Efim Beshmenev
539 lines
19 KiB
C++
539 lines
19 KiB
C++
#pragma once
|
|
|
|
#include "universal_container/hierarchical_size_index.hpp"
|
|
#include "universal_container/ring_block.hpp"
|
|
|
|
#include <algorithm>
|
|
#include <cassert>
|
|
#include <cstddef>
|
|
#include <cstdint>
|
|
#include <limits>
|
|
#include <memory>
|
|
#include <stdexcept>
|
|
#include <type_traits>
|
|
#include <utility>
|
|
#include <vector>
|
|
|
|
namespace uc {
|
|
|
|
struct TieredConfig {
|
|
std::size_t leaf_capacity = 512;
|
|
std::size_t directory_fanout = 64;
|
|
std::size_t directory_levels = 4;
|
|
|
|
friend bool operator==(const TieredConfig&, const TieredConfig&) = default;
|
|
};
|
|
|
|
namespace detail {
|
|
|
|
template <class T>
|
|
class TieredStorage {
|
|
public:
|
|
using value_type = T;
|
|
using leaf_id_type = std::uint32_t;
|
|
static constexpr std::size_t npos = std::numeric_limits<std::size_t>::max();
|
|
|
|
explicit TieredStorage(TieredConfig config = {})
|
|
: config_(normalize(config)),
|
|
directory_(config_.directory_fanout, config_.directory_levels) {}
|
|
|
|
TieredStorage(const TieredStorage& other)
|
|
: config_(other.config_),
|
|
directory_(config_.directory_fanout, config_.directory_levels),
|
|
size_(other.size_),
|
|
next_leaf_id_(other.next_leaf_id_),
|
|
leaf_positions_(other.leaf_positions_) {
|
|
leaves_.reserve(other.leaves_.size());
|
|
for (const auto& leaf : other.leaves_) {
|
|
leaves_.push_back(std::make_unique<Leaf>(*leaf));
|
|
}
|
|
rebuild_directory();
|
|
}
|
|
|
|
TieredStorage(TieredStorage&&) noexcept = default;
|
|
|
|
TieredStorage& operator=(const TieredStorage& other) {
|
|
if (this == &other) {
|
|
return *this;
|
|
}
|
|
TieredStorage copy(other);
|
|
swap(copy);
|
|
return *this;
|
|
}
|
|
|
|
TieredStorage& operator=(TieredStorage&&) noexcept = default;
|
|
~TieredStorage() = default;
|
|
|
|
void swap(TieredStorage& other) noexcept {
|
|
using std::swap;
|
|
swap(config_, other.config_);
|
|
swap(leaves_, other.leaves_);
|
|
swap(directory_, other.directory_);
|
|
swap(size_, other.size_);
|
|
swap(next_leaf_id_, other.next_leaf_id_);
|
|
swap(leaf_positions_, other.leaf_positions_);
|
|
}
|
|
|
|
[[nodiscard]] std::size_t size() const noexcept { return size_; }
|
|
[[nodiscard]] bool empty() const noexcept { return size_ == 0; }
|
|
[[nodiscard]] const TieredConfig& config() const noexcept { return config_; }
|
|
[[nodiscard]] std::size_t leaf_count() const noexcept { return leaves_.size(); }
|
|
[[nodiscard]] std::size_t actual_levels() const noexcept { return directory_.actual_levels(); }
|
|
|
|
// Changing only the directory geometry does not move elements or leaves.
|
|
// It is substantially cheaper than rebuilding with another leaf capacity.
|
|
void reconfigure_directory(std::size_t fanout, std::size_t maximum_levels) {
|
|
auto requested = config_;
|
|
requested.directory_fanout = fanout;
|
|
requested.directory_levels = maximum_levels;
|
|
requested = normalize(requested);
|
|
if (requested.directory_fanout == config_.directory_fanout
|
|
&& requested.directory_levels == config_.directory_levels) {
|
|
return;
|
|
}
|
|
|
|
// Build the complete replacement before changing either the active
|
|
// directory or the public configuration. HierarchicalSizeIndex::rebuild
|
|
// allocates several vectors and may throw; mutating directory_ first
|
|
// would otherwise leave locate() unusable after an allocation failure.
