infiniflow · JinHai-CN · Dec 16, 2025 · Dec 16, 2025 · Dec 19, 2025 · Dec 22, 2025
diff --git a/.github/workflows/tests.yml b/.github/workflows/tests.yml
@@ -105,7 +105,7 @@ jobs:
         uses: astral-sh/ruff-action@v3
         with:
           version: ">=0.14.x"
-          args: "check"
+          args: "check --exclude=**/remote_thrift/infinity_thrift_rpc/**"
 
       - name: Start builder container
         if: ${{ !cancelled() && !failure() }}
@@ -123,6 +123,7 @@ jobs:
               | grep -E '\.(cpp|h|hpp|cppm)$' \
               | grep -v 'third_party/' \
               | grep -v 'network/' \
+              | grep -v 'remote_thrift/infinity_thrift_rpc/' \
               | grep -v 'parser/' || true)
 
             if [ -n "$CHANGED_FILES" ]; then

diff --git a/.gitignore b/.gitignore
@@ -125,7 +125,8 @@ python/infinity_sdk.egg-info
 sift_1m
 
 # ignore all fvecs benchmark file
-test/data/benchmark/*
+#test/data/benchmark/*
+test/data/benchmark/
 
 # ignore valgrind output file
 callgrind.out.*

diff --git a/.hooks/pre-commit b/.hooks/pre-commit
@@ -2,24 +2,44 @@
 
 import subprocess
 
+
 def get_files():
     return subprocess.check_output(['git', 'diff-index', '--cached',
                                     '--name-only', 'HEAD']).split()
+
+
 def apply_code_style():
     files = filter(lambda x: x.find(b'third_party') == -1, get_files())
+    files = filter(lambda x: x.find(b'src/networker/third_party/infinity_thrift') == -1, files)
+    files = filter(lambda x: x.find(b'src/parser/expression_parser.cpp') == -1, files)
+    files = filter(lambda x: x.find(b'src/parser/expression_parser.h') == -1, files)
+    files = filter(lambda x: x.find(b'src/parser/lexer.cpp') == -1, files)
+    files = filter(lambda x: x.find(b'src/parser/lexer.h') == -1, files)
+    files = filter(lambda x: x.find(b'src/parser/parser.cpp') == -1, files)
+    files = filter(lambda x: x.find(b'src/parser/parser.h') == -1, files)
+    files = filter(lambda x: x.find(b'src/parser/search_parser.cpp') == -1, files)
+    files = filter(lambda x: x.find(b'src/parser/search_parser.h') == -1, files)
 
     files = filter(lambda x: x.endswith(b'.c') or
                              x.endswith(b'.h') or
                              x.endswith(b'.hpp') or
                              x.endswith(b'.cpp') or
-                             x.endswith(b'.cppm'),files)
+                             x.endswith(b'.cppm'), files)
     for f in files:
         print("Apply code style to: " + str(f))
-        subprocess.check_output(['clang-format-20', '-i', f])
-        subprocess.check_output(['git', 'add', f])
+        try:
+            subprocess.check_output(['clang-format-20', '-i', f])
+        except subprocess.CalledProcessError as e:
+            print(f"⚠️ clang-format failed for {f}: {e}")
+        try:
+            subprocess.check_output(['git', 'add', f])
+        except subprocess.CalledProcessError as e:
+            print(f"⚠️ git add failed for {f}: {e}")
+
 
 def main():
     apply_code_style()
 
+
 if (__name__ == '__main__'):
     main()
diff --git a/CMakeLists.txt b/CMakeLists.txt
@@ -1,10 +1,10 @@
-cmake_minimum_required(VERSION 4.0.3...4.2.0)
+cmake_minimum_required(VERSION 4.0.3...4.2.1)
 
 cmake_policy(SET CMP0167 OLD)
 
 set(CMAKE_SUPPRESS_DEVELOPER_WARNINGS TRUE)
 
-if (CMAKE_VERSION GREATER_EQUAL 4.0.3 AND CMAKE_VERSION LESS_EQUAL 4.2.0)
+if (CMAKE_VERSION GREATER_EQUAL 4.0.3 AND CMAKE_VERSION LESS_EQUAL 4.2.1)
     set(CMAKE_EXPERIMENTAL_CXX_IMPORT_STD "d0edc3af-4c50-42ea-a356-e2862fe7a444")
 endif ()
 
@@ -52,10 +52,10 @@ if (CLANG_VERSION_STRING VERSION_GREATER_EQUAL 20)
     message(STATUS "Building ${PROJECT_NAME} with CMake version: ${CMAKE_VERSION} On CLANG-${CLANG_VERSION_STRING}")
 
