[kenlm] New Port

ports/kenlm/fix-build-install.patch

@@ -0,0 +1,89 @@
+diff --git a/CMakeLists.txt b/CMakeLists.txt
+index 9d9ed21..2c7aeac 100644
+--- a/CMakeLists.txt
++++ b/CMakeLists.txt
+@@ -1,6 +1,6 @@
+ cmake_minimum_required(VERSION 3.1)
+
+-IF( WIN32 )
++IF( WIN32 AND BUILD_SHARED_LIBS)
+     SET(Boost_USE_STATIC_LIBS OFF)
+     # The auto-linking feature has problems with USE_STATIC_LIBS off, so we use
+     # BOOST_ALL_NO_LIB to turn it off.
+@@ -14,6 +14,7 @@ ENDIF( )
+ project(kenlm)
+
+ option(FORCE_STATIC "Build static executables" OFF)
++option(ENABLE_INTERPOLATE "Build interpolation program" OFF)
+ if (FORCE_STATIC)
+   #presumably overkill, is there a better way?
+   #http://cmake.3232098.n2.nabble.com/Howto-compile-static-executable-td5580269.html
+diff --git a/lm/CMakeLists.txt b/lm/CMakeLists.txt
+index 25d22c3..395815b 100644
+--- a/lm/CMakeLists.txt
++++ b/lm/CMakeLists.txt
+@@ -43,6 +43,9 @@ install(TARGETS kenlm
+   ARCHIVE DESTINATION lib
+ )
+
++file(GLOB SRC_HEADERS ${CMAKE_CURRENT_LIST_DIR}/*.h ${CMAKE_CURRENT_LIST_DIR}/*.hh)
++install(FILES ${SRC_HEADERS} DESTINATION include/kenlm)
++
+ # This directory has children that need to be processed
+ add_subdirectory(builder)
+ add_subdirectory(filter)
+@@ -50,12 +53,14 @@ add_subdirectory(interpolate)
+
+ # Explicitly list the executable files to be compiled
+ set(EXE_LIST
+-  query
+   fragment
+-  build_binary
+   kenlm_benchmark
+ )
+
++if (UNIX)
++    list(APPEND EXE_LIST query build_binary)
++endif()
++
+ set(LM_LIBS kenlm kenlm_util ${Boost_LIBRARIES} Threads::Threads)
+
+ AddExes(EXES ${EXE_LIST}
+diff --git a/lm/interpolate/CMakeLists.txt b/lm/interpolate/CMakeLists.txt
+index 04776be..7375ffd 100644
+--- a/lm/interpolate/CMakeLists.txt
++++ b/lm/interpolate/CMakeLists.txt
+@@ -1,5 +1,6 @@
++if (ENABLE_INTERPOLATE)
+ # Eigen3 less than 3.1.0 has a race condition: http://eigen.tuxfamily.org/bz/show_bug.cgi?id=466
+-find_package(Eigen3 3.1.0 CONFIG)
++find_package(Eigen3 3.1.0 CONFIG REQUIRED)
+
+ if(EIGEN3_FOUND)
+   include_directories(${EIGEN3_INCLUDE_DIR})
+@@ -54,3 +55,4 @@ else()
+       "(cd $HOME; wget -O - https://gitlab.com/libeigen/eigen/-/archive/3.3.7/eigen-3.3.7.tar.bz2 |tar xj)\n"
+       "rm CMakeCache.txt\n")
+ endif()
++endif()
+\ No newline at end of file
+diff --git a/util/CMakeLists.txt b/util/CMakeLists.txt
+index b0a06b1..711aa32 100644
+--- a/util/CMakeLists.txt
++++ b/util/CMakeLists.txt
+@@ -84,6 +84,15 @@ AddExes(EXES probing_hash_table_benchmark
+         LIBRARIES kenlm_util ${Boost_LIBRARIES} Threads::Threads)
+ endif()
+
++install(TARGETS kenlm_util
++  RUNTIME DESTINATION bin
++  LIBRARY DESTINATION lib
++  ARCHIVE DESTINATION lib
++)
++
++file(GLOB UTILS_HEADER ${CMAKE_CURRENT_LIST_DIR}/*.h ${CMAKE_CURRENT_LIST_DIR}/*.hh)
++install(FILES ${UTILS_HEADER} DESTINATION include/kenlm/utils)
++
+ # Only compile and run unit tests if tests should be run
+ if(BUILD_TESTING)
+   set(KENLM_BOOST_TESTS_LIST

