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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594
|
#!/usr/bin/env python
# -*- coding: utf-8 -*
"""
MSNumpress decoder class
Authors:
Manuel Kösters
Christian Fufezan
"""
import sys
import struct
import math
import numpy as np
class MSNumpress:
"""
The library provides implementations of 4 different algorithms,
1 designed to compress first order smooth data like retention time or M/Z
arrays, 1 designed to transform first order smooth data for more efficient
zlib compression, and 2 for compressing non-smooth data with lower
requirements on precision like ion count arrays.
..Note::
This is a Python implementation of the Golomb-Rice encoding,
also known as MS-Numpress. As such it is considerably slower than
the C implementation, which we wrapped into cython. Please make
yourself a favor and use the the wrapper/c implementation and only
as a last result this Python implementation.
"""
def __init__(self, data=None):
"""
Initialize stateful Numpress De- and Encoding.
Args:
array (list): data array with compressed or decompressed data
"""
self.is_little_endian = True if sys.byteorder == "little" else False
self.Filler = {
8: "00000000",
7: "0000000",
6: "000000",
5: "00000",
4: "0000",
3: "000",
2: "00",
1: "0",
0: "",
9: "f",
10: "ff",
11: "fff",
12: "ffff",
13: "fffff",
14: "ffffff",
15: "fffffff",
}
# call check what data was set ...
self._encoded_data = None
self._decoded_data = None
if type(data) == bytearray:
# set data is encoded
self.data_state = "encoded"
self._encoded_data = data
elif type(data) == list:
# set data is decoded
self.data_state = "decoded"
self._decoded_data = data
else:
pass
@property
def decoded_data(self):
if self._decoded_data is None:
raise Exception("decoded data is not set")
return self._decoded_data
@decoded_data.setter
def decoded_data(self, data):
if type(data) is not list and type(data) is not np.ndarray:
raise Exception("data must be list")
self._decoded_data = data
@property
def encoded_data(self):
if self._encoded_data is None:
raise Exception("encoded data is not set")
return self._encoded_data
@encoded_data.setter
def encoded_data(self, data):
"""Set the data."""
if type(data) is not bytearray:
raise Exception("data must be bytearray")
self._encoded_data = data
def _linear_fixed_point(self):
"""
Calculate the optimal predictor coefficient for linear prediction model
for the Numpress Linear algorithm.
Returns:
fixed_point(int): scaling factor for linear prediction
"""
data = self.decoded_data
data_size = len(data)
if data_size == 0:
return 0
if data_size == 1:
return math.floor(0xFFFFFFFF) / data[0]
max_double = max(data[:2])
for i in range(2, data_size):
extrapol = data[i - 1] + (data[i - 1] - data[i - 2])
diff = data[i] - extrapol
max_double = max([max_double, math.ceil(abs(diff) + 1)])
return math.floor(0x7FFFFFFF / max_double)
def _slof_fixed_point(self):
"""
Calculate the optimal predictor coefficient for linear prediction model
for the Numpress Slof algorithm.
Returns:
fixed_point(int): scaling factor for linear prediction
"""
data = self.decoded_data
if len(data) == 0:
return 0
max_double = 1
for i in range(len(data)):
x = math.log(data[i] + 1)
max_double = max([max_double, x])
fp = math.floor(0xFFFF / max_double)
return fp
def _encodeInt(self, integer):
"""
# RENAME FUNCTION :)
Encode a given 4 byte integer by truncating leading 1s and 0s
and saving them in a halfbyte along with the original data
Args:
integer (int): Value to be compressed
Returns:
bytes (bytearray): encoded bytes
"""
mask = 0xF0000000
try:
integer_as_bytes = struct.pack(">i", integer)
except:
integer_as_bytes = None
init = integer & 0xF0000000
results = bytearray()
if integer_as_bytes == b"\x00\x00\x00\x00":
results.append(0x8)
x = 8
elif init == 0:
x = 0
for b in integer_as_bytes:
first_half_byte = (0xFF & b) >> 4
second_half_byte = 0xF & b
if first_half_byte == 0:
x += 1
if second_half_byte == 0:
x += 1
continue
else:
results.append(x)
break
else:
results.append(x)
break
else:
for x in range(8):
m = mask >> 4 * x
if integer & m == m:
continue
else:
results.append(x + 8)
break
if len(results) == 0:
results.append(x + 8)
if x < 8:
for m in range(8 - x):
r_to_l_mask = 0xF
np_byte = (integer >> 4 * m) & r_to_l_mask
results.append(np_byte)
return results
def _decodeInt(self, encodedInt):
"""
Decode an integer from bytes encoded by :py:func:`_encodeInt`
Args:
encodedInt (bytearray): encoded bytes
Returns:
result (int): decoded form of the encoded integer
"""
hex_string = ""
fill = "0"
for pos, byte in enumerate(encodedInt):
if pos == 0:
leading = byte & 0xF
if leading > 8:
leading -= 8
fill = "f"
else:
hex_string = str(hex(byte))[2:] + hex_string
hex_string = fill * leading + hex_string
x = int(hex_string, 16)
if x > 0x7FFFFFFF:
x -= 0x100000000
return x
def encode_linear(self):
"""
Use Linear prediction and Golomb coding to compress an array
of mz values.
