Test constructing DMatrix from cudf
(
input_type: Any, DMatrixT: Type[xgb.DMatrix], missing: float = np.nan
)
| 22 | |
| 23 | |
| 24 | def dmatrix_from_cudf( |
| 25 | input_type: Any, DMatrixT: Type[xgb.DMatrix], missing: float = np.nan |
| 26 | ) -> None: |
| 27 | """Test constructing DMatrix from cudf""" |
| 28 | import pandas as pd |
| 29 | |
| 30 | kRows = 80 |
| 31 | kCols = 3 |
| 32 | |
| 33 | na = np.random.randn(kRows, kCols) |
| 34 | na[:, 0:2] = na[:, 0:2].astype(input_type) |
| 35 | |
| 36 | na[5, 0] = missing |
| 37 | na[3, 1] = missing |
| 38 | |
| 39 | pa = pd.DataFrame({"0": na[:, 0], "1": na[:, 1], "2": na[:, 2].astype(np.int32)}) |
| 40 | |
| 41 | np_label = np.random.randn(kRows).astype(input_type) |
| 42 | pa_label = pd.DataFrame(np_label) |
| 43 | |
| 44 | cd = cudf.from_pandas(pa) |
| 45 | cd_label = cudf.from_pandas(pa_label).iloc[:, 0] |
| 46 | |
| 47 | dtrain = DMatrixT(cd, missing=missing, label=cd_label) |
| 48 | assert dtrain.num_col() == kCols |
| 49 | assert dtrain.num_row() == kRows |
| 50 | |
| 51 | |
| 52 | def _test_from_cudf(DMatrixT: Type[xgb.DMatrix]) -> None: |
no test coverage detected