Defragmenting the existing chunks. We sort by offset to make 
                // sure that we're not clobbering in-use slabs during the copy().
                sort_by_field(to_reuse, [](const std::pair<Index_, Slab*>& x) -> size_t { return x.second->offset; });
                size_t running_offset = 0;
                for (auto& x : to_reuse) {
                    auto& cur_offset = x.second->offset;
                    if (cur_offset != running_offset) {
                        my_holding.shift(cur_offset, my_slab_size, running_offset);
                        cur_offset = running_offset;
                    }
                    running_offset += my_slab_size;
                }
 
                // Collapsing runs of consecutive ranges into a single range;
                // otherwise, making union of ranges. This allows a single TileDb call
                // to populate the contiguous memory pool that we made available after
                // defragmentation; then we just update the slab pointers to refer
                // to the slices of memory corresponding to each slab.
                sort_by_field(to_populate, [](const std::pair<Index_, Slab*>& x) -> Index_ { return x.first; });
 
                serialize([&]() -> void {
                    tiledb::Subarray subarray(my_tdb_comp.ctx, my_tdb_comp.array);
                    int rowdex = my_row;
                    configure(subarray, rowdex);
 
                    // Remember, the slab size is equal to the product of the chunk length and the 
                    // non-target length, so shifting the memory pool offsets by 'slab_size' will 
                    // correspond to a shift of 'chunk_length' on the target dimension. The only
                    // exception is that of the last chunk, but at that point it doesn't matter as 
                    // there's no data following the last chunk.
                    Index_ run_chunk_id = to_populate.front().first;
                    Index_ run_chunk_start = run_chunk_id * my_target_chunk_length;
                    Index_ run_length = std::min(my_target_dim_extent - run_chunk_start, my_target_chunk_length);
 
                    to_populate.front().second->offset = running_offset;
                    auto start_offset = running_offset;
                    running_offset += my_slab_size;
 
                    int dimdex = 1 - rowdex;
                    for (size_t ci = 1, cend = to_populate.size(); ci < cend; ++ci) {
                        auto& current_chunk = to_populate[ci];
                        Index_ current_chunk_id = current_chunk.first;
                        Index_ current_chunk_start = current_chunk_id * my_target_chunk_length;
 
                        if (current_chunk_id - run_chunk_id > 1) { // save the existing run of to_populate as one range, and start a new run.
                            my_tdb_target_dim.add_range(subarray, dimdex, run_chunk_start, run_length);
                            run_chunk_id = current_chunk_id;
                            run_chunk_start = current_chunk_start;
                            run_length = 0;
                        }
 
                        Index_ current_length = std::min(my_target_dim_extent - current_chunk_start, my_target_chunk_length);
                        run_length += current_length;
                        current_chunk.second->offset = running_offset;
                        running_offset += my_slab_size;
                    }
 
                    my_tdb_target_dim.add_range(subarray, dimdex, run_chunk_start, run_length);
                    execute_query(my_tdb_comp, subarray, my_attribute, my_row, my_holding, start_offset, running_offset - start_offset);
                });
            }
        );
 
        size_t final_offset = info.first->offset + my_non_target_length * static_cast<size_t>(info.second); // cast to size_t to avoid overflow
        my_holding.copy(final_offset, my_non_target_length, buffer);
        return buffer;
    }
 
public:
    template<typename Value_>
    const Value_* fetch_block(Index_ i, Index_ block_start, Value_* buffer) {
        return fetch_raw(
            i,
            buffer,
            [&](tiledb::Subarray& subarray, int rowdex) -> void {
                my_tdb_non_target_dim.add_range(subarray, rowdex, block_start, my_non_target_length);
            }
        );
    }
 
    template<typename Value_>
    const Value_* fetch_indices(Index_ i, const std::vector<Index_>& indices, Value_* buffer) {
        return fetch_raw(
            i,
            buffer, 
            [&](tiledb::Subarray& subarray, int rowdex) -> void {
                tatami::process_consecutive_indices<Index_>(
                    indices.data(),
                    indices.size(),
                    [&](Index_ s, Index_ l) -> void {
                        my_tdb_non_target_dim.add_range(subarray, rowdex, s, l);
                    }
                );
            }
        );
    }
};
 
template<bool oracle_, typename Index_>
using DenseCore = typename std::conditional<oracle_, OracularCore<Index_>, MyopicCore<Index_> >::type;
 