     # add_compile_options(-ftime-trace)
-    #    add_compile_options(-fmodule-header)
+    # add_compile_options(-fmodule-header)
     add_link_options(-fuse-ld=mold)
+    # set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wall -Wextra -Wno-unused-parameter -Wno-unused-private-field --rtlib=libgcc --unwindlib=libgcc")
     set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wall -Wextra -Wno-unused-parameter -Wno-unused-private-field")
-
     add_link_options(-L/usr/local/lib)
     set(CMAKE_FIND_PACKAGE_SORT_ORDER NATURAL)
     set(CMAKE_FIND_PACKAGE_SORT_DIRECTION DEC)
@@ -74,47 +74,48 @@ elseif (CMAKE_SYSTEM_PROCESSOR MATCHES "^(x86_64|amd64)$")
     set(X86_64 TRUE)
 endif ()
 
-#strength, for image maintainer
-set(GCC_SEARCH_ROOTS
-        /usr
-        /usr/local
-)
+##strength, for image maintainer
+#set(GCC_SEARCH_ROOTS
+#        /usr
+#        /usr/local
+#)
 
 #if you encounter a problem, try commenting out the code below.
-set(GCC15_SUFFIXES
-        #        lib
-        lib/gcc/15
-        lib/gcc/x86_64-linux-gnu/15
-        lib/gcc/x86_64-pc-linux-gnu/15.1.0
-        lib64/gcc/x86_64-linux-gnu/15
-        lib64/gcc/15
-        lib64/gcc/x86_64-pc-linux-gnu/15.1.0
-)
-
-find_library(STDCXX15_STATIC
-        NAMES libstdc++.a
-        PATHS ${GCC_SEARCH_ROOTS}
-        PATH_SUFFIXES ${GCC15_SUFFIXES}
-        REQUIRED
-        NO_DEFAULT_PATH
-)
-
-find_library(STDCXX15EXP_STATIC
-        NAMES libstdc++exp.a
-        PATHS ${GCC_SEARCH_ROOTS}
-        PATH_SUFFIXES ${GCC15_SUFFIXES}
-        REQUIRED
-        NO_DEFAULT_PATH
-)
-
-get_filename_component(GCC15_LIB_DIR ${STDCXX15_STATIC} DIRECTORY)
-
-set(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} -L${GCC15_LIB_DIR} -static-libstdc++ -static-libgcc")
+#set(GCC15_SUFFIXES
+#        #        lib
+#        lib/gcc/15
+#	    lib/gcc/aarch64-linux-gnu/15
+#        lib/gcc/x86_64-linux-gnu/15
+#        lib/gcc/x86_64-pc-linux-gnu/15.1.0
+#        lib64/gcc/x86_64-linux-gnu/15
+#        lib64/gcc/15
+#        lib64/gcc/x86_64-pc-linux-gnu/15.1.0
+#)
+
+# find_library(GCC15_STATIC
+#         NAMES libgcc.a
+#         PATHS ${GCC_SEARCH_ROOTS}
+#         PATH_SUFFIXES ${GCC15_SUFFIXES}
+#         REQUIRED
+#         NO_DEFAULT_PATH
+# )
+
+#find_library(STDCXX15EXP_STATIC
+#        NAMES libstdc++exp.a
+#        PATHS ${GCC_SEARCH_ROOTS}
+#        PATH_SUFFIXES ${GCC15_SUFFIXES}
+#        REQUIRED
+#        NO_DEFAULT_PATH
+#)
+
+#get_filename_component(GCC15_LIB_DIR "${STDCXX15_STATIC}" DIRECTORY)
+
+set(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} -static-libstdc++ -static-libgcc")
 ########
 
 find_package(absl CONFIG REQUIRED)
 find_package(Arrow CONFIG REQUIRED)
-find_package(Boost REQUIRED COMPONENTS asio)
+find_package(Boost REQUIRED COMPONENTS asio thread)
 find_package(CLI11 CONFIG REQUIRED)
 #find_package(CURL REQUIRED) # wait for s3
 #darts-clone # wait for fix
@@ -142,9 +143,19 @@ find_package(simdjson CONFIG REQUIRED)
 #find_package(spdlog REQUIRED) # wait for fmt:11.2.0
 if (ARM64)
     add_library(sse2neon INTERFACE)
-    # need this after highway and before simdcomp
-    find_path(SSE2NEON_INCLUDE_DIRS "sse2neon/sse2neon.h")
-    target_include_directories(sse2neon SYSTEM ${SSE2NEON_INCLUDE_DIRS})
+    if (EXISTS "${VCPKG_INSTALLED_DIR}/include/sse2neon/sse2neon.h")
+        set(SSE2NEON_INCLUDE_DIRS "${VCPKG_INSTALLED_DIR}/include")
+        message(STATUS "Found sse2neon in vcpkg: ${SSE2NEON_INCLUDE_DIRS}")
+        target_include_directories(sse2neon SYSTEM INTERFACE ${SSE2NEON_INCLUDE_DIRS})
+    else ()
+        find_path(SSE2NEON_INCLUDE_DIRS "sse2neon/sse2neon.h")
+        if (SSE2NEON_INCLUDE_DIRS)
+            message(STATUS "Found sse2neon: ${SSE2NEON_INCLUDE_DIRS}")
+            target_include_directories(sse2neon SYSTEM INTERFACE ${SSE2NEON_INCLUDE_DIRS})
+        else ()
+            message(WARNING "sse2neon not found, proceeding without it.")
+        endif ()
+    endif ()
 endif ()
 find_package(tomlplusplus CONFIG REQUIRED)
 #find_package(turbobase64 CONFIG REQUIRED)
@@ -318,16 +329,9 @@ elseif ("${CMAKE_BUILD_TYPE}" STREQUAL "Debug")
     endif ()
 
