Skip to content

Support UTF8 in fulltext analyzers #26850

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 4 commits into from
Oct 15, 2025
Merged

Support UTF8 in fulltext analyzers #26850

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
faab253
Support russian in fulltext index
kunga Oct 14, 2025
469f979
add digits test
kunga Oct 14, 2025
af15a68
fix invalid cases
kunga Oct 15, 2025
16d5030
simplify IsNonStandard
kunga Oct 15, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
75 changes: 53 additions & 22 deletions ydb/core/base/fulltext.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,5 +1,6 @@
#include "fulltext.h"
#include <regex>
#include <util/charset/utf8.h>
#include <util/generic/xrange.h>

namespace NKikimr::NFulltext {

Expand Down Expand Up @@ -45,35 +46,65 @@ namespace {
return result;
}

// Note: written by llm, can be optimized a lot later
TVector<TString> Tokenize(const TString& text, const Ydb::Table::FulltextIndexSettings::Tokenizer& tokenizer) {
TVector<TString> tokens;
switch (tokenizer) {
case Ydb::Table::FulltextIndexSettings::WHITESPACE: {
std::istringstream stream(text);
TString token;
while (stream >> token) {
tokens.push_back(token);
inline bool IsNonStandard(wchar32 c) {
return !IsAlphabetic(c) && !IsDecdigit(c);
}

void Tokenize(const TString& text, TVector<TString>& tokens, auto isDelimiter) {
const unsigned char* ptr = (const unsigned char*)text.data();
const unsigned char* end = ptr + text.size();

while (ptr < end) {
wchar32 symbol;
size_t symbolBytes = 0;

while (ptr < end) { // skip delimiters
if (SafeReadUTF8Char(symbol, symbolBytes, ptr, end) != RECODE_OK) {
tokens.clear();
return;
}
if (!isDelimiter(symbol)) {
break;
}
ptr += symbolBytes;
}
if (ptr >= end) {
break;
}
case Ydb::Table::FulltextIndexSettings::STANDARD: {
std::regex word_regex(R"(\b\w+\b)"); // match alphanumeric words
std::sregex_iterator it(text.begin(), text.end(), word_regex);
std::sregex_iterator end;
while (it != end) {
tokens.push_back(it->str());
++it;

const unsigned char* tokenPtr = ptr;
while (ptr < end) { // read token
if (SafeReadUTF8Char(symbol, symbolBytes, ptr, end) != RECODE_OK) {
tokens.clear();
return;
}
break;
if (isDelimiter(symbol)) {
break;
}
ptr += symbolBytes;
}
tokens.emplace_back((const char*)tokenPtr, ptr - tokenPtr);
}
}


TVector<TString> Tokenize(const TString& text, const Ydb::Table::FulltextIndexSettings::Tokenizer& tokenizer) {
TVector<TString> tokens;
switch (tokenizer) {
case Ydb::Table::FulltextIndexSettings::WHITESPACE:
Tokenize(text, tokens, IsWhitespace);
break;
case Ydb::Table::FulltextIndexSettings::STANDARD:
Tokenize(text, tokens, IsNonStandard);
break;
case Ydb::Table::FulltextIndexSettings::KEYWORD:
tokens.push_back(text);
if (UTF8Detect(text) != NotUTF8) {
tokens.push_back(text);
}
break;
default:
Y_ENSURE(TStringBuilder() << "Invalid tokenizer: " << static_cast<int>(tokenizer));
}

return tokens;
}

Expand Down Expand Up @@ -131,8 +162,8 @@ TVector<TString> Analyze(const TString& text, const Ydb::Table::FulltextIndexSet
TVector<TString> tokens = Tokenize(text, settings.tokenizer());

if (settings.use_filter_lowercase()) {
for (auto& token : tokens) {
token.to_lower();
for (auto i : xrange(tokens.size())) {
tokens[i] = ToLowerUTF8(tokens[i]);
}
}

Expand Down
57 changes: 53 additions & 4 deletions ydb/core/base/ut/fulltext_ut.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,6 +1,7 @@
#include "fulltext.h"

