[optional] support loading and saving models with protobuf

.travis.yml

@@ -23,6 +23,7 @@ env:
     - TASK=if-else
     - TASK=sdist PYTHON_VERSION=3.4
     - TASK=bdist PYTHON_VERSION=3.5
+    - TASK=proto
     - TASK=gpu METHOD=source
     - TASK=gpu METHOD=pip
 
@@ -38,6 +39,8 @@ matrix:
       env: TASK=pylint
     - os: osx
       env: TASK=check-docs
+    - os: osx
+      env: TASK=proto
 
 before_install:
   - test -n $CC  && unset CC

.travis/test.sh

@@ -48,12 +48,24 @@ if [[ ${TASK} == "if-else" ]]; then
     conda create -q -n test-env python=$PYTHON_VERSION numpy
     source activate test-env
     mkdir build && cd build && cmake .. && make lightgbm || exit -1
-    cd $TRAVIS_BUILD_DIR/tests/cpp_test && ../../lightgbm config=train.conf && ../../lightgbm config=predict.conf output_result=origin.pred || exit -1
+    cd $TRAVIS_BUILD_DIR/tests/cpp_test && ../../lightgbm config=train.conf convert_model_language=cpp convert_model=../../src/boosting/gbdt_prediction.cpp && ../../lightgbm config=predict.conf output_result=origin.pred || exit -1
     cd $TRAVIS_BUILD_DIR/build && make lightgbm || exit -1
     cd $TRAVIS_BUILD_DIR/tests/cpp_test && ../../lightgbm config=predict.conf output_result=ifelse.pred && python test.py || exit -1
     exit 0
 fi
 
+if [[ ${TASK} == "proto" ]]; then
+    conda create -q -n test-env python=$PYTHON_VERSION numpy
+    source activate test-env
+    mkdir build && cd build && cmake .. && make lightgbm || exit -1
+    cd $TRAVIS_BUILD_DIR/tests/cpp_test && ../../lightgbm config=train.conf && ../../lightgbm config=predict.conf output_result=origin.pred || exit -1
+    cd $TRAVIS_BUILD_DIR && git clone https://github.com/google/protobuf && cd protobuf && ./autogen.sh && ./configure && make && sudo make install && sudo ldconfig
+    cd $TRAVIS_BUILD_DIR/build && rm -rf * && cmake -DUSE_PROTO=ON .. && make lightgbm || exit -1
+    cd $TRAVIS_BUILD_DIR/tests/cpp_test && ../../lightgbm config=train.conf model_format=proto && ../../lightgbm config=predict.conf output_result=proto.pred model_format=proto || exit -1
+    cd $TRAVIS_BUILD_DIR/tests/cpp_test && python test.py || exit -1
+    exit 0
+fi
+
 conda create -q -n test-env python=$PYTHON_VERSION numpy nose scipy scikit-learn pandas matplotlib pytest
 source activate test-env
 

CMakeLists.txt

@@ -124,8 +124,24 @@ file(GLOB SOURCES
     src/treelearner/*.cpp
 )
 
-add_executable(lightgbm src/main.cpp ${SOURCES})
-add_library(_lightgbm SHARED src/c_api.cpp src/lightgbm_R.cpp ${SOURCES})
+if (USE_PROTO)
+    find_package(Protobuf REQUIRED)
+    PROTOBUF_GENERATE_CPP(PROTO_SRCS PROTO_HDRS proto/model.proto)
+    include_directories(${PROTOBUF_INCLUDE_DIRS})
+    include_directories(${CMAKE_CURRENT_BINARY_DIR})
+    SET(PROTO_FILES src/proto/gbdt_model_proto.cpp ${PROTO_HDRS} ${PROTO_SRCS})
+else()
+    include_directories(src/proto/not_implemented)
+    SET(PROTO_FILES src/proto/not_implemented/gbdt_model_proto.cpp)
+endif(USE_PROTO)
+
+add_executable(lightgbm src/main.cpp ${SOURCES} ${PROTO_FILES})
+add_library(_lightgbm SHARED src/c_api.cpp src/lightgbm_R.cpp ${SOURCES} ${PROTO_FILES})
+
+if (USE_PROTO)
+    TARGET_LINK_LIBRARIES(lightgbm ${PROTOBUF_LIBRARIES})
+    TARGET_LINK_LIBRARIES(_lightgbm ${PROTOBUF_LIBRARIES})
+endif(USE_PROTO)
 