     set(CMAKE_DEBUG_POSTFIX "")
-
-else ()
-    message(FATAL_ERROR "Only support CMake build type: Debug, RelWithDebInfo, and Release")
 endif ()
 
-if (CLANG_VERSION_STRING VERSION_EQUAL 17)
-    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wall -Werror -Wno-asm-operand-widths -Wno-unused-command-line-argument -Wno-deprecated-declarations -Wno-read-modules-implicitly -Wextra -Wno-unused-parameter -Wno-unused-private-field -pthread -fcolor-diagnostics")
-else ()
-    set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wall -Werror -Wno-asm-operand-widths -Wno-unused-command-line-argument -Wno-deprecated-declarations -Wextra -Wno-unused-parameter -Wno-unused-private-field -pthread -fcolor-diagnostics")
-endif ()
+set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS} -Wall -Werror -Wno-asm-operand-widths -Wno-unused-command-line-argument -Wno-deprecated-declarations -Wextra -Wno-unused-parameter -Wno-unused-private-field -pthread -fcolor-diagnostics")
 
 MESSAGE(STATUS "C++ Compilation flags: " ${CMAKE_CXX_FLAGS})
 
@@ -386,7 +390,6 @@ if (X86_64)
 else ()
     add_definitions(-march=native)
 endif ()
-
 execute_process(
         COMMAND bash -c "zgrep CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y /proc/config.gz 2>/dev/null; grep CONFIG_HAVE_EFFICIENT_UNALIGNED_ACCESS=y /boot/config-$(uname -r) 2>/dev/null"
         OUTPUT_VARIABLE HAVE_EFFICIENT_UNALIGNED_ACCESS

diff --git a/benchmark/local_infinity/CMakeLists.txt b/benchmark/local_infinity/CMakeLists.txt
@@ -15,6 +15,8 @@ target_link_libraries(infinity_benchmark
         jma
         opencc
         dl
+        Boost::asio
+        Boost::thread
         thrift::thrift
         # thriftnb::thriftnb
         libevent::core
@@ -39,7 +41,8 @@ target_link_libraries(infinity_benchmark
         CLI11::CLI11
         magic_enum::magic_enum
         roaring::roaring
-        ${STDCXX15EXP_STATIC}
+#        ${STDCXX15EXP_STATIC}
+        libstdc++exp.a
 )
 
 target_link_directories(infinity_benchmark PUBLIC "${CMAKE_BINARY_DIR}/lib")
@@ -68,6 +71,8 @@ target_link_libraries(knn_import_benchmark
         jma
         opencc
         dl
+        Boost::asio
+        Boost::thread
         thrift::thrift
         # thriftnb::thriftnb
         libevent::core
@@ -92,7 +97,8 @@ target_link_libraries(knn_import_benchmark
         CLI11::CLI11
         magic_enum::magic_enum
         roaring::roaring
-        ${STDCXX15EXP_STATIC}
+#        ${STDCXX15EXP_STATIC}
+        libstdc++exp.a
 )
 
 target_link_directories(knn_import_benchmark PUBLIC "${CMAKE_BINARY_DIR}/lib")
@@ -119,6 +125,8 @@ target_link_libraries(knn_query_benchmark
         jma
         opencc
         dl
+        Boost::asio
+        Boost::thread
         Parquet::parquet_static
         Arrow::arrow_static
         thrift::thrift
@@ -143,7 +151,8 @@ target_link_libraries(knn_query_benchmark
         CLI11::CLI11
         magic_enum::magic_enum
         roaring::roaring
-        ${STDCXX15EXP_STATIC}
+#        ${STDCXX15EXP_STATIC}
+        libstdc++exp.a
 )
 
