572 lines
21 KiB
Python
572 lines
21 KiB
Python
import torch
|
|
import copy
|
|
import pprint
|
|
import warnings
|
|
import numpy as np
|
|
from functools import reduce
|
|
from numbers import Number
|
|
from typing import Any, List, Tuple, Union, Iterator, Optional
|
|
|
|
# Disable pickle warning related to torch, since it has been removed
|
|
# on torch master branch. See Pull Request #39003 for details:
|
|
# https://github.com/pytorch/pytorch/pull/39003
|
|
warnings.filterwarnings(
|
|
"ignore", message="pickle support for Storage will be removed in 1.5.")
|
|
|
|
|
|
def _is_batch_set(data: Any) -> bool:
|
|
if isinstance(data, (list, tuple)):
|
|
if len(data) > 0 and isinstance(data[0], (dict, Batch)):
|
|
return True
|
|
elif isinstance(data, np.ndarray):
|
|
if isinstance(data.item(0), (dict, Batch)):
|
|
return True
|
|
return False
|
|
|
|
|
|
def _valid_bounds(length: int, index: Union[
|
|
slice, int, np.integer, np.ndarray, List[int]]) -> bool:
|
|
if isinstance(index, (int, np.integer)):
|
|
return -length <= index and index < length
|
|
elif isinstance(index, (list, np.ndarray)):
|
|
return _valid_bounds(length, np.min(index)) and \
|
|
_valid_bounds(length, np.max(index))
|
|
elif isinstance(index, slice):
|
|
if index.start is not None:
|
|
start_valid = _valid_bounds(length, index.start)
|
|
else:
|
|
start_valid = True
|
|
if index.stop is not None:
|
|
stop_valid = _valid_bounds(length, index.stop - 1)
|
|
else:
|
|
stop_valid = True
|
|
return start_valid and stop_valid
|
|
|
|
|
|
def _create_value(inst: Any, size: int) -> Union['Batch', np.ndarray]:
|
|
if isinstance(inst, np.ndarray):
|
|
return np.full((size, *inst.shape),
|
|
fill_value=None if inst.dtype == np.object else 0,
|
|
dtype=inst.dtype)
|
|
elif isinstance(inst, torch.Tensor):
|
|
return torch.full((size, *inst.shape),
|
|
fill_value=None if inst.dtype == np.object else 0,
|
|
device=inst.device,
|
|
dtype=inst.dtype)
|
|
elif isinstance(inst, (dict, Batch)):
|
|
zero_batch = Batch()
|
|
for key, val in inst.items():
|
|
zero_batch.__dict__[key] = _create_value(val, size)
|
|
return zero_batch
|
|
elif isinstance(inst, (np.generic, Number)):
|
|
return _create_value(np.asarray(inst), size)
|
|
else: # fall back to np.object
|
|
return np.array([None for _ in range(size)])
|
|
|
|
|
|
class Batch:
|
|
"""Tianshou provides :class:`~tianshou.data.Batch` as the internal data
|
|
structure to pass any kind of data to other methods, for example, a
|
|
collector gives a :class:`~tianshou.data.Batch` to policy for learning.