|
|
std::vector<std::size_t> sizes;
|
|
sizes.reserve(leaves_.size());
|
|
for (const auto& leaf : leaves_) {
|
|
sizes.push_back(leaf->values.size());
|
|
}
|
|
HierarchicalSizeIndex rebuilt(requested.directory_fanout,
|
|
requested.directory_levels);
|
|
rebuilt.rebuild(sizes);
|
|
|
|
static_assert(std::is_nothrow_move_assignable_v<HierarchicalSizeIndex>);
|
|
directory_ = std::move(rebuilt);
|
|
config_.directory_fanout = requested.directory_fanout;
|
|
config_.directory_levels = requested.directory_levels;
|
|
}
|
|
|
|
T& operator[](std::size_t index) noexcept {
|
|
const auto location = locate_unchecked(index);
|
|
return leaves_[location.leaf]->values[location.local];
|
|
}
|
|
|
|
const T& operator[](std::size_t index) const noexcept {
|
|
const auto location = locate_unchecked(index);
|
|
return leaves_[location.leaf]->values[location.local];
|
|
}
|
|
|
|
T& at(std::size_t index) {
|
|
check_index(index);
|
|
return (*this)[index];
|
|
}
|
|
|
|
const T& at(std::size_t index) const {
|
|
check_index(index);
|
|
return (*this)[index];
|
|
}
|
|
|
|
template <class Relocate>
|
|
void push_back(T value, Relocate&& relocate) {
|
|
if (leaves_.empty() || leaves_.back()->values.full()) {
|
|
leaves_.push_back(make_leaf());
|
|
rebuild_positions();
|
|
rebuild_directory();
|
|
}
|
|
auto& leaf = *leaves_.back();
|
|
leaf.values.push_back(std::move(value));
|
|
++size_;
|
|
directory_.update(leaves_.size() - 1, 1);
|
|
relocate(leaf.values[leaf.values.size() - 1], leaf.id, leaf.values.size() - 1);
|
|
}
|
|
|
|
void push_back(T value) {
|
|
push_back(std::move(value), NoRelocate{});
|
|
}
|
|
|
|
template <class Relocate>
|
|
void insert(std::size_t index, T value, Relocate&& relocate) {
|
|
if (index > size_) {
|
|
throw std::out_of_range("tiered insertion index out of range");
|
|
}
|
|
if (index == size_) {
|
|
push_back(std::move(value), std::forward<Relocate>(relocate));
|
|
return;
|
|
}
|
|
|
|
const auto location = directory_.locate(index);
|
|
auto& leaf = *leaves_[location.leaf];
|
|
if (!leaf.values.full()) {
|
|
leaf.values.insert(location.local, std::move(value));
|
|
++size_;
|
|
directory_.update(location.leaf, 1);
|
|
refresh_leaf(location.leaf, relocate);
|
|
return;
|
|
}
|
|
|
|
split_and_insert(location.leaf, location.local, std::move(value), relocate);
|
|
++size_;
|
|
rebuild_directory();
|
|
}
|
|
|
|
void insert(std::size_t index, T value) {
|
|
insert(index, std::move(value), NoRelocate{});
|
|
}
|
|
|
|
template <class Relocate>
|
|
T erase(std::size_t index, Relocate&& relocate) {
|
|
check_index(index);
|
|
const auto location = directory_.locate(index);
|
|
auto removed = leaves_[location.leaf]->values.erase(location.local);
|
|
--size_;
|
|
|
|
if (leaves_[location.leaf]->values.empty() && leaves_.size() > 1) {
|
|
retire_leaf(location.leaf);
|
|
rebuild_positions();
|
|
rebuild_directory();
|
|
return removed;
|
|
}
|
|
|
|
if (try_merge(location.leaf, relocate)) {
|
|
return removed;
|
|
}
|
|
|
|
directory_.update(location.leaf, -1);
|
|
refresh_leaf(location.leaf, relocate);
|
|
return removed;
|
|
}
|
|
|
|
T erase(std::size_t index) {
|
|
return erase(index, NoRelocate{});
|
|
}
|
|
|
|
void clear() noexcept {
|
|
leaves_.clear();
|
|
leaf_positions_.clear();
|
|
size_ = 0;
|
|
next_leaf_id_ = 0;
|
|
directory_.rebuild({});
|
|
}
|
|
|
|
template <class Function>
|
|
void for_each(Function&& function) {
|
|
for (auto& leaf : leaves_) {
|
|
for (std::size_t i = 0; i < leaf->values.size(); ++i) {
