[R-package] adding routine registration in R package (fixes #1910) (#2911)

.ci/test_r_package.sh

@@ -91,7 +91,7 @@ if grep -q -R "WARNING" "$LOG_FILE_NAME"; then
     exit -1
 fi
 
-ALLOWED_CHECK_NOTES=3
+ALLOWED_CHECK_NOTES=2
 NUM_CHECK_NOTES=$(
     cat ${LOG_FILE_NAME} \
         | grep -e '^Status: .* NOTE.*' \

CMakeLists.txt

@@ -235,7 +235,15 @@ file(GLOB SOURCES
 )
 
 add_executable(lightgbm src/main.cpp ${SOURCES})
-add_library(_lightgbm SHARED src/c_api.cpp src/lightgbm_R.cpp ${SOURCES})
+list(APPEND SOURCES "src/c_api.cpp")
+
+# Only build the R part of the library if building for
+# use with the R package
+if(BUILD_FOR_R)
+  list(APPEND SOURCES "src/lightgbm_R.cpp")
+endif(BUILD_FOR_R)
+
+add_library(_lightgbm SHARED ${SOURCES})
 
 if(MSVC)
     set_target_properties(_lightgbm PROPERTIES OUTPUT_NAME "lib_lightgbm")
@@ -308,7 +316,6 @@ if(WIN32 AND (MINGW OR CYGWIN))
 endif()
 
 if(BUILD_FOR_R)
-    TARGET_LINK_LIBRARIES(lightgbm ${LIBR_CORE_LIBRARY})
     TARGET_LINK_LIBRARIES(_lightgbm ${LIBR_CORE_LIBRARY})
 endif(BUILD_FOR_R)
 

R-package/.Rbuildignore

@@ -4,20 +4,22 @@
 ^pkgdown$
 
 # Objects created by compilation
-\.o$
-\.so$
-\.dll$
-\.out$
-\.bin$
+^.*\.o
+^.*\.so
+^.*\.dll
+^.*\.out
+^.*\.bin
 
 # Code copied in at build time
 ^src/CMakeLists.txt$
+^Makefile$
+^src/build/.*$
 
 # unnecessary files from submodules
 ^src/compute/.appveyor.yml$
 ^src/compute/.coveralls.yml$
 ^src/compute/.travis.yml$
-^src/compute/test/$
+^src/compute/test/.*$
 ^src/compute/index.html$
 ^src/compute/.git$
 ^src/compute/.gitignore$

R-package/src/install.libs.R

@@ -43,7 +43,7 @@ if (!use_precompile) {
   # Prepare installation steps
   cmake_cmd <- "cmake "
   build_cmd <- "make _lightgbm"
-  lib_folder <- file.path(R_PACKAGE_SOURCE, "src", fsep = "/")
+  lib_folder <- file.path(source_dir, fsep = "/")
 
   if (use_gpu) {
     cmake_cmd <- paste0(cmake_cmd, " -DUSE_GPU=ON ")
@@ -98,7 +98,7 @@ if (!use_precompile) {
       } else {
         cmake_cmd <- paste0(cmake_cmd, local_vs_def)
         build_cmd <- "cmake --build . --target _lightgbm --config Release"
-        lib_folder <- file.path(R_PACKAGE_SOURCE, "src/Release", fsep = "/")
+        lib_folder <- file.path(source_dir, "Release", fsep = "/")
       }
     }
   }
@@ -110,9 +110,7 @@ if (!use_precompile) {
   # Makefile. We don't need it here anyway since targets are built serially, so trying
   # to remove it with this hack
   generated_makefile <- file.path(
-    R_PACKAGE_SOURCE
-    , "src"
-    , "build"
+    build_dir
     , "Makefile"
   )
   if (file.exists(generated_makefile)) {
@@ -163,12 +161,21 @@ if (!use_precompile) {
   }
 }
 
-# Check installation correctness
+# Packages with install.libs.R need to copy some artifacts into the
+# expected places in the package structure.
+# see https://cran.r-project.org/doc/manuals/r-devel/R-exts.html#Package-subdirectories,
+# especially the paragraph on install.libs.R
 dest <- file.path(R_PACKAGE_DIR, paste0("libs", R_ARCH), fsep = "/")
 dir.create(dest, recursive = TRUE, showWarnings = FALSE)
 if (file.exists(src)) {
   print(paste0("Found library file: ", src, " to move to ", dest))
   file.copy(src, dest, overwrite = TRUE)
+
+  symbols_file <- file.path(source_dir, "symbols.rds")
+  if (file.exists(symbols_file)) {
+    file.copy(symbols_file, dest, overwrite = TRUE)
+  }
+
 } else {
   stop(paste0("Cannot find lib_lightgbm", SHLIB_EXT))
 }

src/lightgbm_R.cpp

@@ -16,6 +16,8 @@
 #include <utility>
 #include <vector>
 
+#include <R_ext/Rdynload.h>
+
 #define COL_MAJOR (0)
 