|
|
Here is the usage:
|
|
::
|
|
|
|
>>> import numpy as np
|
|
>>> from tianshou.data import Batch
|
|
>>> data = Batch(a=4, b=[5, 5], c='2312312')
|
|
>>> data.b
|
|
[5, 5]
|
|
>>> data.b = np.array([3, 4, 5])
|
|
>>> print(data)
|
|
Batch(
|
|
a: 4,
|
|
b: array([3, 4, 5]),
|
|
c: '2312312',
|
|
)
|
|
|
|
In short, you can define a :class:`Batch` with any key-value pair. The
|
|
current implementation of Tianshou typically use 7 reserved keys in
|
|
:class:`~tianshou.data.Batch`:
|
|
|
|
* ``obs`` the observation of step :math:`t` ;
|
|
* ``act`` the action of step :math:`t` ;
|
|
* ``rew`` the reward of step :math:`t` ;
|
|
* ``done`` the done flag of step :math:`t` ;
|
|
* ``obs_next`` the observation of step :math:`t+1` ;
|
|
* ``info`` the info of step :math:`t` (in ``gym.Env``, the ``env.step()``\
|
|
function return 4 arguments, and the last one is ``info``);
|
|
* ``policy`` the data computed by policy in step :math:`t`;
|
|
|
|
:class:`~tianshou.data.Batch` object can be initialized using wide variety
|
|
of arguments, starting with the key/value pairs or dictionary, but also
|
|
list and Numpy arrays of :class:`dict` or Batch instances. In which case,
|
|
each element is considered as an individual sample and get stacked
|
|
together:
|
|
::
|
|
|
|
>>> import numpy as np
|
|
>>> from tianshou.data import Batch
|
|
>>> data = Batch([{'a': {'b': [0.0, "info"]}}])
|
|
>>> print(data[0])
|
|
Batch(
|
|
a: Batch(
|
|
b: array(['0.0', 'info'], dtype='<U32'),
|
|
),
|
|
)
|
|
|
|
:class:`~tianshou.data.Batch` has the same API as a native Python
|
|
:class:`dict`. In this regard, one can access to stored data using string
|
|
key, or iterate over stored data:
|
|
::
|
|
|
|
>>> from tianshou.data import Batch
|
|
>>> data = Batch(a=4, b=[5, 5])
|
|
>>> print(data["a"])
|
|
4
|
|
>>> for key, value in data.items():
|
|
>>> print(f"{key}: {value}")
|
|
a: 4
|
|
b: [5, 5]
|
|
|
|
|
|
:class:`~tianshou.data.Batch` is also reproduce partially the Numpy API for
|
|
arrays. You can access or iterate over the individual samples, if any:
|
|
::
|
|
|
|
>>> import numpy as np
|
|
>>> from tianshou.data import Batch
|
|
>>> data = Batch(a=np.array([[0.0, 2.0], [1.0, 3.0]]), b=[5, -5])
|
|
>>> print(data[0])
|
|
Batch(
|
|
a: np.array([0.0, 2.0])
|
|
b: 5
|
|
)
|
|
>>> for sample in data:
|
|
>>> print(sample.a)
|
|
[0.0, 2.0]
|
|
[1.0, 3.0]
|
|
|
|
Similarly, one can also perform simple algebra on it, and stack, split or
|
|
concatenate multiple instances:
|
|
::
|
|
|
|
>>> import numpy as np
|
|
>>> from tianshou.data import Batch
|
|
>>> data_1 = Batch(a=np.array([0.0, 2.0]), b=5)
|
|
>>> data_2 = Batch(a=np.array([1.0, 3.0]), b=-5)
|
|
>>> data = Batch.stack((data_1, data_2))
|
|
>>> print(data)
|
|
Batch(
|
|
b: array([ 5, -5]),
|
|
a: array([[0., 2.],
|
|
[1., 3.]]),
|
|
)
|
|
>>> print(np.mean(data))
|
|
Batch(
|
|
b: 0.0,
|
|
a: array([0.5, 2.5]),
|
|
)
|
|
>>> data_split = list(data.split(1, False))
|
|
>>> print(list(data.split(1, False)))
|
|
[Batch(
|
|
b: [5],
|
|
a: array([[0., 2.]]),
|
|
),
|
|
Batch(
|
|
b: [-5],
|
|
a: array([[1., 3.]]),
|
|
)]
|
|
>>> data_cat = Batch.cat(data_split)
|
|
>>> print(data_cat)
|
|
Batch(
|
|
b: array([ 5, -5]),
|
|
a: array([[0., 2.],
|
|
[1., 3.]]),
|
|
)
|
|
|
|
Note that stacking of inconsistent data is also supported. In which case,
|
|
None is added in list or :class:`np.ndarray` of objects, 0 otherwise.
