Expose DMatrix API for CUDA columnar and array. (dmlc#7217)

* Use JSON encoded configurations.
* Expose them into header file.

Author: trivialfis

include/xgboost/c_api.h

@@ -210,6 +210,36 @@ XGB_DLL int XGDMatrixCreateFromDT(void** data,
                                   DMatrixHandle* out,
                                   int nthread);
 
+/*!
+ * \brief Create DMatrix from CUDA columnar format. (cuDF)
+ * \param data Array of JSON encoded __cuda_array_interface__ for each column.
+ * \param json_config JSON encoded configuration.  Required values are:
+ *
+ *          - missing
+ *          - nthread
+ *
+ * \param out created dmatrix
+ * \return 0 when success, -1 when failure happens
+ */
+XGB_DLL int XGDMatrixCreateFromCudaColumnar(char const *data,
+                                            char const* json_config,
+                                            DMatrixHandle *out);
+
+/*!
+ * \brief Create DMatrix from CUDA array.
+ * \param data JSON encoded __cuda_array_interface__ for array data.
+ * \param json_config JSON encoded configuration.  Required values are:
+ *
+ *          - missing
+ *          - nthread
+ *
+ * \param out created dmatrix
+ * \return 0 when success, -1 when failure happens
+ */
+XGB_DLL int XGDMatrixCreateFromCudaArrayInterface(char const *data,
+                                                  char const* json_config,
+                                                  DMatrixHandle *out);
+
 /*
  * ========================== Begin data callback APIs =========================
  *

python-package/xgboost/data.py

@@ -421,17 +421,17 @@ def _transform_cudf_df(data, feature_names, feature_types, enable_categorical):
 
 def _from_cudf_df(
     data, missing, nthread, feature_names, feature_types, enable_categorical
-):
+) -> Tuple[ctypes.c_void_p, Any, Any]:
     data, feature_names, feature_types = _transform_cudf_df(
         data, feature_names, feature_types, enable_categorical
     )
     _, interfaces_str = _cudf_array_interfaces(data)
     handle = ctypes.c_void_p()
+    config = bytes(json.dumps({"missing": missing, "nthread": nthread}), "utf-8")
     _check_call(
-        _LIB.XGDMatrixCreateFromArrayInterfaceColumns(
+        _LIB.XGDMatrixCreateFromCudaColumnar(
             interfaces_str,
-            ctypes.c_float(missing),
-            ctypes.c_int(nthread),
+            config,
             ctypes.byref(handle),
         )
     )
@@ -469,11 +469,11 @@ def _from_cupy_array(data, missing, nthread, feature_names, feature_types):
     data = _transform_cupy_array(data)
     interface_str = _cuda_array_interface(data)
     handle = ctypes.c_void_p()
+    config = bytes(json.dumps({"missing": missing, "nthread": nthread}), "utf-8")
     _check_call(
-        _LIB.XGDMatrixCreateFromArrayInterface(
+        _LIB.XGDMatrixCreateFromCudaArrayInterface(
             interface_str,
-            ctypes.c_float(missing),
-            ctypes.c_int(nthread),
+            config,
             ctypes.byref(handle)))
     return handle, feature_names, feature_types
 

src/c_api/c_api.cc

@@ -169,19 +169,17 @@ XGB_DLL int XGDMatrixCreateFromDataIter(
 }
 
 #ifndef XGBOOST_USE_CUDA
-XGB_DLL int XGDMatrixCreateFromArrayInterfaceColumns(char const* c_json_strs,
-                                                     bst_float missing,
-                                                     int nthread,
-                                                     DMatrixHandle* out) {
+XGB_DLL int XGDMatrixCreateFromCudaColumnar(char const *data,
+                                            char const* c_json_config,
+                                            DMatrixHandle *out) {
   API_BEGIN();
   common::AssertGPUSupport();
   API_END();
 }
 
-XGB_DLL int XGDMatrixCreateFromArrayInterface(char const* c_json_strs,
-                                              bst_float missing,
-                                              int nthread,
-                                              DMatrixHandle* out) {
+XGB_DLL int XGDMatrixCreateFromCudaArrayInterface(char const *data,
+                                                  char const* c_json_config,
+                                                  DMatrixHandle *out) {
   API_BEGIN();
   common::AssertGPUSupport();
   API_END();

src/c_api/c_api.cu

@@ -26,23 +26,28 @@ void XGBoostAPIGuard::RestoreGPUAttribute() {
 
 using namespace xgboost;  // NOLINT
 
-XGB_DLL int XGDMatrixCreateFromArrayInterfaceColumns(char const* c_json_strs,
-                                                     bst_float missing,
-                                                     int nthread,
-                                                     DMatrixHandle* out) {
+XGB_DLL int XGDMatrixCreateFromCudaColumnar(char const *data,
+                                            char const* c_json_config,
+                                            DMatrixHandle *out) {
   API_BEGIN();
-  std::string json_str{c_json_strs};
+  std::string json_str{data};
+  auto config = Json::Load(StringView{c_json_config});
+  float missing = GetMissing(config);
+  auto nthread = get<Integer const>(config["nthread"]);
   data::CudfAdapter adapter(json_str);
   *out =
       new std::shared_ptr<DMatrix>(DMatrix::Create(&adapter, missing, nthread));
   API_END();
 }
 
-XGB_DLL int XGDMatrixCreateFromArrayInterface(char const* c_json_strs,
-                                              bst_float missing, int nthread,
-                                              DMatrixHandle* out) {
+XGB_DLL int XGDMatrixCreateFromCudaArrayInterface(char const *data,
+                                                  char const* c_json_config,
+                                                  DMatrixHandle *out) {
   API_BEGIN();
-  std::string json_str{c_json_strs};
+  std::string json_str{data};
+  auto config = Json::Load(StringView{c_json_config});
+  float missing = GetMissing(config);
+  auto nthread = get<Integer const>(config["nthread"]);
   data::CupyAdapter adapter(json_str);
   *out =
       new std::shared_ptr<DMatrix>(DMatrix::Create(&adapter, missing, nthread));