 if(MSVC)
     set_target_properties(_lightgbm PROPERTIES OUTPUT_NAME "lib_lightgbm")

docs/Parameters.rst

@@ -324,6 +324,20 @@ IO Parameters
 
    -  file name of prediction result in ``prediction`` task
 
+-  ``model_format``, default=\ ``text``, type=string
+
+   -  format to save and load model.
+
+   -  ``text``, use text string.
+
+   -  ``proto``, use protocol buffer binary format.
+
+   -  save multiple formats by joining them with comma, like ``text,proto``, in this case, ``model_format`` will be add as suffix after ``output_model``.
+
+   -  not support loading with multiple formats.
+
+   -  Note: you need to cmake with -DUSE_PROTO=ON to use this parameter.
+
 -  ``is_pre_partition``, default=\ ``false``, type=bool
 
    -  used for parallel learning (not include feature parallel)

include/LightGBM/boosting.h

@@ -3,6 +3,7 @@
 
 #include <LightGBM/meta.h>
 #include <LightGBM/config.h>
+#include "model.pb.h"
 
 #include <vector>
 #include <string>
@@ -166,7 +167,7 @@ class LIGHTGBM_EXPORT Boosting {
 
   /*!
   * \brief Save model to file
-  * \param num_used_model Number of model that want to save, -1 means save all
+  * \param num_iterations Number of model that want to save, -1 means save all
   * \param is_finish Is training finished or not
   * \param filename Filename that want to save to
   * \return true if succeeded
@@ -175,7 +176,7 @@ class LIGHTGBM_EXPORT Boosting {
 
   /*!
   * \brief Save model to string
-  * \param num_used_model Number of model that want to save, -1 means save all
+  * \param num_iterations Number of model that want to save, -1 means save all
   * \return Non-empty string if succeeded
   */
   virtual std::string SaveModelToString(int num_iterations) const = 0;
@@ -187,6 +188,20 @@ class LIGHTGBM_EXPORT Boosting {
   */
   virtual bool LoadModelFromString(const std::string& model_str) = 0;
 
+  /*!
+  * \brief Save model with protobuf
+  * \param num_iterations Number of model that want to save, -1 means save all
+  * \param filename Filename that want to save to
+  */
+  virtual void SaveModelToProto(int num_iteration, const char* filename) const = 0;
+
+  /*!
+  * \brief Restore from a serialized protobuf file
+  * \param filename Filename that want to restore from
+  * \return true if succeeded
+  */
+  virtual bool LoadModelFromProto(const char* filename) = 0;
+
   /*!
   * \brief Calculate feature importances
   * \param num_iteration Number of model that want to use for feature importance, -1 means use all
@@ -251,23 +266,17 @@ class LIGHTGBM_EXPORT Boosting {
   /*! \brief Disable copy */
   Boosting(const Boosting&) = delete;
 
-  static bool LoadFileToBoosting(Boosting* boosting, const char* filename);
+  static bool LoadFileToBoosting(Boosting* boosting, const std::string& format, const char* filename);
 
   /*!
   * \brief Create boosting object
   * \param type Type of boosting
+  * \param format Format of model
   * \param config config for boosting
   * \param filename name of model file, if existing will continue to train from this model
   * \return The boosting object
   */
-  static Boosting* CreateBoosting(const std::string& type, const char* filename);
-
-  /*!
-  * \brief Create boosting object from model file
-  * \param filename name of model file
-  * \return The boosting object
-  */
-  static Boosting* CreateBoosting(const char* filename);
+  static Boosting* CreateBoosting(const std::string& type, const std::string& format, const char* filename);
 