|
|
function(leaf->values[i]);
|
|
}
|
|
}
|
|
}
|
|
|
|
template <class Function>
|
|
void for_each(Function&& function) const {
|
|
for (const auto& leaf : leaves_) {
|
|
for (std::size_t i = 0; i < leaf->values.size(); ++i) {
|
|
function(leaf->values[i]);
|
|
}
|
|
}
|
|
}
|
|
|
|
template <class Function>
|
|
void for_each_with_location(Function&& function) const {
|
|
for (const auto& leaf : leaves_) {
|
|
for (std::size_t i = 0; i < leaf->values.size(); ++i) {
|
|
function(leaf->values[i], leaf->id, i);
|
|
}
|
|
}
|
|
}
|
|
|
|
[[nodiscard]] std::vector<T> to_vector_copy() const {
|
|
std::vector<T> result;
|
|
result.reserve(size_);
|
|
for_each([&](const T& value) { result.push_back(value); });
|
|
return result;
|
|
}
|
|
|
|
[[nodiscard]] std::vector<T> to_vector_move() {
|
|
std::vector<T> result;
|
|
result.reserve(size_);
|
|
for_each([&](T& value) { result.push_back(std::move(value)); });
|
|
return result;
|
|
}
|
|
|
|
template <class Relocate>
|
|
static TieredStorage from_vector(std::vector<T>&& source,
|
|
TieredConfig config,
|
|
Relocate&& relocate) {
|
|
TieredStorage result(config);
|
|
if (source.empty()) {
|
|
return result;
|
|
}
|
|
const auto target_occupancy = bulk_target_occupancy(
|
|
result.config_.leaf_capacity);
|
|
const auto count = (source.size() + target_occupancy - 1)
|
|
/ target_occupancy;
|
|
result.leaves_.reserve(count);
|
|
for (auto& value : source) {
|
|
if (result.leaves_.empty()
|
|
|| result.leaves_.back()->values.size() == target_occupancy) {
|
|
result.leaves_.push_back(result.make_leaf());
|
|
}
|
|
result.leaves_.back()->values.push_back(std::move(value));
|
|
++result.size_;
|
|
}
|
|
result.rebuild_positions();
|
|
result.rebuild_directory();
|
|
for (std::size_t leaf = 0; leaf < result.leaves_.size(); ++leaf) {
|
|
result.refresh_leaf(leaf, relocate);
|
|
}
|
|
return result;
|
|
}
|
|
|
|
static TieredStorage from_vector(std::vector<T>&& source, TieredConfig config) {
|
|
return from_vector(std::move(source), config, NoRelocate{});
|
|
}
|
|
|
|
static TieredStorage from_vector_copy(const std::vector<T>& source,
|
|
TieredConfig config) {
|
|
TieredStorage result(config);
|
|
if (source.empty()) {
|
|
return result;
|
|
}
|
|
const auto target_occupancy = bulk_target_occupancy(
|
|
result.config_.leaf_capacity);
|
|
const auto count = (source.size() + target_occupancy - 1)
|
|
/ target_occupancy;
|
|
result.leaves_.reserve(count);
|
|
for (const auto& value : source) {
|
|
if (result.leaves_.empty()
|
|
|| result.leaves_.back()->values.size() == target_occupancy) {
|
|
result.leaves_.push_back(result.make_leaf());
|
|
}
|
|
result.leaves_.back()->values.push_back(value);
|
|
++result.size_;
|
|
}
|
|
result.rebuild_positions();
|
|
result.rebuild_directory();
|
|
return result;
|
|
}
|
|
|
|
static TieredStorage reconfigured_copy(const TieredStorage& source,
|
|
TieredConfig config) {
|
|
TieredStorage result(config);
|
|
if (source.empty()) {
|
|
return result;
|
|
}
|
|
const auto target_occupancy = bulk_target_occupancy(
|
|
result.config_.leaf_capacity);
|
|
const auto count = (source.size() + target_occupancy - 1)
|
|
/ target_occupancy;
|
|
result.leaves_.reserve(count);
|
|
source.for_each([&](const T& value) {
|
|
if (result.leaves_.empty()
|
|
|| result.leaves_.back()->values.size() == target_occupancy) {
|
|
result.leaves_.push_back(result.make_leaf());
|
|
}
|
|
result.leaves_.back()->values.push_back(value);
|
|
++result.size_;
|
|
});
|
|
result.rebuild_positions();
