mandlm/VectorSearch
Archived
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Compare commits

base: mandlm:1441a869c2388eca1b7e073c97947e962d1d5c2a
Branches Tags
mandlm:main
..
compare: mandlm:532e52cc7f035fefb6fd435e9f9b7dcc85f08487
Branches Tags
mandlm:main

No commits in common. "1441a869c2388eca1b7e073c97947e962d1d5c2a" and "532e52cc7f035fefb6fd435e9f9b7dcc85f08487" have entirely different histories.
1441a869c2 ... 532e52cc7f

2 changed files with 5 additions and 15 deletions
Show Stats Expand all files Collapse all files

6
lib_vector_search/include/bucket_finder.h
View File

@ -5,12 +5,6 @@
#include <unordered_map> #include <unordered_map>
/** A Bucket contains a partial WordRefList, split-up and (hash-)mapped by their
* first characters.
*
* It's meant to be read and written by a single thread without the need for
* synchronization.
*/
class Bucket { class Bucket {
private: private:
std::unordered_map<char, WordRefList> directory_; std::unordered_map<char, WordRefList> directory_;

14
lib_vector_search/src/bucket_finder.cpp
View File

@ -7,8 +7,6 @@
#include <thread> #include <thread>
#include <vector> #include <vector>
// class Bucket
void Bucket::insert(const WordList &word_list, size_t first_index, void Bucket::insert(const WordList &word_list, size_t first_index,
size_t last_index) { size_t last_index) {
for (auto index = first_index; index < last_index; ++index) { for (auto index = first_index; index < last_index; ++index) {
@ -33,8 +31,6 @@ WordRefList Bucket::find_prefix(std::string_view search_term) const {
return result; return result;
} }
// class BucketFinder
BucketFinder::BucketFinder(const WordList &word_list) { insert(word_list); } BucketFinder::BucketFinder(const WordList &word_list) { insert(word_list); }
void BucketFinder::insert(const WordList &word_list) { void BucketFinder::insert(const WordList &word_list) {
@ -51,7 +47,7 @@ void BucketFinder::insert(const WordList &word_list) {
std::vector<std::thread> insert_threads; std::vector<std::thread> insert_threads;
for (auto bucket_index = 0; bucket_index < bucket_count; ++bucket_index) { for (auto bucket_index = 0; bucket_index < bucket_count; ++bucket_index) {
auto &thread_bucket = buckets_[bucket_index]; auto &bucket = buckets_[bucket_index];
const bool is_last_bucket = bucket_index == bucket_count - 1; const bool is_last_bucket = bucket_index == bucket_count - 1;
const size_t first_word_index = bucket_index * bucket_size; const size_t first_word_index = bucket_index * bucket_size;
@ -59,7 +55,7 @@ void BucketFinder::insert(const WordList &word_list) {
is_last_bucket ? word_list_size : first_word_index + bucket_size; is_last_bucket ? word_list_size : first_word_index + bucket_size;
insert_threads.emplace_back([&, first_word_index, last_word_index] { insert_threads.emplace_back([&, first_word_index, last_word_index] {
thread_bucket.insert(word_list, first_word_index, last_word_index); bucket.insert(word_list, first_word_index, last_word_index);
}); });
} }
@ -72,9 +68,9 @@ WordRefList BucketFinder::find_prefix(std::string_view search_term) const {
WordRefList search_results; WordRefList search_results;
std::mutex search_results_mutex; std::mutex search_results_mutex;
std::vector<std::thread> search_threads; std::vector<std::thread> threads;
for (const auto &bucket : buckets_) { for (const auto &bucket : buckets_) {
search_threads.emplace_back([&] { threads.emplace_back([&] {
auto thread_search_results = bucket.find_prefix(search_term); auto thread_search_results = bucket.find_prefix(search_term);
if (!thread_search_results.empty()) { if (!thread_search_results.empty()) {
const std::lock_guard result_lock(search_results_mutex); const std::lock_guard result_lock(search_results_mutex);
@ -84,7 +80,7 @@ WordRefList BucketFinder::find_prefix(std::string_view search_term) const {
}); });
} }
for (auto &thread : search_threads) { for (auto &thread : threads) {
thread.join(); thread.join();
} }