some bugs fixed

include/LightGBM/c_api.h

@@ -149,7 +149,7 @@ DllExport int LGBM_DatasetCreateFromMat(const void* data,
 DllExport int LGBM_DatasetGetSubset(
   const DatesetHandle* full_data,
   const int32_t* used_row_indices,
-  const int32_t num_used_row_indices,
+  int32_t num_used_row_indices,
   const char* parameters,
   DatesetHandle* out);
 

include/LightGBM/dataset.h

@@ -47,6 +47,13 @@ public:
   */
   void Init(const char* data_filename, const int num_class);
   /*!
+  * \brief init as subset
+  * \param metadata Filename of data
+  * \param used_indices 
+  * \param num_used_indices
+  */
+  void Init(const Metadata& metadata, const data_size_t* used_indices, data_size_t num_used_indices);
+  /*!
   * \brief Initial with binary memory
   * \param memory Pointer to memory
   */
@@ -77,7 +84,6 @@ public:
   void CheckOrPartition(data_size_t num_all_data,
     const std::vector<data_size_t>& used_data_indices);
 
-
   void SetLabel(const float* label, data_size_t len);
 
   void SetWeights(const float* weights, data_size_t len);

python-package/lightgbm/basic.py

@@ -410,6 +410,10 @@ class Dataset(object):
         params: dict, optional
             other parameters
         """
+        self.__label = None
+        self.__weight = None
+        self.__init_score = None
+        self.__group = None
         if data is None:
             self.handle = None
             return
@@ -453,10 +457,6 @@ class Dataset(object):
                 self.__init_from_csr(csr, params_str, ref_dataset)
             except:
                 raise TypeError('can not initialize Dataset from {}'.format(type(data).__name__))
-        self.__label = None
-        self.__weight = None
-        self.__init_score = None
-        self.__group = None
         if label is not None:
             self.set_label(label)
         if self.get_label() is None:
@@ -505,6 +505,22 @@ class Dataset(object):
         return Dataset(data, label=label, max_bin=self.max_bin, reference=self,
             weight=weight, group_id=group_id, predictor=self.predictor, 
             silent=silent, params=params)
+    def subset(self, used_indices, params=None):
+        used_indices = list_to_1d_numpy(used_indices, np.int32)
+        ret = Dataset(None)
+        ret.handle = ctypes.c_void_p()
+        params_str = dict_to_str(params)
+        _safe_call(_LIB.LGBM_DatasetGetSubset(
+            ctypes.byref(self.handle), 
+            used_indices.data_as(ctypes.POINTER(ctypes.c_int32)),
+            used_indices.shape[0],
+            c_str(params_str), 
+            ctypes.byref(ret.handle)))
+        ret.max_bin = self.max_bin
+        ret.predictor = self.predictor
+        if ret.get_label() is None:
+            raise ValueError("label should not be None")
+        return ret
 
     def __init_from_np2d(self, mat, params_str, ref_dataset):
         """
@@ -1102,7 +1118,7 @@ class Booster(object):
 
     def __inner_eval(self, data_name, data_idx, feval=None):
         """
-        Evaulate traning or validation data
+        Evaulate training  or validation data
         """
         if data_idx >= self.__num_dataset:
             raise ValueError("data_idx should be smaller than number of dataset")

src/c_api.cpp

@@ -387,7 +387,7 @@ DllExport int LGBM_DatasetCreateFromCSC(const void* col_ptr,
 DllExport int LGBM_DatasetGetSubset(
   const DatesetHandle* full_data,
   const int32_t* used_row_indices,
-  const int32_t num_used_row_indices,
+  int32_t num_used_row_indices,
   const char* parameters,
   DatesetHandle* out) {
   API_BEGIN();

src/io/dataset.cpp

@@ -55,6 +55,7 @@ void Dataset::CopyFeatureMapperFrom(const Dataset* dataset, bool is_enable_spars
   num_features_ = static_cast<int>(features_.size());
   num_total_features_ = dataset->num_total_features_;
   feature_names_ = dataset->feature_names_;
+  label_idx_ = dataset->label_idx_;
 }
 
 Dataset* Dataset::Subset(const data_size_t* used_indices, data_size_t num_used_indices, bool is_enable_sparse) const {
@@ -67,6 +68,8 @@ Dataset* Dataset::Subset(const data_size_t* used_indices, data_size_t num_used_i
       ret->features_[fidx]->PushBin(0, i, iterator->Get(used_indices[i]));
     }
   }
+  ret->metadata_.Init(metadata_, used_indices, num_used_indices);
+  return ret.release();
 }
 
 bool Dataset::SetFloatField(const char* field_name, const float* field_data, data_size_t num_element) {

src/io/metadata.cpp

@@ -50,6 +50,69 @@ void Metadata::Init(data_size_t num_data, int num_class, int weight_idx, int que
   }
 }
 
+void Metadata::Init(const Metadata& fullset, const data_size_t* used_indices, data_size_t num_used_indices) {
+  num_data_ = num_used_indices;
+  num_class_ = fullset.num_class_;
+
+  label_ = std::vector<float>(num_used_indices);
+  for (data_size_t i = 0; i < num_used_indices; i++) {
+    label_[i] = fullset.label_[used_indices[i]];
+  }
+
+  if (fullset.weights_.size() > 0) {
+    weights_ = std::vector<float>(num_used_indices);
+    num_weights_ = num_used_indices;
+    for (data_size_t i = 0; i < num_used_indices; i++) {
+      weights_[i] = fullset.weights_[used_indices[i]];
+    }
+  } else {
+    num_weights_ = 0;
+  }
+
+  if (fullset.init_score_.size() > 0) {
+    init_score_ = std::vector<float>(num_used_indices);
+    num_init_score_ = num_used_indices;
+    for (data_size_t i = 0; i < num_used_indices; i++) {
+      init_score_[i] = fullset.init_score_[used_indices[i]];
+    }
+  } else {
+    num_init_score_ = 0;
+  }
+
+  if (fullset.query_boundaries_.size() > 0) {
+    std::vector<data_size_t> used_query;
+    data_size_t data_idx = 0;
+    for (data_size_t qid = 0; qid < num_queries_ && data_idx < num_used_indices; ++qid) {
+      data_size_t start = fullset.query_boundaries_[qid];
+      data_size_t end = fullset.query_boundaries_[qid + 1];
+      data_size_t len = end - start;
+      if (used_indices[data_idx] > start) {
+        continue;
+      } else if (used_indices[data_idx] == start) {
+        if (num_used_indices >= data_idx + len && used_indices[data_idx + len - 1] == end - 1) {
+          used_query.push_back(qid);
+          data_idx += len;
+        } else {
+          Log::Fatal("Data partition error, data didn't match queries");
+        }
+      } else {
+        Log::Fatal("Data partition error, data didn't match queries");
+      }
+    }
+    query_boundaries_ = std::vector<data_size_t>(used_query.size() + 1);
+    num_queries_ = static_cast<data_size_t>(used_query.size());
+    query_boundaries_[0] = 0;
+    for (data_size_t i = 0; i < num_queries_; ++i) {
+      data_size_t qid = used_query[i];
+      data_size_t len = fullset.query_boundaries_[qid + 1] - fullset.query_boundaries_[qid];
+      query_boundaries_[i + 1] = query_boundaries_[i] + len;
+    }
+  } else {
+    num_queries_ = 0;
+  }
+
+}
+
 void Metadata::PartitionLabel(const std::vector<data_size_t>& used_indices) {
   if (used_indices.size() <= 0) {
     return;