/***************************
 *** Concrete subclasses ***
 ***************************/
 
template<bool oracle_, typename Value_, typename Index_>
class Full : public tatami::DenseExtractor<oracle_, Value_, Index_> {
public:
    Full(
        const Components& tdb_comp,
        const std::string& attribute, 
        bool row,
        Index_ target_dim_extent,
        const Dimension& tdb_target_dim,
        const Dimension& tdb_non_target_dim,
        tiledb_datatype_t tdb_type,
        tatami::MaybeOracle<oracle_, Index_> oracle,
        Index_ non_target_dim,
        const CacheParameters<Index_>& cache_stats) :
        my_core(
            tdb_comp,
            attribute,
            row,
            target_dim_extent,
            tdb_target_dim,
            tdb_non_target_dim,
            tdb_type,
            non_target_dim,
            std::move(oracle),
            cache_stats 
        )
    {}
 
    const Value_* fetch(Index_ i, Value_* buffer) {
        return my_core.fetch_block(i, 0, buffer);
    }
 
private:
    DenseCore<oracle_, Index_> my_core;
};
 
template<bool oracle_, typename Value_, typename Index_>
class Block : public tatami::DenseExtractor<oracle_, Value_, Index_> {
public:
    Block(
        const Components& tdb_comp,
        const std::string& attribute, 
        bool row,
        Index_ target_dim_extent,
        const Dimension& tdb_target_dim,
        const Dimension& tdb_non_target_dim,
        tiledb_datatype_t tdb_type,
        tatami::MaybeOracle<oracle_, Index_> oracle,
        Index_ block_start,
        Index_ block_length,
        const CacheParameters<Index_>& cache_stats) :
        my_core( 
            tdb_comp,
            attribute,
            row,
            target_dim_extent,
            tdb_target_dim,
            tdb_non_target_dim,
            tdb_type,
            block_length,
            std::move(oracle),
            cache_stats
        ),
        my_block_start(block_start)
    {}
 
    const Value_* fetch(Index_ i, Value_* buffer) {
        return my_core.fetch_block(i, my_block_start, buffer);
    }
 
private:
    DenseCore<oracle_, Index_> my_core;
    Index_ my_block_start;
};
 
template<bool oracle_, typename Value_, typename Index_>
class Index : public tatami::DenseExtractor<oracle_, Value_, Index_> {
public:
    Index(
        const Components& tdb_comp,
        const std::string& attribute, 
        bool row,
        Index_ target_dim_extent,
        const Dimension& tdb_target_dim,
        const Dimension& tdb_non_target_dim,
        tiledb_datatype_t tdb_type,
        tatami::MaybeOracle<oracle_, Index_> oracle,
        tatami::VectorPtr<Index_> indices_ptr,
        const CacheParameters<Index_>& cache_stats) :
        my_core(
            tdb_comp,
            attribute,
            row,
            target_dim_extent,
            tdb_target_dim,
            tdb_non_target_dim,
            tdb_type,
            indices_ptr->size(),
            std::move(oracle),
            cache_stats
        ),
        my_indices_ptr(std::move(indices_ptr))
    {}
 
    const Value_* fetch(Index_ i, Value_* buffer) {
        return my_core.fetch_indices(i, *my_indices_ptr, buffer);
    }
 
private:
    DenseCore<oracle_, Index_> my_core;
    tatami::VectorPtr<Index_> my_indices_ptr; 
};
 
}
template<typename Value_, typename Index_>
class DenseMatrix : public tatami::Matrix<Value_, Index_> {
public:
    DenseMatrix(const std::string& uri, std::string attribute, tiledb::Context ctx, const DenseMatrixOptions& options) : my_attribute(std::move(attribute)) {
        initialize(uri, std::move(ctx), options);
    }
 