ports/kenlm/portfile.cmake

@@ -0,0 +1,48 @@
+vcpkg_check_linkage(ONLY_STATIC_LIBRARY)
+
+vcpkg_from_github(
+    OUT_SOURCE_PATH SOURCE_PATH
+    REPO kpu/kenlm
+    REF ac454207c69f293315ae9be3aff2238fc8c999a0
+    SHA512 96f35c46237870ce71c04e20783b4d410e7dfec6eb10673e29aa9ba27f95c5ad77aafd443a1583cd7b757a8c360fc16f236ef8ecf965f317c24c8f3b45547722
+    HEAD_REF master
+    PATCHES fix-build-install.patch
+)
+
+file(REMOVE ${SOURCE_PATH}/cmake/modules/FindEigen3.cmake)
+
+vcpkg_check_features(OUT_FEATURE_OPTIONS FEATURE_OPTIONS
+    interpolate ENABLE_INTERPOLATE
+)
+
+if ("interpolate" IN_LIST FEATURES AND VCPKG_TARGET_IS_WINDOWS)
+    message(FATAL_ERROR "Feature interpolate only support unix.")
+endif()
+
+vcpkg_configure_cmake(
+    SOURCE_PATH ${SOURCE_PATH}
+    PREFER_NINJA
+    OPTIONS
+        ${FEATURE_OPTIONS}
+        -DFORCE_STATIC=OFF #already handled by vcpkg
+        -DENABLE_PYTHON=OFF # kenlm.lib(bhiksha.cc.obj) : fatal error LNK1000: Internal error during IMAGE::Pass2
+        -DCOMPILE_TESTS=OFF
+)
+vcpkg_install_cmake()
+
+set(KENLM_TOOLS count_ngrams filter fragment kenlm_benchmark lmplz phrase_table_vocab)
+if (NOT VCPKG_TARGET_IS_WINDOWS)
+    list(APPEND KENLM_TOOLS probing_hash_table_benchmark query build_binary streaming_example)
+    if ("interpolate" IN_LIST FEATURES)
+        list(APPEND KENLM_TOOLS interpolate)
+    endif()
+endif()
+vcpkg_copy_tools(TOOL_NAMES ${KENLM_TOOLS} AUTO_CLEAN)
+
+vcpkg_copy_pdbs()
+
+file(REMOVE_RECURSE "${CURRENT_PACKAGES_DIR}/debug/include")
+
+# Copyright and License
+file(INSTALL ${SOURCE_PATH}/COPYING DESTINATION ${CURRENT_PACKAGES_DIR}/share/${PORT} RENAME copyright)
+file(INSTALL ${SOURCE_PATH}/LICENSE DESTINATION ${CURRENT_PACKAGES_DIR}/share/${PORT} RENAME license)

ports/kenlm/vcpkg.json

@@ -0,0 +1,32 @@
+{
+  "name": "kenlm",
+  "version-string": "20200924",
+  "description": "KenLM: Faster and Smaller Language Model Queries",
+  "supports": "!(arm64 & windows)",
+  "dependencies": [
+    "boost-interprocess",
+    "boost-program-options",
+    "boost-ptr-container",
+    "boost-system",
+    "boost-test",
+    "boost-thread",
+    "bzip2",
+    {
+      "name": "eigen3",
+      "platform": "!windows"
+    },
+    "liblzma",
+    "zlib"
+  ],
+  "features": {
+    "interpolate": {
+      "description": "Build interpolation program",
+      "dependencies": [
+        {
+          "name": "eigen3",
+          "platform": "!windows"
+        }
+      ]
+    }
+  }
+}