-
Notifications
You must be signed in to change notification settings - Fork 154
/
autodecode.py
569 lines (456 loc) · 15.1 KB
/
autodecode.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
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
#
# Copyright (c) 2017-2021 NVIDIA CORPORATION. All rights reserved.
# This file is part of the WebDataset library.
# See the LICENSE file for licensing terms (BSD-style).
#
"""Automatically decode webdataset samples."""
import io
import json
import os
import pickle
import re
import tempfile
from functools import partial
import numpy as np
# Obtained with:
# ```
# import PIL.Image
# IMAGE_EXTENSIONS = []
# PIL.Image.init()
# for ext, format in PIL.Image.EXTENSION.items():
# if format in PIL.Image.OPEN:
# IMAGE_EXTENSIONS.append(ext[1:])
# ```
"""Extensions passed on to the image decoder."""
IMAGE_EXTENSIONS = [
"blp",
"bmp",
"dib",
"bufr",
"cur",
"pcx",
"dcx",
"dds",
"ps",
"eps",
"fit",
"fits",
"fli",
"flc",
"ftc",
"ftu",
"gbr",
"gif",
"grib",
"h5",
"hdf",
"png",
"apng",
"jp2",
"j2k",
"jpc",
"jpf",
"jpx",
"j2c",
"icns",
"ico",
"im",
"iim",
"tif",
"tiff",
"jfif",
"jpe",
"jpg",
"jpeg",
"mpg",
"mpeg",
"msp",
"pcd",
"pxr",
"pbm",
"pgm",
"ppm",
"pnm",
"psd",
"bw",
"rgb",
"rgba",
"sgi",
"ras",
"tga",
"icb",
"vda",
"vst",
"webp",
"wmf",
"emf",
"xbm",
"xpm",
]
################################################################
# handle basic datatypes
################################################################
def torch_loads(data: bytes):
"""Function: torch_loads
Description:
This function loads data using torch.loads. It first imports torch only if necessary. Then it decodes the input data using torch.load.
Parameters:
- data (bytes): The data to be decoded.
Returns:
It returns the decoded input data.
Example:
data = b'...'
output = torch_loads(data)
"""
import io
import torch
stream = io.BytesIO(data)
return torch.load(stream)
def tenbin_loads(data):
from . import tenbin
return tenbin.decode_buffer(data)
def msgpack_loads(data):
import msgpack
return msgpack.unpackb(data)
def npy_loads(data):
import numpy.lib.format
stream = io.BytesIO(data)
return numpy.lib.format.read_array(stream)
def npz_loads(data):
import numpy.lib.format
stream = io.BytesIO(data)
return dict(np.load(stream))
def cbor_loads(data):
import cbor
return cbor.loads(data)
decoders = {
"txt": lambda data: data.decode("utf-8"),
"text": lambda data: data.decode("utf-8"),
"transcript": lambda data: data.decode("utf-8"),
"cls": lambda data: int(data),
"cls2": lambda data: int(data),
"index": lambda data: int(data),
"inx": lambda data: int(data),
"id": lambda data: int(data),
"json": lambda data: json.loads(data),
"jsn": lambda data: json.loads(data),
"pyd": lambda data: pickle.loads(data),
"pickle": lambda data: pickle.loads(data),
"pth": lambda data: torch_loads(data),
"ten": tenbin_loads,
"tb": tenbin_loads,
"mp": msgpack_loads,
"msg": msgpack_loads,
"npy": npy_loads,
"npz": npz_loads,
"cbor": cbor_loads,
}
def basichandlers(key, data):
"""Handle basic file decoding.
This function is usually part of the post= decoders.
This handles the following forms of decoding:
- txt -> unicode string
- cls cls2 class count index inx id -> int
- json jsn -> JSON decoding
- pyd pickle -> pickle decoding
- pth -> torch.loads
- ten tenbin -> fast tensor loading
- mp messagepack msg -> messagepack decoding
- npy -> Python NPY decoding
:param key: file name extension
:param data: binary data to be decoded
"""
extension = re.sub(r".*[.]", "", key)
if extension in decoders:
return decoders[extension](data)
return None
################################################################
# Generic extension handler.
################################################################
def call_extension_handler(key, data, f, extensions):
"""Call the function f with the given data if the key matches the extensions.
:param key: actual key found in the sample
:param data: binary data
:param f: decoder function
:param extensions: list of matching extensions
"""
extension = key.lower().split(".")
for target in extensions:
target = target.split(".")
if len(target) > len(extension):
continue
if extension[-len(target) :] == target:
return f(data)
return None
def handle_extension(extensions, f):
"""Return a decoder function for the list of extensions.