    DenseMatrix(const std::string& uri, std::string attribute, const DenseMatrixOptions& options) : my_attribute(std::move(attribute)) {
        initialize(uri, false, options);
    }
 
    DenseMatrix(const std::string& uri, std::string attribute) : DenseMatrix(uri, std::move(attribute), DenseMatrixOptions()) {}
 
private:
    template<class PossibleContext_>
    void initialize(const std::string& uri, PossibleContext_ ctx, const DenseMatrixOptions& options) {
        serialize([&]() -> void {
            my_tdb_comp.reset(
                [&]{
                    // If we have to create our own Context_ object, we do so inside the serialized
                    // section, rather than using a delegating constructor.
                    if constexpr(std::is_same<PossibleContext_, tiledb::Context>::value) {
                        return new DenseMatrix_internal::Components(std::move(ctx), uri);
                    } else {
                        return new DenseMatrix_internal::Components(uri);
                    }
                }(),
                [](DenseMatrix_internal::Components* ptr) -> void {
                    // Serializing the deleter, for completeness's sake.
                    serialize([&]() -> void {
                        delete ptr;
                    });
                }
            );
 
            auto schema = my_tdb_comp->array.schema();
            if (schema.array_type() != TILEDB_DENSE) {
                throw std::runtime_error("TileDB array should be dense");
            }
 
            if (!schema.has_attribute(my_attribute)) {
                throw std::runtime_error("no attribute '" + my_attribute + "' is present in the TileDB array");
            }
            auto attr = schema.attribute(my_attribute);
            my_tdb_type = attr.type();
 
            my_cache_size_in_elements = options.maximum_cache_size / internal::determine_type_size(my_tdb_type);
            my_require_minimum_cache = options.require_minimum_cache;
 
            tiledb::Domain domain = schema.domain();
            if (domain.ndim() != 2) {
                throw std::runtime_error("TileDB array should have exactly two dimensions");
            }
 
            tiledb::Dimension first_dim = domain.dimension(0);
            my_tdb_first_dim.reset(first_dim);
            Index_ first_extent = my_tdb_first_dim.extent<Index_>();
            Index_ first_tile = my_tdb_first_dim.tile<Index_>();
            my_firstdim_stats = tatami_chunked::ChunkDimensionStats<Index_>(first_extent, first_tile);
 
            tiledb::Dimension second_dim = domain.dimension(1);
            my_tdb_second_dim.reset(second_dim);
            Index_ second_extent = my_tdb_second_dim.extent<Index_>();
            Index_ second_tile = my_tdb_second_dim.tile<Index_>();
            my_seconddim_stats = tatami_chunked::ChunkDimensionStats<Index_>(second_extent, second_tile);
 
            // Favoring extraction on the dimension that involves pulling out fewer chunks per dimension element.
            auto tiles_per_firstdim = (second_extent / second_tile) + (second_extent % second_tile > 0);
            auto tiles_per_seconddim = (first_extent / first_tile) + (first_extent % first_tile > 0);
            my_prefer_firstdim = tiles_per_firstdim <= tiles_per_seconddim;
        });
    }
 
private:
    std::shared_ptr<DenseMatrix_internal::Components> my_tdb_comp;
 
    DenseMatrix_internal::Dimension my_tdb_first_dim, my_tdb_second_dim;
    tiledb_datatype_t my_tdb_type;
 
    std::string my_attribute;
    size_t my_cache_size_in_elements;
    bool my_require_minimum_cache;
 
    int my_first_offset, my_second_offset;
    tatami_chunked::ChunkDimensionStats<Index_> my_firstdim_stats, my_seconddim_stats;
    bool my_prefer_firstdim;
 
private:
    Index_ nrow_internal() const {
        return my_firstdim_stats.dimension_extent;
    }
 