|
|
result.rebuild_directory();
|
|
return result;
|
|
}
|
|
|
|
[[nodiscard]] std::size_t logical_index(leaf_id_type leaf_id,
|
|
std::size_t local) const {
|
|
if (leaf_id >= leaf_positions_.size()) {
|
|
throw std::out_of_range("unknown tiered leaf id");
|
|
}
|
|
const auto position = leaf_positions_[leaf_id];
|
|
if (position == npos || local >= leaves_[position]->values.size()) {
|
|
throw std::out_of_range("stale tiered location");
|
|
}
|
|
return directory_.prefix_before(position) + local;
|
|
}
|
|
|
|
[[nodiscard]] std::size_t allocated_bytes() const noexcept {
|
|
std::size_t bytes = leaves_.capacity() * sizeof(typename decltype(leaves_)::value_type)
|
|
+ leaf_positions_.capacity() * sizeof(std::size_t)
|
|
+ directory_.allocated_bytes();
|
|
for (const auto& leaf : leaves_) {
|
|
bytes += sizeof(Leaf) + leaf->values.allocated_bytes();
|
|
}
|
|
return bytes;
|
|
}
|
|
|
|
private:
|
|
struct NoRelocate {
|
|
void operator()(const T&, leaf_id_type, std::size_t) const noexcept {}
|
|
};
|
|
|
|
struct Leaf {
|
|
Leaf(leaf_id_type leaf_id, std::size_t capacity)
|
|
: id(leaf_id), values(capacity) {}
|
|
|
|
leaf_id_type id;
|
|
RingBlock<T> values;
|
|
};
|
|
|
|
static TieredConfig normalize(TieredConfig config) {
|
|
config.leaf_capacity = std::clamp<std::size_t>(config.leaf_capacity, 4, 1U << 20U);
|
|
config.directory_fanout = std::clamp<std::size_t>(config.directory_fanout, 2, 1U << 16U);
|
|
config.directory_levels = std::clamp<std::size_t>(config.directory_levels, 1, 8);
|
|
return config;
|
|
}
|
|
|
|
// Bulk-loading completely full leaves makes the first random insertion
|
|
// into almost every leaf pay for a split and a directory rebuild. Keep a
|
|
// small, deterministic reserve instead. Seven eighths preserves compact
|
|
// memory usage while giving each freshly built leaf enough room for the
|
|
// short edit bursts that motivate switching away from vector storage.
|
|
[[nodiscard]] static std::size_t
|
|
bulk_target_occupancy(std::size_t capacity) noexcept {
|
|
return std::max<std::size_t>(1, capacity - capacity / 8);
|
|
}
|
|
|
|
[[nodiscard]] std::unique_ptr<Leaf> make_leaf() {
|
|
if (next_leaf_id_ == std::numeric_limits<leaf_id_type>::max()) {
|
|
throw std::length_error("tiered leaf id space exhausted");
|
|
}
|
|
const auto id = next_leaf_id_++;
|
|
if (leaf_positions_.size() <= id) {
|
|
leaf_positions_.resize(static_cast<std::size_t>(id) + 1, npos);
|
|
}
|
|
return std::make_unique<Leaf>(id, config_.leaf_capacity);
|
|
}
|
|
|
|
void check_index(std::size_t index) const {
|
|
if (index >= size_) {
|
|
throw std::out_of_range("tiered index out of range");
|
|
}
|
|
}
|
|
|
|
[[nodiscard]] typename HierarchicalSizeIndex::Location
|
|
locate_unchecked(std::size_t index) const noexcept {
|
|
assert(index < size_);
|
|
try {
|
|
return directory_.locate(index);
|
|
} catch (...) {
|
|
std::terminate();
|
|
}
|
|
}
|
|
|
|
void rebuild_positions() {
|
|
std::fill(leaf_positions_.begin(), leaf_positions_.end(), npos);
|
|
for (std::size_t i = 0; i < leaves_.size(); ++i) {
|
|
const auto id = leaves_[i]->id;
|
|
if (leaf_positions_.size() <= id) {
|
|
leaf_positions_.resize(static_cast<std::size_t>(id) + 1, npos);
|
|
}
|
|
leaf_positions_[id] = i;
|
|
}
|
|
}
|
|
|
|
void rebuild_directory() {
|
|
std::vector<std::size_t> sizes;
|
|
sizes.reserve(leaves_.size());
|
|
for (const auto& leaf : leaves_) {
|
|
sizes.push_back(leaf->values.size());
|
|
}
|
|
directory_.rebuild(sizes);
|
|