 };
 

include/LightGBM/config.h

@@ -105,6 +105,7 @@ struct IOConfig: public ConfigBase {
   std::string output_result = "LightGBM_predict_result.txt";
   std::string convert_model = "gbdt_prediction.cpp";
   std::string input_model = "";
+  std::string model_format = "text";
   int verbosity = 1;
   int num_iteration_predict = -1;
   bool is_pre_partition = false;
@@ -446,7 +447,7 @@ struct ParameterAlias {
     const std::unordered_set<std::string> parameter_set({
       "config", "config_file", "task", "device",
       "num_threads", "seed", "boosting_type", "objective", "data",
-      "output_model", "input_model", "output_result", "valid_data",
+      "output_model", "input_model", "output_result", "model_format", "valid_data",
       "is_enable_sparse", "is_pre_partition", "is_training_metric",
       "ndcg_eval_at", "min_data_in_leaf", "min_sum_hessian_in_leaf",
       "num_leaves", "feature_fraction", "num_iterations",

include/LightGBM/tree.h

@@ -3,6 +3,7 @@
 
 #include <LightGBM/meta.h>
 #include <LightGBM/dataset.h>
+#include "model.pb.h"
 
 #include <string>
 #include <vector>
@@ -31,6 +32,12 @@ class Tree {
   */
   explicit Tree(const std::string& str);
 
+  /*!
+  * \brief Construtor, from a protobuf object
+  * \param model_tree Model protobuf object
+  */
+  explicit Tree(const LightGBM::Model_Tree& model_tree);
+
   ~Tree();
 
   /*!
@@ -165,6 +172,9 @@ class Tree {
   /*! \brief Serialize this object to if-else statement*/
   std::string ToIfElse(int index, bool is_predict_leaf_index) const;
 
+  /*! \brief Serialize this object to protobuf object*/
+  void ToProto(Model_Tree& model_tree) const;
+
   inline static bool IsZero(double fval) {
     if (fval > -kZeroAsMissingValueRange && fval <= kZeroAsMissingValueRange) {
       return true;

proto/model.proto

@@ -0,0 +1,33 @@
+syntax = "proto3";
+
+package LightGBM;
+
+message Model {
+    string name = 1;
+    uint32 num_class = 2;
+    uint32 num_tree_per_iteration = 3;
+    uint32 label_index = 4;
+    uint32 max_feature_idx = 5;
+    string objective = 6;
+    bool average_output = 7;
+    repeated string feature_names = 8;
+    repeated string feature_infos = 9;
+    message Tree {
+        uint32 num_leaves = 1;
+        uint32 num_cat = 2;
+        repeated uint32 split_feature = 3;
+        repeated double split_gain = 4;
+        repeated double threshold = 5;
+        repeated uint32 decision_type = 6;
+        repeated sint32 left_child = 7;
+        repeated sint32 right_child = 8;
+        repeated double leaf_value = 9;
+        repeated uint32 leaf_count = 10;
+        repeated double internal_value = 11;
+        repeated double internal_count = 12;
+        repeated sint32 cat_boundaries = 13;
+        repeated uint32 cat_threshold = 14;
+        double shrinkage = 15;
+    }
+    repeated Tree trees = 10;
+}

src/application/application.cpp

@@ -180,6 +180,7 @@ void Application::InitTrain() {
   // create boosting
   boosting_.reset(
     Boosting::CreateBoosting(config_.boosting_type,
+                             config_.io_config.model_format.c_str(),
                              config_.io_config.input_model.c_str()));
   // create objective function
   objective_fun_.reset(
@@ -203,6 +204,22 @@ void Application::InitTrain() {
 void Application::Train() {
   Log::Info("Started training...");
   boosting_->Train(config_.io_config.snapshot_freq, config_.io_config.output_model);
+  std::vector<std::string> model_formats = Common::Split(config_.io_config.model_format.c_str(), ',');
+  bool save_with_multiple_format = (model_formats.size() > 1);
+  for (auto model_format: model_formats) {
+    std::string save_file_name = config_.io_config.output_model;
+    if (save_with_multiple_format) {
+      // use suffix to distinguish different model format
+      save_file_name += "." + model_format;
+    }
+    if (model_format == std::string("text")) {
+      boosting_->SaveModelToFile(-1, save_file_name.c_str());
+    } else if (model_format == std::string("proto")) {
+      boosting_->SaveModelToProto(-1, save_file_name.c_str());
+    } else {
+      Log::Fatal("Unknown model format during saving: %s", model_format.c_str());
+    }
+  }
   // convert model to if-else statement code
   if (config_.convert_model_language == std::string("cpp")) {
     boosting_->SaveModelToIfElse(-1, config_.io_config.convert_model.c_str());
@@ -223,13 +240,15 @@ void Application::Predict() {
 