    Index_ ncol_internal() const {
        return my_seconddim_stats.dimension_extent;
    }
 
public:
    Index_ nrow() const {
        return nrow_internal();
    }
 
    Index_ ncol() const {
        return ncol_internal();
    }
 
    bool is_sparse() const {
        return false;
    }
 
    double is_sparse_proportion() const {
        return 0;
    }
 
    bool prefer_rows() const {
        return my_prefer_firstdim;
    }
 
    double prefer_rows_proportion() const {
        return static_cast<double>(my_prefer_firstdim);
    }
 
    bool uses_oracle(bool) const {
        // The oracle won't necessarily be used if the cache size is non-zero,
        // but if the cache is empty, the oracle definitely _won't_ be used.
        return my_cache_size_in_elements > 0;
    }
 
private:
    template<bool oracle_, template<bool, typename, typename> class Extractor_, typename ... Args_>
    std::unique_ptr<tatami::DenseExtractor<oracle_, Value_, Index_> > populate(bool row, Index_ non_target_length, tatami::MaybeOracle<oracle_, Index_> oracle, Args_&& ... args) const {
        const auto& target_dim_stats = (row ? my_firstdim_stats : my_seconddim_stats);
        const auto& tdb_target_dim = (row ? my_tdb_first_dim : my_tdb_second_dim);
        const auto& tdb_non_target_dim = (row ? my_tdb_second_dim : my_tdb_first_dim);
 
        tatami_chunked::SlabCacheStats slab_stats(
            target_dim_stats.chunk_length,
            non_target_length,
            target_dim_stats.num_chunks,
            my_cache_size_in_elements,
            my_require_minimum_cache
        );
 
        // No need to have a dedicated SoloCore for uncached extraction,
        // because it would still need to hold a single Workspace. We instead
        // reuse the existing code with a chunk length of 1 to achieve the same
        // memory usage. This has a mild perf hit from the caching machinery
        // but perf already sucks without caching so who cares.
        DenseMatrix_internal::CacheParameters<Index_> cache_params;
        if (slab_stats.max_slabs_in_cache > 0) {
            cache_params.chunk_length = target_dim_stats.chunk_length;
            cache_params.slab_size_in_elements = slab_stats.slab_size_in_elements;
            cache_params.max_slabs_in_cache = slab_stats.max_slabs_in_cache;
        } else {
            cache_params.chunk_length = 1;
            cache_params.slab_size_in_elements = non_target_length;
            cache_params.max_slabs_in_cache = 1;
        }
 
        return std::make_unique<Extractor_<oracle_, Value_, Index_> >(
            *my_tdb_comp,
            my_attribute, 
            row,
            target_dim_stats.dimension_extent,
            tdb_target_dim,
            tdb_non_target_dim,
            my_tdb_type,
            std::move(oracle),
            std::forward<Args_>(args)...,
            cache_params
        );
    }
 
    /********************
     *** Myopic dense ***
     ********************/
public:
    std::unique_ptr<tatami::MyopicDenseExtractor<Value_, Index_> > dense(bool row, const tatami::Options&) const {
        Index_ full_non_target = (row ? ncol_internal() : nrow_internal());
        return populate<false, DenseMatrix_internal::Full>(row, full_non_target, false, full_non_target);
    }
 
    std::unique_ptr<tatami::MyopicDenseExtractor<Value_, Index_> > dense(bool row, Index_ block_start, Index_ block_length, const tatami::Options&) const {
        return populate<false, DenseMatrix_internal::Block>(row, block_length, false, block_start, block_length);
    }
 
    std::unique_ptr<tatami::MyopicDenseExtractor<Value_, Index_> > dense(bool row, tatami::VectorPtr<Index_> indices_ptr, const tatami::Options&) const {
        auto nidx = indices_ptr->size();
        return populate<false, DenseMatrix_internal::Index>(row, nidx, false, std::move(indices_ptr));
    }
 