Extensions can be a space separated list of extensions.
Extensions can contain dots, in which case the corresponding number
of extension components must be present in the key given to f.
Comparisons are case insensitive.
Examples:
handle_extension("jpg jpeg", my_decode_jpg) # invoked for any file.jpg
handle_extension("seg.jpg", special_case_jpg) # invoked only for file.seg.jpg
"""
extensions = extensions.lower().split()
return partial(call_extension_handler, f=f, extensions=extensions)
################################################################
# handle images
################################################################
imagespecs = {
"l8": ("numpy", "uint8", "l"),
"rgb8": ("numpy", "uint8", "rgb"),
"rgba8": ("numpy", "uint8", "rgba"),
"l": ("numpy", "float", "l"),
"rgb": ("numpy", "float", "rgb"),
"rgba": ("numpy", "float", "rgba"),
"torchl8": ("torch", "uint8", "l"),
"torchrgb8": ("torch", "uint8", "rgb"),
"torchrgba8": ("torch", "uint8", "rgba"),
"torchl": ("torch", "float", "l"),
"torchrgb": ("torch", "float", "rgb"),
"torch": ("torch", "float", "rgb"),
"torchrgba": ("torch", "float", "rgba"),
"pill": ("pil", None, "l"),
"pil": ("pil", None, "rgb"),
"pilrgb": ("pil", None, "rgb"),
"pilrgba": ("pil", None, "rgba"),
}
class ImageHandler:
"""Decode image data using the given `imagespec`.
The `imagespec` specifies whether the image is decoded
to numpy/torch/pi, decoded to uint8/float, and decoded
to l/rgb/rgba:
- l8: numpy uint8 l
- rgb8: numpy uint8 rgb
- rgba8: numpy uint8 rgba
- l: numpy float l
- rgb: numpy float rgb
- rgba: numpy float rgba
- torchl8: torch uint8 l
- torchrgb8: torch uint8 rgb
- torchrgba8: torch uint8 rgba
- torchl: torch float l
- torchrgb: torch float rgb
- torch: torch float rgb
- torchrgba: torch float rgba
- pill: pil None l
- pil: pil None rgb
- pilrgb: pil None rgb
- pilrgba: pil None rgba
"""
def __init__(self, imagespec, extensions=IMAGE_EXTENSIONS):
"""Create an image handler.
:param imagespec: short string indicating the type of decoding
:param extensions: list of extensions the image handler is invoked for
"""
if imagespec not in list(imagespecs.keys()):
raise ValueError("Unknown imagespec: %s" % imagespec)
self.imagespec = imagespec.lower()
self.extensions = extensions
def __call__(self, key, data):
"""Perform image decoding.
:param key: file name extension
:param data: binary data
"""
import PIL.Image
extension = re.sub(r".*[.]", "", key)
if extension.lower() not in self.extensions:
return None
imagespec = self.imagespec
atype, etype, mode = imagespecs[imagespec]
with io.BytesIO(data) as stream:
img = PIL.Image.open(stream)
img.load()
img = img.convert(mode.upper())
if atype == "pil":
if mode == "l":
img = img.convert("L")
return img
elif mode == "rgb":
img = img.convert("RGB")
return img
elif mode == "rgba":
img = img.convert("RGBA")
return img
else:
raise ValueError("Unknown mode: %s" % mode)
result = np.asarray(img)
if etype == "float":
result = result.astype(np.float32) / 255.0
assert result.ndim in [2, 3], result.shape
assert mode in ["l", "rgb", "rgba"], mode
if mode == "l":
if result.ndim == 3:
result = np.mean(result[:, :, :3], axis=2)
elif mode == "rgb":
if result.ndim == 2:
result = np.repeat(result[:, :, np.newaxis], 3, axis=2)
elif result.shape[2] == 4:
result = result[:, :, :3]
elif mode == "rgba":
if result.ndim == 2:
result = np.repeat(result[:, :, np.newaxis], 4, axis=2)
result[:, :, 3] = 255
elif result.shape[2] == 3:
result = np.concatenate(
[result, 255 * np.ones(result.shape[:2])], axis=2
)
assert atype in ["numpy", "torch"], atype
if atype == "numpy":
return result
elif atype == "torch":
import torch
if result.ndim == 3:
return torch.from_numpy(result.transpose(2, 0, 1).copy())
else:
return torch.from_numpy(result.copy())
return None
def imagehandler(imagespec, extensions=IMAGE_EXTENSIONS):
"""Create an image handler.
