from __future__ import annotations
from copy import deepcopy
from pathlib import PurePosixPath
from typing import Any
import fsspec
import torch
from kedro.io.core import (
AbstractVersionedDataset,
DatasetError,
Version,
get_filepath_str,
get_protocol_and_path,
)
[docs]
class PyTorchDataset(AbstractVersionedDataset[Any, Any]):
"""``PyTorchDataset`` loads and saves PyTorch models' `state_dict`
using PyTorch's recommended zipfile serialization protocol. To avoid
security issues with Pickle.
.. code-block:: yaml
model:
type: pytorch.PyTorchDataset
filepath: data/06_models/model.pt
.. code-block:: pycon
>>> from kedro_datasets_experimental.pytorch import PyTorchDataset
>>> import torch
>>>
>>> model: torch.nn.Module
>>> model = torch.nn.Sequential(torch.nn.Linear(10, 10), torch.nn.ReLU())
>>> dataset = PyTorchDataset(filepath=tmp_path / "model.pt")
>>> dataset.save(model)
>>> reloaded = TheModelClass(*args, **kwargs)
>>> reloaded.load_state_dict(dataset.load())
"""
DEFAULT_LOAD_ARGS: dict[str, Any] = {}
DEFAULT_SAVE_ARGS: dict[str, Any] = {}
def __init__( # noqa: PLR0913
self,
*,
filepath,
load_args: dict[str, Any] = None,
save_args: dict[str, Any] = None,
version: Version | None = None,
credentials: dict[str, Any] = None,
fs_args: dict[str, Any] = None,
metadata: dict[str, Any] = None,
):
_fs_args = deepcopy(fs_args) or {}
_fs_open_args_load = _fs_args.pop("open_args_load", {})
_fs_open_args_save = _fs_args.pop("open_args_save", {})
_credentials = deepcopy(credentials) or {}
protocol, path = get_protocol_and_path(filepath, version)
if protocol == "file":
_fs_args.setdefault("auto_mkdir", True)
self._protocol = protocol
self._fs = fsspec.filesystem(self._protocol, **_credentials, **_fs_args)
self.metadata = metadata
super().__init__(
filepath=PurePosixPath(path),
version=version,
exists_function=self._fs.exists,
glob_function=self._fs.glob,
)
# Handle default load and save arguments
self._load_args = deepcopy(self.DEFAULT_LOAD_ARGS)
if load_args is not None:
self._load_args.update(load_args)
self._save_args = deepcopy(self.DEFAULT_SAVE_ARGS)
if save_args is not None:
self._save_args.update(save_args)
self._fs_open_args_load = _fs_open_args_load
self._fs_open_args_save = _fs_open_args_save
def _describe(self) -> dict[str, Any]:
return {
"filepath": self._filepath,
"protocol": self._protocol,
"load_args": self._load_args,
"save_args": self._save_args,
"version": self._version,
}
[docs]
def load(self) -> Any:
load_path = get_filepath_str(self._get_load_path(), self._protocol)
return torch.load(load_path, **self._fs_open_args_load) #nosec: B614
[docs]
def save(self, data: torch.nn.Module) -> None:
save_path = get_filepath_str(self._get_save_path(), self._protocol)
torch.save(data.state_dict(), save_path, **self._fs_open_args_save) #nosec: B614
self._invalidate_cache()
def _exists(self):
try:
load_path = get_filepath_str(self._get_load_path(), self._protocol)
except DatasetError:
return False
return self._fs.exists(load_path)
def _release(self) -> None:
super()._release()
self._invalidate_cache()
def _invalidate_cache(self) -> None:
"""Invalidate underlying filesystem caches."""
filepath = get_filepath_str(self._filepath, self._protocol)
self._fs.invalidate_cache(filepath)