|
|
::
|
|
|
|
>>> import numpy as np
|
|
>>> from tianshou.data import Batch
|
|
>>> data_1 = Batch(a=np.array([0.0, 2.0]))
|
|
>>> data_2 = Batch(a=np.array([1.0, 3.0]), b='done')
|
|
>>> data = Batch.stack((data_1, data_2))
|
|
>>> print(data)
|
|
Batch(
|
|
a: array([[0., 2.],
|
|
[1., 3.]]),
|
|
b: array([None, 'done'], dtype=object),
|
|
)
|
|
|
|
:meth:`~tianshou.data.Batch.size` and :meth:`~tianshou.data.Batch.__len__`
|
|
methods are also provided to respectively get the length and the size of
|
|
a :class:`Batch` instance. It mimics the Numpy API for Numpy arrays, which
|
|
means that getting the length of a scalar Batch raises an exception, while
|
|
the size is 1. The size is only 0 if empty. Note that the size and length
|
|
are the identical if multiple samples are stored:
|
|
::
|
|
|
|
>>> import numpy as np
|
|
>>> from tianshou.data import Batch
|
|
>>> data = Batch(a=[5., 4.], b=np.zeros((2, 3, 4)))
|
|
>>> data.size
|
|
2
|
|
>>> len(data)
|
|
2
|
|
>>> data[0].size
|
|
1
|
|
>>> len(data[0])
|
|
TypeError: Object of type 'Batch' has no len()
|
|
|
|
Convenience helpers are available to convert in-place the stored data into
|
|
Numpy arrays or Torch tensors.
|
|
|
|
Finally, note that :class:`~tianshou.data.Batch` instance are serializable
|
|
and therefore Pickle compatible. This is especially important for
|
|
distributed sampling.
|
|
"""
|
|
|
|
def __init__(self,
|
|
batch_dict: Optional[Union[
|
|
dict, 'Batch', Tuple[Union[dict, 'Batch']],
|
|
List[Union[dict, 'Batch']], np.ndarray]] = None,
|
|
**kwargs) -> None:
|
|
if _is_batch_set(batch_dict):
|
|
self.stack_(batch_dict)
|
|
elif isinstance(batch_dict, (dict, Batch)):
|
|
for k, v in batch_dict.items():
|
|
if isinstance(v, dict) or _is_batch_set(v):
|
|
self.__dict__[k] = Batch(v)
|
|
else:
|
|
self.__dict__[k] = v
|
|
if len(kwargs) > 0:
|
|
self.__init__(kwargs)
|
|
|
|
def __getstate__(self):
|
|
"""Pickling interface. Only the actual data are serialized
|
|
for both efficiency and simplicity.
|
|
"""
|
|
state = {}
|
|
for k, v in self.items():
|
|
if isinstance(v, Batch):
|
|
v = v.__getstate__()
|
|
state[k] = v
|
|
return state
|
|
|
|
def __setstate__(self, state):
|
|
"""Unpickling interface. At this point, self is an empty Batch
|
|
instance that has not been initialized, so it can safely be
|
|
initialized by the pickle state.
|
|
"""
|
|
self.__init__(**state)
|
|
|
|
def __getitem__(self, index: Union[
|
|
str, slice, int, np.integer, np.ndarray, List[int]]) -> 'Batch':
|
|
"""Return self[index]."""
|
|
if isinstance(index, str):
|
|
return self.__dict__[index]
|
|
|
|
if not _valid_bounds(len(self), index):
|
|
raise IndexError(
|
|
f"Index {index} out of bounds for Batch of len {len(self)}.")
|
|
else:
|
|
b = Batch()
|
|
is_index_scalar = isinstance(index, (int, np.integer)) or \
|
|
(isinstance(index, np.ndarray) and index.ndim == 0)
|
|
for k, v in self.items():
|
|
if isinstance(v, Batch) and len(v.__dict__) == 0:
|
|
b.__dict__[k] = Batch()
|
|
elif is_index_scalar or not isinstance(v, list):
|
|
b.__dict__[k] = v[index]
|
|
else:
|
|
b.__dict__[k] = [v[i] for i in index]
|
|
return b
|
|
|
|
def __setitem__(self, index: Union[
|
|
str, slice, int, np.integer, np.ndarray, List[int]],
|
|
value: Any) -> None:
|
|
"""Assign value to self[index]."""
