From b2ff2f55218707f01cdb014c03ed5079806f8537 Mon Sep 17 00:00:00 2001
From: PHILO-HE <feilong.he@intel.com>
Date: Thu, 29 Aug 2024 08:56:29 +0800
Subject: [PATCH] [GLUTEN-7049][VL] Install lib stemmer through vcpkg (#7050)

---
 cpp/velox/CMakeLists.txt                  | 19 ++++++++++++++----
 dev/vcpkg/ports/libstemmer/CMakeLists.txt | 22 +++++++++++++++++++++
 dev/vcpkg/ports/libstemmer/portfile.cmake | 24 +++++++++++++++++++++++
 dev/vcpkg/ports/libstemmer/vcpkg.json     | 12 ++++++++++++
 dev/vcpkg/vcpkg.json                      |  3 ++-
 5 files changed, 75 insertions(+), 5 deletions(-)
 create mode 100644 dev/vcpkg/ports/libstemmer/CMakeLists.txt
 create mode 100644 dev/vcpkg/ports/libstemmer/portfile.cmake
 create mode 100644 dev/vcpkg/ports/libstemmer/vcpkg.json

diff --git a/cpp/velox/CMakeLists.txt b/cpp/velox/CMakeLists.txt
index 202863b2a2e9..1d7e8548aa04 100644
--- a/cpp/velox/CMakeLists.txt
+++ b/cpp/velox/CMakeLists.txt
@@ -299,12 +299,22 @@ target_link_libraries(velox PUBLIC Folly::folly)
 find_re2()
 target_link_libraries(velox PUBLIC ${RE2_LIBRARY})
 
-import_library(external::stemmer
-               ${VELOX_BUILD_PATH}/_deps/libstemmer/src/libstemmer/libstemmer.a)
-target_link_libraries(velox PUBLIC external::stemmer)
-
 set(CMAKE_FIND_LIBRARY_SUFFIXES_BCK ${CMAKE_FIND_LIBRARY_SUFFIXES})
 set(CMAKE_FIND_LIBRARY_SUFFIXES ".a")
+
+find_library(STEMMER_LIB NAMES libstemmer.a)
+if(STEMMER_LIB)
+  message(STATUS "Found stemmer: ${STEMMER_LIB}")
+  add_library(external::stemmer STATIC IMPORTED)
+  set_target_properties(external::stemmer PROPERTIES IMPORTED_LOCATION
+                                                     ${STEMMER_LIB})
+else()
+  import_library(
+    external::stemmer
+    ${VELOX_BUILD_PATH}/_deps/libstemmer/src/libstemmer/libstemmer.a)
+endif()
+target_link_libraries(velox PUBLIC external::stemmer)
+
 find_package(simdjson CONFIG)
 if(simdjson_FOUND AND TARGET simdjson::simdjson)
   target_link_libraries(velox PUBLIC simdjson::simdjson)
@@ -313,6 +323,7 @@ else()
                  ${VELOX_BUILD_PATH}/_deps/simdjson-build/libsimdjson.a)
   target_link_libraries(velox PUBLIC external::simdjson)
 endif()
+
 set(CMAKE_FIND_LIBRARY_SUFFIXES ${CMAKE_FIND_LIBRARY_SUFFIXES_BCK})
 
 if(BUILD_TESTS)
diff --git a/dev/vcpkg/ports/libstemmer/CMakeLists.txt b/dev/vcpkg/ports/libstemmer/CMakeLists.txt
new file mode 100644
index 000000000000..787bb26ec3f7
--- /dev/null
+++ b/dev/vcpkg/ports/libstemmer/CMakeLists.txt
@@ -0,0 +1,22 @@
+cmake_minimum_required(VERSION 3.8.0)
+
+project(libstemmer)
+
+file(GLOB SNOWBALL_SOURCES
+  src_c/*.c
+  runtime/api.c
+  runtime/utilities.c
+  libstemmer/libstemmer.c
+)
+
+add_library(stemmer ${SNOWBALL_SOURCES})
+
+install(
+  TARGETS stemmer
+  RUNTIME DESTINATION bin
+  LIBRARY DESTINATION lib
+  ARCHIVE DESTINATION lib
+)
+if(NOT DISABLE_INSTALL_HEADERS)
+  install(FILES include/libstemmer.h DESTINATION include)
+endif()
diff --git a/dev/vcpkg/ports/libstemmer/portfile.cmake b/dev/vcpkg/ports/libstemmer/portfile.cmake
new file mode 100644
index 000000000000..253027ad72e0
--- /dev/null
+++ b/dev/vcpkg/ports/libstemmer/portfile.cmake
@@ -0,0 +1,24 @@
+vcpkg_check_linkage(ONLY_STATIC_LIBRARY)
+string(SUBSTRING "${VERSION}" 5 -1 VERSION)
+
+vcpkg_download_distfile(ARCHIVE
+    URLS "https://snowballstem.org/dist/libstemmer_c-${VERSION}.tar.gz"
+    FILENAME "libstemmer_c-${VERSION}.tar.gz"
+    SHA512 a61a06a046a6a5e9ada12310653c71afb27b5833fa9e21992ba4bdf615255de991352186a8736d0156ed754248a0ffb7b7712e8d5ea16f75174d1c8ddd37ba4a
+)
+
+vcpkg_extract_source_archive(
+    SOURCE_PATH
+    ARCHIVE "${ARCHIVE}"
+)
+
+file(COPY "${CMAKE_CURRENT_LIST_DIR}/CMakeLists.txt" DESTINATION "${SOURCE_PATH}")
+
+vcpkg_cmake_configure(
+    SOURCE_PATH "${SOURCE_PATH}"
+    OPTIONS_DEBUG -DDISABLE_INSTALL_HEADERS=ON
+)
+
+vcpkg_cmake_install()
+
+vcpkg_install_copyright(FILE_LIST "${SOURCE_PATH}/COPYING")
diff --git a/dev/vcpkg/ports/libstemmer/vcpkg.json b/dev/vcpkg/ports/libstemmer/vcpkg.json
new file mode 100644
index 000000000000..03e8275d0b25
--- /dev/null
+++ b/dev/vcpkg/ports/libstemmer/vcpkg.json
@@ -0,0 +1,12 @@
+{
+  "name": "libstemmer",
+  "version": "2021.2.2.0",
+  "description": "Snowball is a small string processing language designed for creating stemming algorithms for use in Information Retrieval",
+  "homepage": "https://snowballstem.org/",
+  "dependencies": [
+    {
+      "name": "vcpkg-cmake",
+      "host": true
+    }
+  ]
+}
diff --git a/dev/vcpkg/vcpkg.json b/dev/vcpkg/vcpkg.json
index 6eec2190d043..618293ff63f2 100644
--- a/dev/vcpkg/vcpkg.json
+++ b/dev/vcpkg/vcpkg.json
@@ -52,7 +52,8 @@
         "benchmark",
         "jemalloc",
         "icu",
-        "thrift"
+        "thrift",
+        "libstemmer"
       ]
     },
     "velox-s3": {