-
Notifications
You must be signed in to change notification settings - Fork 38
/
datasets.py
331 lines (285 loc) · 12.2 KB
/
datasets.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
from collections.abc import Iterable
from typing import TYPE_CHECKING, Any
import torch
from datasets import Dataset, IterableDataset
from datasets.features.features import Features, Value
from datasets.fingerprint import Hasher
from pandas import DataFrame
from ..datasets.utils import get_column_names
from ..pickling import unpickle_transform
if TYPE_CHECKING: # pragma: no cover
from ..steps import Step
class OutputDatasetMixin:
@property
def step(self) -> "Step":
"""The step that produced the dataset."""
return self._step # type:ignore[attr-defined]
@property
def column_names(self) -> list[str]:
"""The column names in the dataset."""
return get_column_names(self.dataset) # type:ignore[attr-defined]
@property
def num_columns(self) -> int:
"""The number of columns in the dataset."""
return len(self.column_names)
@property
def info(self) -> Any:
return self.dataset.info # type:ignore[attr-defined]
@property
def _features(self) -> Features:
if self.info and self.info.features:
return self.info.features
else:
return Features()
def __iter__(self):
if self._pickled or self._pickled_inferred: # type:ignore[attr-defined]
for row in iter(self.dataset): # type:ignore[attr-defined]
yield unpickle_transform(row, features=self._features, batched=False)
else:
yield from iter(self.dataset) # type:ignore[attr-defined]
def __getitem__(self, key: int | slice | str | Iterable[int]) -> Any:
"""Get a row or column from the dataset.
Args:
key: The index or name of the column to get.
Returns:
The row or column from the dataset.
"""
if isinstance(key, str):
feature = self._features.get(key, None)
feature_is_pickle_type = False
if isinstance(feature, Value) and feature.dtype == "binary":
feature_is_pickle_type = True
if isinstance(self.dataset, Dataset): # type:ignore[attr-defined]
return OutputDatasetColumn(
self._step, # type:ignore[attr-defined]
self.dataset.select_columns([key]), # type:ignore[attr-defined]
pickled=self._pickled # type:ignore[attr-defined]
and feature_is_pickle_type,
)
else:
return OutputIterableDatasetColumn(
self._step, # type:ignore[attr-defined]
self.dataset.select_columns([key]), # type:ignore[attr-defined]
pickled=self._pickled # type:ignore[attr-defined]
and feature_is_pickle_type,
total_num_rows=self.total_num_rows, # type:ignore[attr-defined]
)
if self._pickled or self._pickled_inferred: # type:ignore[attr-defined]
if isinstance(key, int):
return unpickle_transform(
self.dataset[key], # type:ignore[attr-defined]
features=self._features,
batched=False,
)
else:
return unpickle_transform(
self.dataset[key], # type:ignore[attr-defined]
features=self._features,
batched=True,
)
else:
return self.dataset[key] # type:ignore[attr-defined]
@property
def fingerprint(self) -> Any:
return Hasher.hash((self.step.fingerprint, self.column_names))
def head(self, n=5, shuffle=False, seed=None, buffer_size=1000) -> DataFrame:
if isinstance(self.dataset, Dataset): # type:ignore[attr-defined]
iterable_dataset = (
self.dataset.to_iterable_dataset() # type:ignore[attr-defined]
)
else:
iterable_dataset = self.dataset # type:ignore[attr-defined]
if shuffle:
iterable_dataset = iterable_dataset.shuffle(
seed=seed, buffer_size=buffer_size
)
return DataFrame.from_records(list(iterable_dataset.take(n)))
def __repr__(self) -> str:
if isinstance(self, OutputDataset):
return (
f"{type(self).__name__}("
f"column_names={str(self.column_names)}, "
f"num_rows={len(self)}, "
f"dataset=<{type(self.dataset).__name__} @ {id(self.dataset)}>"
")"
)
elif isinstance(self, OutputIterableDataset):
return (
f"{type(self).__name__}("
f"column_names={str(self.column_names)}, "
f"num_rows={str(self.total_num_rows).replace('None', 'Unknown')}, "
f"dataset=<{type(self.dataset).__name__} @ {id(self.dataset)}>"
")"
)
else:
return super().__repr__() # pragma: no cover
class OutputDatasetColumnMixin:
def __iter__(self):
column_name = self.column_names[0] # type:ignore[attr-defined]
for row in super().__iter__(): # type:ignore[misc]
yield row[column_name]
def __getitem__(self, key: int | slice | str | Iterable[int]) -> Any:
column_name = self.column_names[0] # type:ignore[attr-defined]
if isinstance(key, str):
if isinstance(self, OutputIterableDatasetColumn):
if key != column_name:
raise KeyError(
f"Column '{key}' is not valid. This OutputIterableDatasetColumn"
f" object only has a single column named '{column_name}'."