|
|
if isinstance(index, str):
|
|
self.__dict__[index] = value
|
|
return
|
|
if not isinstance(value, (dict, Batch)):
|
|
raise TypeError("Batch does not supported value type "
|
|
f"{type(value)} for item assignment.")
|
|
if not set(value.keys()).issubset(self.__dict__.keys()):
|
|
raise KeyError(
|
|
"Creating keys is not supported by item assignment.")
|
|
for key, val in self.items():
|
|
try:
|
|
self.__dict__[key][index] = value[key]
|
|
except KeyError:
|
|
if isinstance(val, Batch):
|
|
self.__dict__[key][index] = Batch()
|
|
elif isinstance(val, np.ndarray) and \
|
|
val.dtype == np.integer:
|
|
# Fallback for np.array of integer,
|
|
# since neither None or nan is supported.
|
|
self.__dict__[key][index] = 0
|
|
else:
|
|
self.__dict__[key][index] = None
|
|
|
|
def __iadd__(self, other: Union['Batch', Number]):
|
|
"""Algebraic addition with another :class:`~tianshou.data.Batch`
|
|
instance in-place."""
|
|
if isinstance(other, Batch):
|
|
for (k, r), v in zip(self.__dict__.items(),
|
|
other.__dict__.values()):
|
|
if r is None:
|
|
continue
|
|
elif isinstance(r, list):
|
|
self.__dict__[k] = [r_ + v_ for r_, v_ in zip(r, v)]
|
|
else:
|
|
self.__dict__[k] += v
|
|
return self
|
|
elif isinstance(other, Number):
|
|
for k, r in self.items():
|
|
if r is None:
|
|
continue
|
|
elif isinstance(r, list):
|
|
self.__dict__[k] = [r_ + other for r_ in r]
|
|
else:
|
|
self.__dict__[k] += other
|
|
return self
|
|
else:
|
|
raise TypeError("Only addition of Batch or number is supported.")
|
|
|
|
def __add__(self, other: Union['Batch', Number]):
|
|
"""Algebraic addition with another :class:`~tianshou.data.Batch`
|
|
instance out-of-place."""
|
|
return copy.deepcopy(self).__iadd__(other)
|
|
|
|
def __imul__(self, val: Number):
|
|
"""Algebraic multiplication with a scalar value in-place."""
|
|
assert isinstance(val, Number), \
|
|
"Only multiplication by a number is supported."
|
|
for k in self.__dict__.keys():
|
|
self.__dict__[k] *= val
|
|
return self
|
|
|
|
def __mul__(self, val: Number):
|
|
"""Algebraic multiplication with a scalar value out-of-place."""
|
|
return copy.deepcopy(self).__imul__(val)
|
|
|
|
def __itruediv__(self, val: Number):
|
|
"""Algebraic division wibyth a scalar value in-place."""
|
|
assert isinstance(val, Number), \
|
|
"Only division by a number is supported."
|
|
for k in self.__dict__.keys():
|
|
self.__dict__[k] /= val
|
|
return self
|
|
|
|
def __truediv__(self, val: Number):
|
|
"""Algebraic division wibyth a scalar value out-of-place."""
|
|
return copy.deepcopy(self).__itruediv__(val)
|
|
|
|
def __repr__(self) -> str:
|
|
"""Return str(self)."""
|
|
s = self.__class__.__name__ + '(\n'
|
|
flag = False
|
|
for k, v in self.items():
|
|
rpl = '\n' + ' ' * (6 + len(k))
|
|
obj = pprint.pformat(v).replace('\n', rpl)
|
|
s += f' {k}: {obj},\n'
|
|
flag = True
|
|
if flag:
|
|
s += ')'
|
|
else:
|
|
s = self.__class__.__name__ + '()'
|
|
return s
|
|
|
|
def keys(self) -> List[str]:
|
|
"""Return self.keys()."""