 #define R_API_BEGIN() \
@@ -656,3 +658,54 @@ LGBM_SE LGBM_BoosterDumpModel_R(LGBM_SE handle,
   EncodeChar(out_str, inner_char_buf.data(), buffer_len, actual_len, static_cast<size_t>(out_len));
   R_API_END();
 }
+
+// .Call() calls
+static const R_CallMethodDef CallEntries[] = {
+  {"LGBM_GetLastError_R"              , (DL_FUNC) &LGBM_GetLastError_R              , 3},
+  {"LGBM_DatasetCreateFromFile_R"     , (DL_FUNC) &LGBM_DatasetCreateFromFile_R     , 5},
+  {"LGBM_DatasetCreateFromCSC_R"      , (DL_FUNC) &LGBM_DatasetCreateFromCSC_R      , 10},
+  {"LGBM_DatasetCreateFromMat_R"      , (DL_FUNC) &LGBM_DatasetCreateFromMat_R      , 7},
+  {"LGBM_DatasetGetSubset_R"          , (DL_FUNC) &LGBM_DatasetGetSubset_R          , 6},
+  {"LGBM_DatasetSetFeatureNames_R"    , (DL_FUNC) &LGBM_DatasetSetFeatureNames_R    , 3},
+  {"LGBM_DatasetGetFeatureNames_R"    , (DL_FUNC) &LGBM_DatasetGetFeatureNames_R    , 5},
+  {"LGBM_DatasetSaveBinary_R"         , (DL_FUNC) &LGBM_DatasetSaveBinary_R         , 3},
+  {"LGBM_DatasetFree_R"               , (DL_FUNC) &LGBM_DatasetFree_R               , 2},
+  {"LGBM_DatasetSetField_R"           , (DL_FUNC) &LGBM_DatasetSetField_R           , 5},
+  {"LGBM_DatasetGetFieldSize_R"       , (DL_FUNC) &LGBM_DatasetGetFieldSize_R       , 4},
+  {"LGBM_DatasetGetField_R"           , (DL_FUNC) &LGBM_DatasetGetField_R           , 4},
+  {"LGBM_DatasetUpdateParamChecking_R", (DL_FUNC) &LGBM_DatasetUpdateParamChecking_R, 3},
+  {"LGBM_DatasetGetNumData_R"         , (DL_FUNC) &LGBM_DatasetGetNumData_R         , 3},
+  {"LGBM_DatasetGetNumFeature_R"      , (DL_FUNC) &LGBM_DatasetGetNumFeature_R      , 3},
+  {"LGBM_BoosterCreate_R"             , (DL_FUNC) &LGBM_BoosterCreate_R             , 4},
+  {"LGBM_BoosterFree_R"               , (DL_FUNC) &LGBM_BoosterFree_R               , 2},
+  {"LGBM_BoosterCreateFromModelfile_R", (DL_FUNC) &LGBM_BoosterCreateFromModelfile_R, 3},
+  {"LGBM_BoosterLoadModelFromString_R", (DL_FUNC) &LGBM_BoosterLoadModelFromString_R, 3},
+  {"LGBM_BoosterMerge_R"              , (DL_FUNC) &LGBM_BoosterMerge_R              , 3},
+  {"LGBM_BoosterAddValidData_R"       , (DL_FUNC) &LGBM_BoosterAddValidData_R       , 3},
+  {"LGBM_BoosterResetTrainingData_R"  , (DL_FUNC) &LGBM_BoosterResetTrainingData_R  , 3},
+  {"LGBM_BoosterResetParameter_R"     , (DL_FUNC) &LGBM_BoosterResetParameter_R     , 3},
+  {"LGBM_BoosterGetNumClasses_R"      , (DL_FUNC) &LGBM_BoosterGetNumClasses_R      , 3},
+  {"LGBM_BoosterUpdateOneIter_R"      , (DL_FUNC) &LGBM_BoosterUpdateOneIter_R      , 2},
+  {"LGBM_BoosterUpdateOneIterCustom_R", (DL_FUNC) &LGBM_BoosterUpdateOneIterCustom_R, 5},
+  {"LGBM_BoosterRollbackOneIter_R"    , (DL_FUNC) &LGBM_BoosterRollbackOneIter_R    , 2},
+  {"LGBM_BoosterGetCurrentIteration_R", (DL_FUNC) &LGBM_BoosterGetCurrentIteration_R, 3},
+  {"LGBM_BoosterGetUpperBoundValue_R" , (DL_FUNC) &LGBM_BoosterGetUpperBoundValue_R , 3},
+  {"LGBM_BoosterGetLowerBoundValue_R" , (DL_FUNC) &LGBM_BoosterGetLowerBoundValue_R , 3},
+  {"LGBM_BoosterGetEvalNames_R"       , (DL_FUNC) &LGBM_BoosterGetEvalNames_R       , 5},
+  {"LGBM_BoosterGetEval_R"            , (DL_FUNC) &LGBM_BoosterGetEval_R            , 4},
+  {"LGBM_BoosterGetNumPredict_R"      , (DL_FUNC) &LGBM_BoosterGetNumPredict_R      , 4},
+  {"LGBM_BoosterGetPredict_R"         , (DL_FUNC) &LGBM_BoosterGetPredict_R         , 4},
+  {"LGBM_BoosterPredictForFile_R"     , (DL_FUNC) &LGBM_BoosterPredictForFile_R     , 10},
+  {"LGBM_BoosterCalcNumPredict_R"     , (DL_FUNC) &LGBM_BoosterCalcNumPredict_R     , 8},
+  {"LGBM_BoosterPredictForCSC_R"      , (DL_FUNC) &LGBM_BoosterPredictForCSC_R      , 14},
+  {"LGBM_BoosterPredictForMat_R"      , (DL_FUNC) &LGBM_BoosterPredictForMat_R      , 11},
+  {"LGBM_BoosterSaveModel_R"          , (DL_FUNC) &LGBM_BoosterSaveModel_R          , 4},
+  {"LGBM_BoosterSaveModelToString_R"  , (DL_FUNC) &LGBM_BoosterSaveModelToString_R  , 6},
+  {"LGBM_BoosterDumpModel_R"          , (DL_FUNC) &LGBM_BoosterDumpModel_R          , 6},
+  {NULL, NULL, 0}
+};
+
+void R_init_lightgbm(DllInfo *dll) {
+  R_registerRoutines(dll, NULL, CallEntries, NULL, NULL);
+  R_useDynamicSymbols(dll, FALSE);
+}

