mandlm/VectorSearch
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

refactor: integrate finder-threads as lambdas

Browse source
This commit is contained in:
Michael Mandl 2024-03-23 11:24:06 +01:00
parent 2ebc2107ba
commit c1b4fc67e9
Signed by: mandlm
GPG key ID: 4AA25D647AA54CC7
4 changed files with 50 additions and 82 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

15
lib_vector_search/include/word_list.h
View file

@ -1,12 +1,9 @@
#pragma once
#include <filesystem>
#include <mutex>
#include <string>
#include <vector>
class WordRefList;
class WordList : public std::vector<std::string> {
public:
WordList &multiply(size_t factor);
@ -15,22 +12,10 @@ public:
static WordList oneCap();
static WordList fourCaps();
static WordList fromFile(const std::filesystem::path &path);
static void find_prefix_in_range(const WordList &word_list,
const std::string_view &search_prefix,
size_t start_index, size_t end_index,
WordRefList &result,
std::mutex &result_mutex);
};
class WordRefList : public std::vector<const std::string *> {
public:
WordRefList() = default;
WordRefList(const WordList &source);
static void find_prefix_in_range(const WordRefList &word_list,
const std::string_view &search_prefix,
size_t start_index, size_t end_index,
WordRefList &result,
std::mutex &result_mutex);
};

40
lib_vector_search/src/grouped_finder.cpp
View file

@ -19,31 +19,43 @@ WordRefList GroupedFinder::find_prefix(string_view search_prefix) const {
}
const auto word_list = group->second;
const auto word_list_size = word_list.size();
const auto thread_count =
const size_t word_list_size = word_list.size();
const size_t thread_count =
std::min<size_t>(std::thread::hardware_concurrency(), word_list_size);
const size_t words_per_thread = word_list_size / thread_count;
WordRefList result;
mutex result_mutex;
WordRefList search_results;
mutex search_results_mutex;
vector<thread> search_threads;
for (size_t thread_index = 0; thread_index < thread_count; ++thread_index) {
const size_t first_index = thread_index * (word_list_size / thread_count);
const bool is_last_thread = thread_index == thread_count - 1;
const size_t start_index = thread_index * words_per_thread;
const size_t end_index =
is_last_thread ? word_list_size : start_index + words_per_thread;
const size_t last_index =
(thread_index == thread_count - 1)
? word_list_size
: (thread_index + 1) * (word_list_size / thread_count);
search_threads.emplace_back([&, start_index, end_index] {
WordRefList local_results;
search_threads.emplace_back(
WordRefList::find_prefix_in_range, cref(word_list), cref(search_prefix),
first_index, last_index, ref(result), ref(result_mutex));
for (size_t word_index = start_index; word_index < end_index;
++word_index) {
const auto *current_word = word_list[word_index];
if (current_word->starts_with(search_prefix)) {
local_results.push_back(current_word);
}
}
if (!local_results.empty()) {
const std::lock_guard lock(search_results_mutex);
std::move(local_results.begin(), local_results.end(),
std::back_inserter(search_results));
}
});
}
for (auto &thread : search_threads) {
thread.join();
}
return result;
return search_results;
}

35
lib_vector_search/src/parallel_finder.cpp
View file

@ -9,30 +9,43 @@ ParallelFinder::ParallelFinder(const WordList &word_list)
: word_list_(word_list) {}
WordRefList ParallelFinder::find_prefix(string_view search_prefix) const {
WordRefList result;
mutex result_mutex;
WordRefList search_results;
mutex search_results_mutex;
const size_t word_list_size = word_list_.size();
const size_t thread_count =
std::min<size_t>(thread::hardware_concurrency(), word_list_size);
const size_t words_per_thread = word_list_size / thread_count;
vector<thread> search_threads;
for (size_t thread_index = 0; thread_index < thread_count; ++thread_index) {
const size_t first_index = thread_index * (word_list_size / thread_count);
const bool is_last_thread = thread_index == thread_count - 1;
const size_t start_index = thread_index * words_per_thread;
const size_t end_index =
is_last_thread ? word_list_size : start_index + words_per_thread;
const size_t last_index =
(thread_index == thread_count - 1)
? word_list_size
: (thread_index + 1) * (word_list_size / thread_count);
search_threads.emplace_back([&, start_index, end_index] {
WordRefList local_results;
search_threads.emplace_back(
WordList::find_prefix_in_range, cref(word_list_), cref(search_prefix),
first_index, last_index, ref(result), ref(result_mutex));
for (size_t word_index = start_index; word_index < end_index;
++word_index) {
const auto &current_word = word_list_[word_index];
if (current_word.starts_with(search_prefix)) {
local_results.push_back(&current_word);
}
}
if (!local_results.empty()) {
const std::lock_guard lock(search_results_mutex);
std::move(local_results.begin(), local_results.end(),
std::back_inserter(search_results));
}
});
}
for (auto &thread : search_threads) {
thread.join();
}
return result;
return search_results;
}

42
lib_vector_search/src/word_list.cpp
View file

@ -71,50 +71,8 @@ WordList WordList::fromFile(const std::filesystem::path &path) {
return word_list;
}
void WordList::find_prefix_in_range(const WordList &word_list,
const std::string_view &search_prefix,
size_t start_index, size_t end_index,
WordRefList &result,
std::mutex &result_mutex) {
WordRefList local_results;
for (size_t index = start_index; index < end_index; ++index) {
const auto &current_word = word_list[index];
if (current_word.starts_with(search_prefix)) {
local_results.push_back(&current_word);
}
}
if (!local_results.empty()) {
const std::lock_guard<std::mutex> lock(result_mutex);
std::move(local_results.begin(), local_results.end(),
std::back_inserter(result));
}
};
WordRefList::WordRefList(const WordList &source) {
for (const auto &word : source) {
push_back(&word);
}
}
void WordRefList::find_prefix_in_range(const WordRefList &word_list,
const std::string_view &search_prefix,
size_t start_index, size_t end_index,
WordRefList &result,
std::mutex &result_mutex) {
WordRefList local_results;
for (size_t index = start_index; index < end_index; ++index) {
const auto *current_word = word_list[index];
if (current_word->starts_with(search_prefix)) {
local_results.push_back(current_word);
}
}
if (!local_results.empty()) {
const std::lock_guard<std::mutex> lock(result_mutex);
std::move(local_results.begin(), local_results.end(),
std::back_inserter(result));
}
};