 target_link_directories(knn_query_benchmark PUBLIC "${CMAKE_BINARY_DIR}/lib")
@@ -171,6 +180,8 @@ target_link_libraries(fulltext_benchmark
         jma
         opencc
         dl
+        Boost::asio
+        Boost::thread
         thrift::thrift
         # thriftnb::thriftnb
         libevent::core
@@ -195,7 +206,8 @@ target_link_libraries(fulltext_benchmark
         CLI11::CLI11
         magic_enum::magic_enum
         roaring::roaring
-        ${STDCXX15EXP_STATIC}
+#        ${STDCXX15EXP_STATIC}
+        libstdc++exp.a
 )
 
 target_link_directories(fulltext_benchmark PUBLIC "${CMAKE_BINARY_DIR}/lib")
@@ -221,6 +233,8 @@ target_link_libraries(sparse_benchmark
         jma
         opencc
         dl
+        Boost::asio
+        Boost::thread
         Parquet::parquet_static
         Arrow::arrow_static
         ${JEMALLOC_STATIC_LIB}
@@ -241,7 +255,8 @@ target_link_libraries(sparse_benchmark
         CLI11::CLI11
         magic_enum::magic_enum
         roaring::roaring
-        ${STDCXX15EXP_STATIC}
+#        ${STDCXX15EXP_STATIC}
+        libstdc++exp.a
 )
 
 target_link_directories(sparse_benchmark PUBLIC "${CMAKE_BINARY_DIR}/lib")
@@ -266,6 +281,8 @@ target_link_libraries(bmp_benchmark
         jma
         opencc
         dl
+        Boost::asio
+        Boost::thread
         thrift::thrift
         # thriftnb::thriftnb
         libevent::core
@@ -290,7 +307,8 @@ target_link_libraries(bmp_benchmark
         CLI11::CLI11
         magic_enum::magic_enum
         roaring::roaring
-        ${STDCXX15EXP_STATIC}
+#        ${STDCXX15EXP_STATIC}
+        libstdc++exp.a
 )
 
 target_link_directories(bmp_benchmark PUBLIC "${CMAKE_BINARY_DIR}/lib")
@@ -315,6 +333,8 @@ target_link_libraries(hnsw_benchmark
         jma
         opencc
         dl
+        Boost::asio
+        Boost::thread
         Parquet::parquet_static
         Arrow::arrow_static
         thrift::thrift
@@ -339,7 +359,8 @@ target_link_libraries(hnsw_benchmark
         CLI11::CLI11
         magic_enum::magic_enum
         roaring::roaring
-        ${STDCXX15EXP_STATIC}
+#        ${STDCXX15EXP_STATIC}
+        libstdc++exp.a
 )
 
 target_link_directories(hnsw_benchmark PUBLIC "${CMAKE_BINARY_DIR}/lib")

diff --git a/benchmark/local_infinity/knn/knn_query_benchmark.cpp b/benchmark/local_infinity/knn/knn_query_benchmark.cpp
@@ -219,9 +219,9 @@ int main(int argc, char *argv[]) {
                 infinity
                     ->Search(db_name, table_name, search_expr, nullptr, nullptr, nullptr, output_columns, nullptr, nullptr, nullptr, nullptr, false);
             {
-                auto &cv = result.result_table_->GetDataBlockById(0)->column_vectors;
+                auto &cv = result.result_table_->GetDataBlockById(0)->column_vectors_;
                 auto &column = *cv[0];
-                auto data = reinterpret_cast<const RowID *>(column.data());
+                auto data = reinterpret_cast<const RowID *>(column.data().get());
                 auto cnt = column.Size();
                 for (size_t i = 0; i < cnt; ++i) {
                     query_results[query_idx].emplace_back(data[i].ToUint64());
@@ -286,12 +286,12 @@ int main(int argc, char *argv[]) {
     std::shared_ptr<Infinity> infinity = Infinity::LocalConnect();
     QueryResult cache_result = infinity->ShowCache();
 
-    auto &vectors = cache_result.result_table_->GetDataBlockById(0)->column_vectors;
+    auto &vectors = cache_result.result_table_->GetDataBlockById(0)->column_vectors_;
     std::cout << "columns: " << vectors.size() << std::endl;
-    auto column1 = reinterpret_cast<const u64 *>(vectors[1]->data());
-    auto column2 = reinterpret_cast<const u64 *>(vectors[2]->data());
-    auto column3 = reinterpret_cast<const u64 *>(vectors[3]->data());
-    auto column4 = reinterpret_cast<const double *>(vectors[4]->data());
+    auto column1 = reinterpret_cast<const u64 *>(vectors[1]->data().get());
+    auto column2 = reinterpret_cast<const u64 *>(vectors[2]->data().get());
+    auto column3 = reinterpret_cast<const u64 *>(vectors[3]->data().get());
+    auto column4 = reinterpret_cast<const double *>(vectors[4]->data().get());
 
     std::cout << "Cache db, items: " << column1[0] << ", hits: " << column2[0] << ", request: " << column3[0] << ", hit rate: " << column4[0]
               << std::endl;