[kenlm] New Port (#13692)

3 files changed, 92 insertions, 0 deletions

diff --git a/ports/kenlm/fix-boost.patch b/ports/kenlm/fix-boost.patch
new file mode 100644
index 000000000..3eded1c69
--- /dev/null
+++ b/ports/kenlm/fix-boost.patch
@@ -0,0 +1,12 @@
+diff --git a/CMakeLists.txt b/CMakeLists.txt
+index bec81d4..44fd48a 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.
diff --git a/ports/kenlm/portfile.cmake b/ports/kenlm/portfile.cmake
new file mode 100644
index 000000000..38fe9652f
--- /dev/null
+++ b/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 1f054617eca14eae921e987b4b4eeb2b1d91de6b
+    SHA512 c18f9c22fbbb1f54ebe9c3b771fb2d7c09d502141d1b3645cff9db44cc51b3c976311ff0db79b60f410622579d043f185c56a4c7386e1b0ba8708e433238968b
+    HEAD_REF master
+    PATCHES fix-boost.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 "The interpolate feature does not support Windows.")
+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 query build_binary)
+if (NOT VCPKG_TARGET_IS_WINDOWS)
+    list(APPEND KENLM_TOOLS probing_hash_table_benchmark)
+    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)
diff --git a/ports/kenlm/vcpkg.json b/ports/kenlm/vcpkg.json
new file mode 100644
index 000000000..0542aa6ed
--- /dev/null
+++ b/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"
+        }
+      ]
+    }
+  }
+}