|
|
return self.__dict__.keys()
|
|
|
|
def values(self) -> List[Any]:
|
|
"""Return self.values()."""
|
|
return self.__dict__.values()
|
|
|
|
def items(self) -> List[Tuple[str, Any]]:
|
|
"""Return self.items()."""
|
|
return self.__dict__.items()
|
|
|
|
def get(self, k: str, d: Optional[Any] = None) -> Union['Batch', Any]:
|
|
"""Return self[k] if k in self else d. d defaults to None."""
|
|
return self.__dict__.get(k, d)
|
|
|
|
def to_numpy(self) -> None:
|
|
"""Change all torch.Tensor to numpy.ndarray. This is an in-place
|
|
operation.
|
|
"""
|
|
for k, v in self.items():
|
|
if isinstance(v, torch.Tensor):
|
|
self.__dict__[k] = v.detach().cpu().numpy()
|
|
elif isinstance(v, Batch):
|
|
v.to_numpy()
|
|
|
|
def to_torch(self,
|
|
dtype: Optional[torch.dtype] = None,
|
|
device: Union[str, int, torch.device] = 'cpu'
|
|
) -> None:
|
|
"""Change all numpy.ndarray to torch.Tensor. This is an in-place
|
|
operation.
|
|
"""
|
|
if not isinstance(device, torch.device):
|
|
device = torch.device(device)
|
|
|
|
for k, v in self.items():
|
|
if isinstance(v, (np.generic, np.ndarray)):
|
|
v = torch.from_numpy(v).to(device)
|
|
if dtype is not None:
|
|
v = v.type(dtype)
|
|
self.__dict__[k] = v
|
|
if isinstance(v, torch.Tensor):
|
|
if dtype is not None and v.dtype != dtype:
|
|
must_update_tensor = True
|
|
elif v.device.type != device.type:
|
|
must_update_tensor = True
|
|
elif device.index is not None and \
|
|
device.index != v.device.index:
|
|
must_update_tensor = True
|
|
else:
|
|
must_update_tensor = False
|
|
if must_update_tensor:
|
|
if dtype is not None:
|
|
v = v.type(dtype)
|
|
self.__dict__[k] = v.to(device)
|
|
elif isinstance(v, Batch):
|
|
v.to_torch(dtype, device)
|
|
|
|
def append(self, batch: 'Batch') -> None:
|
|
warnings.warn('Method append will be removed soon, please use '
|
|
':meth:`~tianshou.data.Batch.cat`')
|
|
return self.cat_(batch)
|
|
|
|
def cat_(self, batch: 'Batch') -> None:
|
|
"""Concatenate a :class:`~tianshou.data.Batch` object into
|
|
current batch.
|
|
"""
|
|
assert isinstance(batch, Batch), \
|
|
'Only Batch is allowed to be concatenated in-place!'
|
|
for k, v in batch.items():
|
|
if v is None:
|
|
continue
|
|
if not hasattr(self, k) or self.__dict__[k] is None:
|
|
self.__dict__[k] = copy.deepcopy(v)
|
|
elif isinstance(v, np.ndarray) and v.ndim > 0:
|
|
self.__dict__[k] = np.concatenate([self.__dict__[k], v])
|
|
elif isinstance(v, torch.Tensor):
|
|
self.__dict__[k] = torch.cat([self.__dict__[k], v])
|
|
elif isinstance(v, list):
|
|
self.__dict__[k] += copy.deepcopy(v)
|
|
elif isinstance(v, Batch):
|
|
self.__dict__[k].cat_(v)
|
|
else:
|
|
s = 'No support for method "cat" with type '\
|
|
f'{type(v)} in class Batch.'
|
|
raise TypeError(s)
|
|
|
|
@classmethod
|
|
def cat(cls, batches: List['Batch']) -> 'Batch':
|
|
"""Concatenate a :class:`~tianshou.data.Batch` object into a
|
|
single new batch.