This is just a lower case alias for ImageHander.
:param imagespec: textual image spec
:param extensions: list of extensions the handler should be applied for
"""
return ImageHandler(imagespec, extensions)
################################################################
# torch video
################################################################
def torch_video(key, data):
"""Decode video using the torchvideo library.
:param key: file name extension
:param data: data to be decoded
"""
extension = re.sub(r".*[.]", "", key)
if extension not in "mp4 ogv mjpeg avi mov h264 mpg webm wmv".split():
return None
import torchvision.io
with tempfile.TemporaryDirectory() as dirname:
fname = os.path.join(dirname, f"file.{extension}")
with open(fname, "wb") as stream:
stream.write(data)
return torchvision.io.read_video(fname, pts_unit="sec")
################################################################
# torchaudio
################################################################
def torch_audio(key, data):
"""Decode audio using the torchaudio library.
:param key: file name extension
:param data: data to be decoded
"""
extension = re.sub(r".*[.]", "", key)
if extension not in ["flac", "mp3", "sox", "wav", "m4a", "ogg", "wma"]:
return None
import torchaudio
with tempfile.TemporaryDirectory() as dirname:
fname = os.path.join(dirname, f"file.{extension}")
with open(fname, "wb") as stream:
stream.write(data)
return torchaudio.load(fname)
################################################################
# special class for continuing decoding
################################################################
class Continue:
"""Special class for continuing decoding.
This is mostly used for decompression, as in:
def decompressor(key, data):
if key.endswith(".gz"):
return Continue(key[:-3], decompress(data))
return None
"""
def __init__(self, key, data):
"""__init__.
:param key:
:param data:
"""
self.key, self.data = key, data
def gzfilter(key, data):
"""Decode .gz files.
This decodes compressed files and the continues decoding.
:param key: file name extension
:param data: binary data
"""
import gzip
if not key.endswith(".gz"):
return None
decompressed = gzip.open(io.BytesIO(data)).read()
return Continue(key[:-3], decompressed)
################################################################
# decode entire training amples
################################################################
default_pre_handlers = [gzfilter]
default_post_handlers = [basichandlers]
class Decoder:
"""Decode samples using a list of handlers.
For each key/data item, this iterates through the list of
handlers until some handler returns something other than None.
"""
def __init__(self, handlers, pre=None, post=None, only=None, partial=False):
"""Create a Decoder.
:param handlers: main list of handlers
:param pre: handlers called before the main list (.gz handler by default)
:param post: handlers called after the main list (default handlers by default)
:param only: a list of extensions; when give, only ignores files with those extensions
:param partial: allow partial decoding (i.e., don't decode fields that aren't of type bytes)
"""
assert isinstance(handlers, list), f"handlers = {handlers} must be a list"
if isinstance(only, str):
only = only.split()
self.only = only if only is None else set(only)
if pre is None:
pre = default_pre_handlers
if post is None:
post = default_post_handlers
assert all(callable(h) for h in handlers), f"one of {handlers} not callable"
assert all(callable(h) for h in pre), f"one of {pre} not callable"
assert all(callable(h) for h in post), f"one of {post} not callable"
self.handlers = pre + handlers + post
self.partial = partial
def decode1(self, key, data):
"""Decode a single field of a sample.
:param key: file name extension
:param data: binary data
"""
key = "." + key
for f in self.handlers:
result = f(key, data)
if isinstance(result, Continue):
key, data = result.key, result.data
continue
if result is not None:
return result
return data
def decode(self, sample):
"""Decode an entire sample.
:param sample: the sample, a dictionary of key value pairs
"""
result = {}
assert isinstance(sample, dict), sample
for k, v in list(sample.items()):
if k[:2] == "__":
if isinstance(v, bytes):
try:
v = v.decode("utf-8")
except Exception:
print(f"Can't decode v of k = {k} as utf-8: v = {v}")
result[k] = v
continue
if self.only is not None and k not in self.only:
result[k] = v
continue
assert v is not None
if self.partial:
if isinstance(v, bytes):
result[k] = self.decode1(k, v)
else:
result[k] = v
else:
assert isinstance(v, bytes), f"k,v = {k}, {v}"
result[k] = self.decode1(k, v)
return result
def __call__(self, sample):
"""Decode an entire sample.
:param sample: the sample
"""
assert isinstance(sample, dict), (len(sample), sample)
return self.decode(sample)