gluonts.dataset.arrow.file 模块

class gluonts.dataset.arrow.file.ArrowFile(path: pathlib.Path, _start: int = 0, _take: Union[int, NoneType] = None)

基类: gluonts.dataset.arrow.file.File

property batch_offsets

decoder: gluonts.dataset.arrow.dec.ArrowDecoder

iter_batches()

location_for(idx)

metadata() → Dict[str, str]

path: pathlib.Path

reader: pyarrow.ipc.RecordBatchFileReader

property schema

class gluonts.dataset.arrow.file.ArrowStreamFile(path: pathlib.Path, _start: int = 0, _take: Union[int, NoneType] = None)

基类: gluonts.dataset.arrow.file.File

metadata() → Dict[str, str]

path: pathlib.Path

class gluonts.dataset.arrow.file.File

基类: object

SUFFIXES = {'.arrow', '.feather', '.parquet'}

static infer(path: pathlib.Path) → Union[gluonts.dataset.arrow.file.ArrowFile, gluonts.dataset.arrow.file.ArrowStreamFile, gluonts.dataset.arrow.file.ParquetFile]

通过检查提供的路径返回ArrowFile、ArrowStreamFile或ParquetFile。

Arrow 的 random-access 格式以 ARROW1 开头，因此我们查看提供的文件以查找它。

abstract metadata() → Dict[str, str]

class gluonts.dataset.arrow.file.ParquetFile(path: pathlib.Path, _start: int = 0, _take: Union[int, NoneType] = None, _row_group_sizes: List[int] = <factory>)

基类: gluonts.dataset.arrow.file.File

location_for(idx)

metadata() → Dict[str, str]

path: pathlib.Path

reader: pyarrow.parquet.core.ParquetFile