windows/LightGBM.vcxproj

@@ -237,13 +237,11 @@
     <ClInclude Include="..\include\LightGBM\dataset.h" />
     <ClInclude Include="..\include\LightGBM\dataset_loader.h" />
     <ClInclude Include="..\include\LightGBM\feature_group.h" />
-    <ClInclude Include="..\include\LightGBM\lightgbm_R.h" />
     <ClInclude Include="..\include\LightGBM\meta.h" />
     <ClInclude Include="..\include\LightGBM\metric.h" />
     <ClInclude Include="..\include\LightGBM\network.h" />
     <ClInclude Include="..\include\LightGBM\objective_function.h" />
     <ClInclude Include="..\include\LightGBM\prediction_early_stop.h" />
-    <ClInclude Include="..\include\LightGBM\R_object_helper.h" />
     <ClInclude Include="..\include\LightGBM\tree.h" />
     <ClInclude Include="..\include\LightGBM\tree_learner.h" />
     <ClInclude Include="..\include\LightGBM\utils\array_args.h" />
@@ -308,7 +306,6 @@
     <ClCompile Include="..\src\io\metadata.cpp" />
     <ClCompile Include="..\src\io\parser.cpp" />
     <ClCompile Include="..\src\io\tree.cpp" />
-    <ClCompile Include="..\src\lightgbm_R.cpp" />
     <ClCompile Include="..\src\metric\dcg_calculator.cpp" />
     <ClCompile Include="..\src\metric\metric.cpp" />
     <ClCompile Include="..\src\network\network.cpp" />

windows/LightGBM.vcxproj.filters

@@ -183,12 +183,6 @@
     <ClInclude Include="..\src\metric\map_metric.hpp">
       <Filter>src\metric</Filter>
     </ClInclude>
-    <ClInclude Include="..\include\LightGBM\lightgbm_R.h">
-      <Filter>include\LightGBM</Filter>
-    </ClInclude>
-    <ClInclude Include="..\include\LightGBM\R_object_helper.h">
-      <Filter>include\LightGBM</Filter>
-    </ClInclude>
     <ClInclude Include="..\src\boosting\rf.hpp">
       <Filter>src\boosting</Filter>
     </ClInclude>
@@ -296,9 +290,6 @@
     <ClCompile Include="..\src\boosting\prediction_early_stop.cpp">
       <Filter>src\boosting</Filter>
     </ClCompile>
-    <ClCompile Include="..\src\lightgbm_R.cpp">
-      <Filter>src</Filter>
-    </ClCompile>
     <ClCompile Include="..\src\boosting\gbdt_model_text.cpp">
       <Filter>src\boosting</Filter>
     </ClCompile>