#include <library/cpp/testing/unittest/registar.h>
#include <util/generic/xrange.h>

namespace NKikimr::NFulltext {

Expand Down Expand Up @@ -100,20 +101,68 @@ Y_UNIT_TEST_SUITE(NFulltext) {

Y_UNIT_TEST(Analyze) {
Ydb::Table::FulltextIndexSettings::Analyzers analyzers;
TString text = "apple WaLLet spaced-dog";
TString text = "apple WaLLet spaced-dog_cat 0123,456@";

analyzers.set_tokenizer(Ydb::Table::FulltextIndexSettings::WHITESPACE);
UNIT_ASSERT_VALUES_EQUAL(Analyze(text, analyzers), (TVector<TString>{"apple", "WaLLet", "spaced-dog"}));
UNIT_ASSERT_VALUES_EQUAL(Analyze(text, analyzers), (TVector<TString>{"apple", "WaLLet", "spaced-dog_cat", "0123,456@"}));

analyzers.set_tokenizer(Ydb::Table::FulltextIndexSettings::STANDARD);
UNIT_ASSERT_VALUES_EQUAL(Analyze(text, analyzers), (TVector<TString>{"apple", "WaLLet", "spaced", "dog"}));
UNIT_ASSERT_VALUES_EQUAL(Analyze(text, analyzers), (TVector<TString>{"apple", "WaLLet", "spaced", "dog", "cat", "0123", "456"}));

analyzers.set_tokenizer(Ydb::Table::FulltextIndexSettings::KEYWORD);
UNIT_ASSERT_VALUES_EQUAL(Analyze(text, analyzers), (TVector<TString>{text}));

analyzers.set_tokenizer(Ydb::Table::FulltextIndexSettings::WHITESPACE);
analyzers.set_use_filter_lowercase(true);
UNIT_ASSERT_VALUES_EQUAL(Analyze(text, analyzers), (TVector<TString>{"apple", "wallet", "spaced-dog"}));
UNIT_ASSERT_VALUES_EQUAL(Analyze(text, analyzers), (TVector<TString>{"apple", "wallet", "spaced-dog_cat", "0123,456@"}));
}

Y_UNIT_TEST(AnalyzeRu) {
Ydb::Table::FulltextIndexSettings::Analyzers analyzers;
TString text = "Привет, это test123 и слово Ёлка ёль!";

analyzers.set_tokenizer(Ydb::Table::FulltextIndexSettings::WHITESPACE);
UNIT_ASSERT_VALUES_EQUAL(Analyze(text, analyzers), (TVector<TString>{"Привет,", "это", "test123", "и", "слово", "Ёлка", "ёль!"}));

analyzers.set_tokenizer(Ydb::Table::FulltextIndexSettings::STANDARD);
UNIT_ASSERT_VALUES_EQUAL(Analyze(text, analyzers), (TVector<TString>{"Привет", "это", "test123", "и", "слово", "Ёлка", "ёль"}));

analyzers.set_tokenizer(Ydb::Table::FulltextIndexSettings::KEYWORD);
UNIT_ASSERT_VALUES_EQUAL(Analyze(text, analyzers), (TVector<TString>{text}));

analyzers.set_tokenizer(Ydb::Table::FulltextIndexSettings::STANDARD);
analyzers.set_use_filter_lowercase(true);
UNIT_ASSERT_VALUES_EQUAL(Analyze(text, analyzers), (TVector<TString>{"привет", "это", "test123", "и", "слово", "ёлка", "ёль"}));
}

Y_UNIT_TEST(AnalyzeInvalid) {
Ydb::Table::FulltextIndexSettings::Analyzers analyzers;

TVector<TString> texts = {
"\xC2\x41", // Invalid continuation byte
"\xC0\x81", // Overlong encoding
"\x80", // Lone continuation byte
"\xF4\x90\x80\x80", // Outside Unicode range
"\xE3\x81", // Truncated (incomplete)
};

for (auto i : xrange(texts.size())) {
TString testCase = TStringBuilder() << "case #" << i;
auto& text = texts[i];

analyzers.set_tokenizer(Ydb::Table::FulltextIndexSettings::WHITESPACE);
UNIT_ASSERT_VALUES_EQUAL_C(Analyze(text, analyzers), (TVector<TString>{}), testCase);

analyzers.set_tokenizer(Ydb::Table::FulltextIndexSettings::STANDARD);
UNIT_ASSERT_VALUES_EQUAL_C(Analyze(text, analyzers), (TVector<TString>{}), testCase);

analyzers.set_tokenizer(Ydb::Table::FulltextIndexSettings::KEYWORD);
UNIT_ASSERT_VALUES_EQUAL_C(Analyze(text, analyzers), (TVector<TString>{}), testCase);

analyzers.set_tokenizer(Ydb::Table::FulltextIndexSettings::KEYWORD);
analyzers.set_use_filter_lowercase(true);
UNIT_ASSERT_VALUES_EQUAL_C(Analyze(text, analyzers), (TVector<TString>{}), testCase);
}
}
}

Expand Down
Loading