assert(directory_.total() == size_);
|
|
}
|
|
|
|
template <class Relocate>
|
|
void refresh_leaf(std::size_t leaf_index, Relocate& relocate) {
|
|
auto& leaf = *leaves_[leaf_index];
|
|
for (std::size_t i = 0; i < leaf.values.size(); ++i) {
|
|
relocate(leaf.values[i], leaf.id, i);
|
|
}
|
|
}
|
|
|
|
template <class Relocate>
|
|
void split_and_insert(std::size_t leaf_index,
|
|
std::size_t local,
|
|
T value,
|
|
Relocate& relocate) {
|
|
auto& old = *leaves_[leaf_index];
|
|
std::vector<T> combined;
|
|
combined.reserve(old.values.size() + 1);
|
|
for (std::size_t i = 0; i < old.values.size(); ++i) {
|
|
combined.push_back(std::move(old.values[i]));
|
|
}
|
|
combined.insert(combined.begin() + static_cast<std::ptrdiff_t>(local), std::move(value));
|
|
|
|
const auto left_id = old.id;
|
|
auto left = std::make_unique<Leaf>(left_id, config_.leaf_capacity);
|
|
auto right = make_leaf();
|
|
const auto middle = combined.size() / 2;
|
|
for (std::size_t i = 0; i < middle; ++i) {
|
|
left->values.push_back(std::move(combined[i]));
|
|
}
|
|
for (std::size_t i = middle; i < combined.size(); ++i) {
|
|
right->values.push_back(std::move(combined[i]));
|
|
}
|
|
|
|
leaves_[leaf_index] = std::move(left);
|
|
leaves_.insert(leaves_.begin() + static_cast<std::ptrdiff_t>(leaf_index + 1),
|
|
std::move(right));
|
|
rebuild_positions();
|
|
refresh_leaf(leaf_index, relocate);
|
|
refresh_leaf(leaf_index + 1, relocate);
|
|
}
|
|
|
|
void retire_leaf(std::size_t leaf_index) {
|
|
const auto id = leaves_[leaf_index]->id;
|
|
if (id < leaf_positions_.size()) {
|
|
leaf_positions_[id] = npos;
|
|
}
|
|
leaves_.erase(leaves_.begin() + static_cast<std::ptrdiff_t>(leaf_index));
|
|
}
|
|
|
|
template <class Relocate>
|
|
bool try_merge(std::size_t leaf_index, Relocate& relocate) {
|
|
if (leaves_.size() < 2) {
|
|
return false;
|
|
}
|
|
const auto threshold = std::max<std::size_t>(1, config_.leaf_capacity / 4);
|
|
if (leaves_[leaf_index]->values.size() >= threshold) {
|
|
return false;
|
|
}
|
|
|
|
std::size_t left_index = leaf_index;
|
|
std::size_t right_index = leaf_index + 1;
|
|
if (right_index >= leaves_.size()) {
|
|
left_index = leaf_index - 1;
|
|
right_index = leaf_index;
|
|
}
|
|
const auto combined_size = leaves_[left_index]->values.size()
|
|
+ leaves_[right_index]->values.size();
|
|
if (combined_size > config_.leaf_capacity) {
|
|
return false;
|
|
}
|
|
|
|
const auto survivor_id = leaves_[left_index]->id;
|
|
auto merged = std::make_unique<Leaf>(survivor_id, config_.leaf_capacity);
|
|
for (std::size_t i = 0; i < leaves_[left_index]->values.size(); ++i) {
|
|
merged->values.push_back(std::move(leaves_[left_index]->values[i]));
|
|
}
|
|
for (std::size_t i = 0; i < leaves_[right_index]->values.size(); ++i) {
|
|
merged->values.push_back(std::move(leaves_[right_index]->values[i]));
|
|
}
|
|
|
|
const auto retired_id = leaves_[right_index]->id;
|
|
leaves_[left_index] = std::move(merged);
|
|
leaves_.erase(leaves_.begin() + static_cast<std::ptrdiff_t>(right_index));
|
|
if (retired_id < leaf_positions_.size()) {
|
|
leaf_positions_[retired_id] = npos;
|
|
}
|
|
rebuild_positions();
|
|
rebuild_directory();
|
|
refresh_leaf(left_index, relocate);
|
|
return true;
|
|
}
|
|
|
|
TieredConfig config_;
|
|
std::vector<std::unique_ptr<Leaf>> leaves_;
|
|
HierarchicalSizeIndex directory_;
|
|
std::size_t size_ = 0;
|
|
leaf_id_type next_leaf_id_ = 0;
|
|
std::vector<std::size_t> leaf_positions_;
|
|
};
|
|
|
|
} // namespace detail
|
|
} // namespace uc
|