Returns:
result (bytearray): Golomb encoded bytearray
EXPAMPLE
IN:
[ 12.23123, 14.21431, 23.22222 ]
OUT:
(b'\x50\x0c\x0f\x08\x51\x02\x03\x01\x23\x00') <class 'bytearray'>
"""
ints = [0 for x in range(3)]
data = self.decoded_data
fp = self._linear_fixed_point()
encoded_fixed_point = self._encode_fixed_point(fp)
result = bytearray(encoded_fixed_point)
if len(data) == 0:
return 8
ints[1] = int(round(data[0] * fp))
for i in range(4):
result.append((ints[1] >> (i * 8)) & 0xFF)
if len(data) == 1:
return 12
ints[2] = int(round(data[1] * fp))
for i in range(4):
result.append((ints[2] >> (i * 8)) & 0xFF)
enc_diff = bytearray()
for i in range(2, len(data)):
ints[0] = ints[1]
ints[1] = ints[2]
ints[2] = int(round(data[i] * fp))
extrapol = ints[1] + (ints[1] - ints[0])
diff = ints[2] - extrapol
enc_diff += self._encodeInt(diff)
for i in range(1, len(enc_diff), 2):
final_byte = enc_diff[i] & 0xF | enc_diff[i - 1] << 4
result.append(final_byte)
if len(enc_diff) % 2 != 0:
enc_diff = bytearray([enc_diff[-1]])
else:
enc_diff = bytearray()
self.encoded_data = result
return result
def decode_linear(self):
"""
Decode a Golomb encoded bytearray to its corresponding numpy array.
Returns:
result (np.ndarray): NumLin decoded numpy array of the
original data
"""
ints = [0 for x in range(3)]
data = self.encoded_data
if len(data) < 8:
raise Exception("Corrupt input data.\nnot enough bytes to read fixed point")
enc_fp = data
fp = self._decode_fixed_point(enc_fp)
if len(data) < 12:
raise Exception("Corrupt input data.\nnot enough bytes to read first value")
for i in range(4):
ints[1] = ints[1] | ((0xFF & (data[8 + i])) << (i * 8))
i1 = ints[1] / fp
if len(data) < 16:
raise Exception("Corrupt input data\nnot enough bytes to read second value")
for i in range(4):
ints[2] = ints[2] | ((0xFF & (data[12 + i])) << (i * 8))
i2 = ints[2] / fp
diff_vals = self._decode_ints_from_bytearray(16, fp)
result = [0 for x in range(len(diff_vals) + 2)]
result[0] = i1
result[1] = i2
i = 2
for diff in diff_vals:
ints[0] = ints[1]
ints[1] = ints[2]
ints[2] = diff
extrapol = ints[1] + (ints[1] - ints[0])
val = extrapol + ints[2]
result[i] = val / fp
ints[2] = val
i += 1
return result
def _decode_ints_from_bytearray(self, pos, normalization=1):
"""
Decode truncated integer encoded values and end position
in bytearray.
Args:
pos (int) : start position in array
Returns:
results (list): ...
"""
# results = np.empty(len(self.encoded_data) * 2)
results = [0 for x in range(len(self.encoded_data) * 2)]
leading_count = None
i = 0
while pos < len(self.encoded_data):
current_byte = "{0:02x}".format(self.encoded_data[pos])
for byte_pos in [0, 1]:
hb = current_byte[byte_pos]
if leading_count is None:
if hb == "8":
results[i] = 0
i += 1
else:
leading_count = int(hb, 16)
hex_str = ""
cb = leading_count - (8 * math.floor(leading_count / 8))
else:
hex_str = "{0}{1}".format(hb, hex_str)
cb += 1
try:
1 / (cb - 8)
except:
final_int = int(self.Filler[leading_count] + hex_str, 16)
if final_int > 0x7FFFFFFF:
final_int -= 0x100000000
results[i] = final_int
i += 1
leading_count = None
pos += 1
return results[:i]
def encode_pic(self):
"""
Encode Ion count data by rounding to the next integer and
store in a truncated form.