 void Application::InitPredict() {
   boosting_.reset(
-    Boosting::CreateBoosting(config_.io_config.input_model.c_str()));
+    Boosting::CreateBoosting("gbdt", config_.io_config.model_format.c_str(),
+                             config_.io_config.input_model.c_str()));
   Log::Info("Finished initializing prediction");
 }
 
 void Application::ConvertModel() {
   boosting_.reset(
     Boosting::CreateBoosting(config_.boosting_type,
+                             config_.io_config.model_format.c_str(),
                              config_.io_config.input_model.c_str()));
   boosting_->SaveModelToIfElse(-1, config_.io_config.convert_model.c_str());
 }

src/boosting/boosting.cpp

@@ -12,21 +12,30 @@ std::string GetBoostingTypeFromModelFile(const char* filename) {
   return type;
 }
 
-bool Boosting::LoadFileToBoosting(Boosting* boosting, const char* filename) {
+bool Boosting::LoadFileToBoosting(Boosting* boosting, const std::string& format, const char* filename) {
   if (boosting != nullptr) {
-    TextReader<size_t> model_reader(filename, true);
-    model_reader.ReadAllLines();
-    std::stringstream str_buf;
-    for (auto& line : model_reader.Lines()) {
-      str_buf << line << '\n';
+    if (format == std::string("text")) {
+      TextReader<size_t> model_reader(filename, true);
+      model_reader.ReadAllLines();
+      std::stringstream str_buf;
+      for (auto& line : model_reader.Lines()) {
+        str_buf << line << '\n';
+      }
+      if (!boosting->LoadModelFromString(str_buf.str())) {
+        return false;
+      }
+    } else if (format == std::string("proto")) {
+      if (!boosting->LoadModelFromProto(filename)) {
+        return false;
+      }
+    } else {
+      Log::Fatal("Unknown model format during loading: %s", format.c_str());
     }
-    if (!boosting->LoadModelFromString(str_buf.str()))
-      return false;
   }
   return true;
 }
 
-Boosting* Boosting::CreateBoosting(const std::string& type, const char* filename) {
+Boosting* Boosting::CreateBoosting(const std::string& type, const std::string& format, const char* filename) {
   if (filename == nullptr || filename[0] == '\0') {
     if (type == std::string("gbdt")) {
       return new GBDT();
@@ -41,8 +50,7 @@ Boosting* Boosting::CreateBoosting(const std::string& type, const char* filename
     }
   } else {
     std::unique_ptr<Boosting> ret;
-    auto type_in_file = GetBoostingTypeFromModelFile(filename);
-    if (type_in_file == std::string("tree")) {
+    if (format == std::string("proto") || GetBoostingTypeFromModelFile(filename) == std::string("tree")) {
       if (type == std::string("gbdt")) {
         ret.reset(new GBDT());
       } else if (type == std::string("dart")) {
@@ -54,24 +62,12 @@ Boosting* Boosting::CreateBoosting(const std::string& type, const char* filename
       } else {
         Log::Fatal("unknown boosting type %s", type.c_str());
       }
-      LoadFileToBoosting(ret.get(), filename);
+      LoadFileToBoosting(ret.get(), format, filename);
     } else {
-      Log::Fatal("unknown submodel type in model file %s", filename);
+      Log::Fatal("unknown model format or submodel type in model file %s", filename);
     }
     return ret.release();
   }
 }
 
-Boosting* Boosting::CreateBoosting(const char* filename) {
-  auto type = GetBoostingTypeFromModelFile(filename);
-  std::unique_ptr<Boosting> ret;
-  if (type == std::string("tree")) {
-    ret.reset(new GBDT());
-  } else {
-    Log::Fatal("unknown submodel type in model file %s", filename);
-  }
-  LoadFileToBoosting(ret.get(), filename);
-  return ret.release();
-}
-
 }  // namespace LightGBM

src/boosting/gbdt.cpp

@@ -352,7 +352,6 @@ void GBDT::Train(int snapshot_freq, const std::string& model_output_path) {
       SaveModelToFile(-1, snapshot_out.c_str());
     }
   }
-  SaveModelToFile(-1, model_output_path.c_str());
 }
 
 double GBDT::BoostFromAverage() {