|
|
"""
|
|
batch = cls()
|
|
for batch_ in batches:
|
|
batch.cat_(batch_)
|
|
return batch
|
|
|
|
def stack_(self,
|
|
batches: List[Union[dict, 'Batch']],
|
|
axis: int = 0) -> None:
|
|
"""Stack a :class:`~tianshou.data.Batch` object i into current
|
|
batch.
|
|
"""
|
|
if len(self.__dict__) > 0:
|
|
batches = [self] + list(batches)
|
|
keys_map = list(map(lambda e: set(e.keys()), batches))
|
|
keys_shared = set.intersection(*keys_map)
|
|
values_shared = [
|
|
[e[k] for e in batches] for k in keys_shared]
|
|
for k, v in zip(keys_shared, values_shared):
|
|
if isinstance(v[0], (dict, Batch)):
|
|
self.__dict__[k] = Batch.stack(v, axis)
|
|
elif isinstance(v[0], torch.Tensor):
|
|
self.__dict__[k] = torch.stack(v, axis)
|
|
else:
|
|
self.__dict__[k] = np.stack(v, axis)
|
|
keys_partial = reduce(set.symmetric_difference, keys_map)
|
|
for k in keys_partial:
|
|
for i, e in enumerate(batches):
|
|
val = e.get(k, None)
|
|
if val is not None:
|
|
try:
|
|
self.__dict__[k][i] = val
|
|
except KeyError:
|
|
self.__dict__[k] = \
|
|
_create_value(val, len(batches))
|
|
self.__dict__[k][i] = val
|
|
|
|
@staticmethod
|
|
def stack(batches: List['Batch'], axis: int = 0) -> 'Batch':
|
|
"""Stack a :class:`~tianshou.data.Batch` object into a
|
|
single new batch.
|
|
"""
|
|
batch = Batch()
|
|
batch.stack_(batches, axis)
|
|
return batch
|
|
|
|
def __len__(self) -> int:
|
|
"""Return len(self)."""
|
|
r = []
|
|
for v in self.__dict__.values():
|
|
if isinstance(v, Batch) and len(v.__dict__) == 0:
|
|
continue
|
|
elif hasattr(v, '__len__') and (not isinstance(
|
|
v, (np.ndarray, torch.Tensor)) or v.ndim > 0):
|
|
r.append(len(v))
|
|
else:
|
|
raise TypeError("Object of type 'Batch' has no len()")
|
|
if len(r) == 0:
|
|
raise TypeError("Object of type 'Batch' has no len()")
|
|
return min(r)
|
|
|
|
@property
|
|
def size(self) -> int:
|
|
"""Return self.size."""
|
|
if len(self.__dict__.keys()) == 0:
|
|
return 0
|
|
else:
|
|
r = []
|
|
for v in self.__dict__.values():
|
|
if isinstance(v, Batch):
|
|
r.append(v.size)
|
|
elif hasattr(v, '__len__') and (not isinstance(
|
|
v, (np.ndarray, torch.Tensor)) or v.ndim > 0):
|
|
r.append(len(v))
|
|
else:
|
|
r.append(1)
|
|
return min(r) if len(r) > 0 else 0
|
|
|
|
def split(self, size: Optional[int] = None,
|
|
shuffle: bool = True) -> Iterator['Batch']:
|
|
"""Split whole data into multiple small batch.
|
|
|
|
:param int size: if it is ``None``, it does not split the data batch;
|
|
otherwise it will divide the data batch with the given size.
|
|
Default to ``None``.
|
|
:param bool shuffle: randomly shuffle the entire data batch if it is
|
|
``True``, otherwise remain in the same. Default to ``True``.
|
|
"""
|
|
length = len(self)
|
|
if size is None:
|
|
size = length
|
|
if shuffle:
|
|
indices = np.random.permutation(length)
|
|
else:
|
|
indices = np.arange(length)
|
|
for idx in np.arange(0, length, size):
|
|
yield self[indices[idx:(idx + size)]]
|