)
return iter(self)
else:
dataset = self.dataset # type:ignore[attr-defined]
if self._pickled or self._pickled_inferred: # type:ignore[attr-defined]
return (
unpickle_transform(
{key: dataset[key]},
features=self._features, # type:ignore[attr-defined]
batched=True,
)
)[key]
else:
return dataset[key]
else:
return super().__getitem__(key)[column_name] # type:ignore[misc]
def __repr__(self) -> str:
if isinstance(self, OutputDatasetColumn):
return (
f"{type(self).__name__}("
f"column_name={repr(self.column_names[0])}, "
f"num_rows={len(self)}, "
f"dataset=<{type(self.dataset).__name__} @ {id(self.dataset)}>"
")"
)
elif isinstance(self, OutputIterableDatasetColumn):
return (
f"{type(self).__name__}("
f"column_name={repr(self.column_names[0])}, "
f"num_rows={str(self.total_num_rows).replace('None', 'Unknown')}, "
f"dataset=<{type(self.dataset).__name__} @ {id(self.dataset)}>"
")"
)
else:
return super().__repr__() # pragma: no cover
class OutputIterableDataset(OutputDatasetMixin):
def __init__(
self,
step: "Step",
dataset: IterableDataset,
pickled: bool = False,
total_num_rows: None | int = None,
):
from ..steps import Step
if not isinstance(step, Step):
raise ValueError(f"Expected Step, got {type(step)}.")
if not isinstance(dataset, IterableDataset):
raise ValueError(f"Expected IterableDataset, got {type(dataset)}.")
self._step: "Step" = step
self._dataset: IterableDataset = dataset
self._pickled: bool = pickled
self._pickled_inferred: bool = False
self.total_num_rows: None | int = total_num_rows
for f in self._features.values():
if isinstance(f, Value) and f.dtype == "binary":
self._pickled_inferred = True
break
@property
def dataset(self) -> IterableDataset:
"""The underlying Hugging Face :py:class:`~datasets.IterableDataset`."""
return self._dataset
@property
def num_rows(self) -> None | int:
"""The number of rows in the dataset."""
return self.total_num_rows
class OutputDataset(OutputDatasetMixin):
def __init__(self, step: "Step", dataset: Dataset, pickled: bool = False):
from ..steps import Step
if not isinstance(step, Step):
raise ValueError(f"Expected Step, got {type(step)}.")
if not isinstance(dataset, Dataset):
raise ValueError(f"Expected Dataset, got {type(dataset)}.")
self._step: "Step" = step
self._dataset: Dataset = dataset
self._pickled: bool = pickled
self._pickled_inferred: bool = False
@property
def dataset(self) -> Dataset:
"""The underlying Hugging Face :py:class:`~datasets.Dataset`."""
return self._dataset
def save_to_disk(
self, path: str, num_proc: None | int, num_shards: None | int
) -> None:
self._dataset.save_to_disk(
path,
num_proc=min(num_proc if num_proc is not None else 1, len(self._dataset)),
num_shards=num_shards,
)
self._dataset = Dataset.load_from_disk(path)
@property
def num_rows(self) -> int:
"""The number of rows in the dataset."""
return len(self)
def __len__(self):
return len(self.dataset)
class OutputIterableDatasetColumn(OutputDatasetColumnMixin, OutputIterableDataset):
def __init__(
self,
step: "Step",
dataset: IterableDataset,
pickled: bool = False,
total_num_rows: None | int = None,
):
from ..steps import Step
if not isinstance(step, Step):
raise ValueError(f"Expected Step, got {type(step)}.")
if not isinstance(dataset, IterableDataset):
raise ValueError(f"Expected IterableDataset, got {type(dataset)}.")
self._step: "Step" = step
self._dataset: IterableDataset = dataset
self._pickled: bool = pickled
self._pickled_inferred: bool = False
self.total_num_rows: None | int = total_num_rows
for f in self._features.values():
if isinstance(f, Value) and f.dtype == "binary":
self._pickled_inferred = True
break
if len(self.column_names) != 1:
raise ValueError(f"Expected single column only, got {self.column_names}")
class OutputDatasetColumn(OutputDatasetColumnMixin, OutputDataset):
def __init__(self, step: "Step", dataset: Dataset, pickled: bool = False):
from ..steps import Step
if not isinstance(step, Step):
raise ValueError(f"Expected Step, got {type(step)}.")
if not isinstance(dataset, Dataset):
raise ValueError(f"Expected Dataset, got {type(dataset)}.")
self._step: "Step" = step
self._dataset: Dataset = dataset
self._pickled: bool = pickled
self._pickled_inferred: bool = False
if len(self.column_names) != 1:
raise ValueError(f"Expected single column only, got {self.column_names}")
class _SizedIterableDataset(torch.utils.data.IterableDataset):
def __init__(self, dataset: IterableDataset, total_num_rows: int):
self.dataset = dataset
self.total_num_rows = total_num_rows
@property
def features(self): # pragma: no cover
return self.dataset.features # type:ignore[attr-defined]
def cast_column(
self, *args, **kwargs
) -> "_SizedIterableDataset": # pragma: no cover
return _SizedIterableDataset(
dataset=self.dataset.cast_column(*args, **kwargs),
total_num_rows=self.total_num_rows,
)
def __iter__(self):
return iter(self.dataset)
def __len__(self) -> int:
return self.total_num_rows
def get_sized_dataset(
dataset: Dataset | IterableDataset, total_num_rows: None | int
) -> Dataset | _SizedIterableDataset | IterableDataset:
if isinstance(dataset, IterableDataset) and total_num_rows is not None:
return _SizedIterableDataset(dataset=dataset, total_num_rows=total_num_rows)
else:
return dataset