Returns:
result (bytearray): array with rounded
ion count data in bytes
"""
# need to be optimized
res = bytearray()
final = bytearray()
for val in self.decoded_data:
val = int(round(val)) # simply round
enc_int = self._encodeInt(val)
res.extend(enc_int)
# assure res has an even length
if len(res) % 2 != 0:
res.append(0)
# pull halfbytes together
for i in range(1, len(res), 2):
val = (res[i - 1] << 4) | res[i] & 0xF
final.append(val)
self.encoded_data = final
return final
def decode_pic(self):
"""
Decode ion count data compressed with :py:func:`encodePic`
Returns:
result (array): decoded Ion count data as numpy array
"""
results = []
data = self.encoded_data
read_first = True
hb_to_read = 0
current_value = bytearray()
# TODO fix problem when 2 or more zeroes appear after each other
# => if first hbc is 8, the next number is a hbc again and not data
for val in data:
if hb_to_read == 0:
if read_first is True:
current_value = bytearray()
count = (0xFF & val) >> 4
current_value.append(count)
hb_to_read = 8 - count
read_first = False
else:
current_value = bytearray()
count = 0xF & val
current_value.append(count)
hb_to_read = 8 - count
read_first = True
if hb_to_read != 0:
if hb_to_read == 1:
if read_first is True:
hb = (0xFF & val) >> 4
current_value.append(hb)
hb_to_read -= 1
dec_int = self._decodeInt(current_value)
current_value = bytearray()
results.append(dec_int)
count = 0xF & val
current_value.append(count)
hb_to_read = 8 - count
continue
else:
hb = 0xF & val
current_value.append(hb)
hb_to_read -= 1
dec_int = self._decodeInt(current_value)
current_value = bytearray()
results.append(dec_int)
read_first = True
continue
if read_first is True:
hb1 = (0xFF & val) >> 4
hb2 = 0xF & val
current_value.append(hb1)
current_value.append(hb2)
hb_to_read -= 2
if hb_to_read == 0:
dec_int = self._decodeInt(current_value)
results.append(dec_int)
current_value = bytearray()
continue
else:
hb = 0xF & val
current_value.append(hb)
read_first = True
hb_to_read -= 1
self.decoded_data = results
return results
def encode_slof(self):
"""
encode ion count data by multiplying the logarithm of each value with
a scaling factor, rounding to nearest integer, and saving the 2 LSB.
Returns:
result (bytearray): short logged float compressed bytearray
"""
data = self.decoded_data
self.fixed_point = self._slof_fixed_point()
res = self._encode_fixed_point(self.fixed_point)
for i in range(len(data)):
x = math.floor((math.log(data[i] + 1)) * self.fixed_point + 0.5)
res.append(0xFF & x)
res.append(x >> 8)
self.encoded_data = res
return res
def decode_slof(self):
"""
Decode short logged float compressed data.
Returns:
result (bytearray): array with decoded ion count data
"""
data = self.encoded_data
if len(data) < 8:
return -1
res = list()
fixed_point = self._decode_fixed_point(data)
for i in range(8, len(data), 2):
x = 0xFF & data[i] | ((0xFF & data[i + 1]) << 8)
res.append(math.exp((0xFFFF & x) / fixed_point) - 1)
self.decoded_data = res
return res
def encode_safe(self):
"""
Transform values by linear prediction without scaling or truncated
representation.
Returns:
result (bytearray): transformed data in bytes
"""
# result = bytearray()
# assert data format is correct
# save_first_2_values_unencoded()
# for val in self.data[2:]:
# pred = do_linear_prediction()
# enc = do_Golomb_encoding()
# result.extend(enc)
# return result
pass
def decode_safe(self):
# if not self.data_state == 'decoded':
# raise Exception('No encoded data found\nPlease set encoded Data')
# result = np.ndarray()
# int1 = read_int1(8)
# int2 = read_int2(8)
# add ints to result
# for x in range(0, len(self.data, 2)):
# enc_int = self.data[x:x+8]
# result.append(dec_int)
pass
def _encode_fixed_point(self, value):
"""
Save a given float as bytes.
Args:
value(float): fixed point
Returns:
res (bytearray): encoded fixed point
"""
res = bytearray(8)
fp = value
fp = struct.pack("<d", fp)
fp = struct.unpack("<q", fp)[0]
if self.is_little_endian:
for i in range(8):
res[7 - i] = (fp >> (8 * i)) & 0xFF
else:
for i in range(7, -1, -1):
res[7 - i] = (fp >> (8 * i)) & 0xFF
return res
def _decode_fixed_point(self, value):
"""
Decode a given bytearray to float.
Args:
value(bytearray): encoded fixed point
Returns:
res (float): encoded fixed point
"""
fp = 0
if self.is_little_endian:
for i in range(8):
fp = fp | ((0x00FF & value[7 - i]) << (8 * i))
else:
for i in range(7, -1, -1):
fp = fp | ((0x00FF & value[7 - i]) << (8 * i))
fp = struct.pack("<q", fp)
fp = struct.unpack("<d", fp)[0]
return fp
if __name__ == "__main__":
print(__doc__)
|