Revert "Add word_stem Presto function (facebookincubator#9363)"

This reverts commit 4797041.
zhli1142015 · Jun 1, 2024 · b0698b1 · b0698b1
1 parent ed35a5e
commit b0698b1
Show file tree

Hide file tree

Showing 10 changed files with 1 addition and 343 deletions.
diff --git a/CMake/resolve_dependency_modules/README.md b/CMake/resolve_dependency_modules/README.md
@@ -37,7 +37,6 @@ by Velox. See details on bundling below.
 | wangle            | v2024.04.01.00  | No       |
 | mvfst             | v2024.04.01.00  | No       |
 | fbthrift          | v2024.04.01.00  | No       |
-| libstemmer        | 2.2.0           | Yes      |
 | DuckDB (testing)  | 0.8.1           | Yes      |
 | cpr (testing)     | 1.10.15         | Yes      |
 

diff --git a/CMake/resolve_dependency_modules/libstemmer/Makefile.patch b/CMake/resolve_dependency_modules/libstemmer/Makefile.patch
diff --git a/CMake/resolve_dependency_modules/stemmer.cmake b/CMake/resolve_dependency_modules/stemmer.cmake
diff --git a/CMakeLists.txt b/CMakeLists.txt
@@ -550,9 +550,6 @@ endif()
 set_source(xsimd)
 resolve_dependency(xsimd 10.0.0)
 
-set(stemmer_SOURCE BUNDLED)
-resolve_dependency(stemmer)
-
 if(VELOX_BUILD_TESTING)
   set(BUILD_TESTING ON)
   include(CTest) # include after project() but before add_subdirectory()

diff --git a/velox/docs/functions/presto/string.rst b/velox/docs/functions/presto/string.rst
@@ -255,43 +255,6 @@ String Functions
 
     Converts ``string`` to uppercase.
 
-.. function:: word_stem(word) -> varchar
-
-    Returns the stem of ``word`` in the English language. If the ``word`` is not an English word,
-    the ``word`` in lowercase is returned.
-
-.. function:: word_stem(word, lang) -> varchar
-
-    Returns the stem of ``word`` in the ``lang`` language. This function supports the following languages:
-
-    =========== ================
-    lang        Language
-    =========== ================
-    ``ca``      ``Catalan``
-    ``da``      ``Danish``
-    ``de``      ``German``
-    ``en``      ``English``
-    ``es``      ``Spanish``
-    ``eu``      ``Basque``
-    ``fi``      ``Finnish``
-    ``fr``      ``French``
-    ``hu``      ``Hungarian``
-    ``hy``      ``Armenian``
-    ``ir``      ``Irish``
-    ``it``      ``Italian``
-    ``lt``      ``Lithuanian``
-    ``nl``      ``Dutch``
-    ``no``      ``Norwegian``
-    ``pt``      ``Portuguese``
-    ``ro``      ``Romanian``
-    ``ru``      ``Russian``
-    ``sv``      ``Swedish``
-    ``tr``      ``Turkish``
-    =========== ================
-
-    If the specified ``lang`` is not supported, this function throws a user error.
-
-
 Unicode Functions
 -----------------
 

diff --git a/velox/functions/prestosql/CMakeLists.txt b/velox/functions/prestosql/CMakeLists.txt
@@ -72,8 +72,7 @@ target_link_libraries(
   velox_type_tz
   velox_presto_types
   velox_functions_util
-  Folly::folly
-  stemmer::stemmer)
+  Folly::folly)
 
 set_property(TARGET velox_functions_prestosql_impl PROPERTY JOB_POOL_COMPILE
                                                             high_memory_pool)

diff --git a/velox/functions/prestosql/WordStem.h b/velox/functions/prestosql/WordStem.h
diff --git a/velox/functions/prestosql/registration/StringFunctionsRegistration.cpp b/velox/functions/prestosql/registration/StringFunctionsRegistration.cpp
@@ -19,7 +19,6 @@
 #include "velox/functions/prestosql/SplitPart.h"
 #include "velox/functions/prestosql/SplitToMap.h"
 #include "velox/functions/prestosql/StringFunctions.h"
-#include "velox/functions/prestosql/WordStem.h"
 
 namespace facebook::velox::functions {
 
@@ -131,10 +130,5 @@ void registerStringFunctions(const std::string& prefix) {
       {prefix + "strrpos"});
   registerFunction<StrRPosFunction, int64_t, Varchar, Varchar, int64_t>(
       {prefix + "strrpos"});
-
-  // word_stem function
-  registerFunction<WordStemFunction, Varchar, Varchar>({prefix + "word_stem"});
-  registerFunction<WordStemFunction, Varchar, Varchar, Varchar>(
-      {prefix + "word_stem"});
 }
 } // namespace facebook::velox::functions
diff --git a/velox/functions/prestosql/tests/CMakeLists.txt b/velox/functions/prestosql/tests/CMakeLists.txt
@@ -98,7 +98,6 @@ add_executable(
   URLFunctionsTest.cpp
   Utf8Test.cpp
   WidthBucketArrayTest.cpp
-  WordStemTest.cpp
   ZipTest.cpp
   ZipWithTest.cpp)
 

diff --git a/velox/functions/prestosql/tests/WordStemTest.cpp b/velox/functions/prestosql/tests/WordStemTest.cpp