    /*********************
     *** Myopic sparse ***
     *********************/
public:
    std::unique_ptr<tatami::MyopicSparseExtractor<Value_, Index_> > sparse(bool row, const tatami::Options& opt) const {
        Index_ full_non_target = (row ? ncol_internal() : nrow_internal());
        return std::make_unique<tatami::FullSparsifiedWrapper<false, Value_, Index_> >(dense(row, opt), full_non_target, opt);
    }
 
    std::unique_ptr<tatami::MyopicSparseExtractor<Value_, Index_> > sparse(bool row, Index_ block_start, Index_ block_length, const tatami::Options& opt) const {
        return std::make_unique<tatami::BlockSparsifiedWrapper<false, Value_, Index_> >(dense(row, block_start, block_length, opt), block_start, block_length, opt);
    }
 
    std::unique_ptr<tatami::MyopicSparseExtractor<Value_, Index_> > sparse(bool row, tatami::VectorPtr<Index_> indices_ptr, const tatami::Options& opt) const {
        auto ptr = dense(row, indices_ptr, opt);
        return std::make_unique<tatami::IndexSparsifiedWrapper<false, Value_, Index_> >(std::move(ptr), std::move(indices_ptr), opt);
    }
 
    /**********************
     *** Oracular dense ***
     **********************/
public:
    std::unique_ptr<tatami::OracularDenseExtractor<Value_, Index_> > dense(
        bool row,
        std::shared_ptr<const tatami::Oracle<Index_> > oracle,
        const tatami::Options&) 
    const {
        Index_ full_non_target = (row ? ncol_internal() : nrow_internal());
        return populate<true, DenseMatrix_internal::Full>(row, full_non_target, std::move(oracle), full_non_target);
    }
 
    std::unique_ptr<tatami::OracularDenseExtractor<Value_, Index_> > dense(
        bool row, 
        std::shared_ptr<const tatami::Oracle<Index_> > oracle, 
        Index_ block_start, 
        Index_ block_length, 
        const tatami::Options&) 
    const {
        return populate<true, DenseMatrix_internal::Block>(row, block_length, std::move(oracle), block_start, block_length);
    }
 
    std::unique_ptr<tatami::OracularDenseExtractor<Value_, Index_> > dense(
        bool row, 
        std::shared_ptr<const tatami::Oracle<Index_> > oracle, 
        tatami::VectorPtr<Index_> indices_ptr, 
        const tatami::Options&) 
    const {
        auto nidx = indices_ptr->size();
        return populate<true, DenseMatrix_internal::Index>(row, nidx, std::move(oracle), std::move(indices_ptr));
    }
 
    /***********************
     *** Oracular sparse ***
     ***********************/
public:
    std::unique_ptr<tatami::OracularSparseExtractor<Value_, Index_> > sparse(
        bool row, 
        std::shared_ptr<const tatami::Oracle<Index_> > oracle, 
        const tatami::Options& opt) 
    const {
        Index_ full_non_target = (row ? ncol_internal() : nrow_internal());
        return std::make_unique<tatami::FullSparsifiedWrapper<true, Value_, Index_> >(dense(row, std::move(oracle), opt), full_non_target, opt);
    }
 
    std::unique_ptr<tatami::OracularSparseExtractor<Value_, Index_> > sparse(
        bool row, 
        std::shared_ptr<const tatami::Oracle<Index_> > oracle, 
        Index_ block_start, 
        Index_ block_length, 
        const tatami::Options& opt) 
    const {
        return std::make_unique<tatami::BlockSparsifiedWrapper<true, Value_, Index_> >(dense(row, std::move(oracle), block_start, block_length, opt), block_start, block_length, opt);
    }
 
    std::unique_ptr<tatami::OracularSparseExtractor<Value_, Index_> > sparse(
        bool row, 
        std::shared_ptr<const tatami::Oracle<Index_> > oracle, 
        tatami::VectorPtr<Index_> indices_ptr, 
        const tatami::Options& opt) 
    const {
        auto ptr = dense(row, std::move(oracle), indices_ptr, opt);
        return std::make_unique<tatami::IndexSparsifiedWrapper<true, Value_, Index_> >(std::move(ptr), std::move(indices_ptr), opt);
    }
};
 
}
 
#endif