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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954
|
# -*- coding: utf-8 -*-
"""
ninjotiff.py
Created on Mon Apr 15 13:41:55 2013
A big amount of the tiff writer are (PFE) from
https://github.com/davidh-ssec/polar2grid by David Hoese
License:
Copyright (C) 2013 Space Science and Engineering Center (SSEC),
University of Wisconsin-Madison.
Lars Ørum Rasmussen, DMI.
This program is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program. If not, see <http://www.gnu.org/licenses/>.
Original scripts and automation included as part of this package are
distributed under the GNU GENERAL PUBLIC LICENSE agreement version 3.
Binary executable files included as part of this software package are
copyrighted and licensed by their respective organizations, and
distributed consistent with their licensing terms.
"""
import os
import copy
import logging
import calendar
from datetime import datetime
import numpy as np
import mpop.imageo.formats.libtiff as libtiff
from mpop.imageo.formats.libtiff import TIFF, TIFFFieldInfo, TIFFDataType, FIELD_CUSTOM
log = logging.getLogger(__name__)
#-------------------------------------------------------------------------------
#
# Ninjo tiff tags from DWD
#
#-------------------------------------------------------------------------------
# Geotiff tags
GTF_ModelPixelScale = 33550
GTF_ModelTiepoint = 33922
NTD_Magic = 40000
NTD_SatelliteNameID = 40001
NTD_DateID = 40002
NTD_CreationDateID = 40003
NTD_ChannelID = 40004
NTD_HeaderVersion = 40005
NTD_FileName = 40006
NTD_DataType = 40007
NTD_SatelliteNumber = 40008
NTD_ColorDepth = 40009
NTD_DataSource = 40010
NTD_XMinimum = 40011
NTD_XMaximum = 40012
NTD_YMinimum = 40013
NTD_YMaximum = 40014
NTD_Projection = 40015
NTD_MeridianWest = 40016
NTD_MeridianEast = 40017
NTD_EarthRadiusLarge = 40018
NTD_EarthRadiusSmall = 40019
NTD_GeodeticDate = 40020
NTD_ReferenceLatitude1 = 40021
NTD_ReferenceLatitude2 = 40022
NTD_CentralMeridian = 40023
NTD_PhysicValue = 40024
NTD_PhysicUnit = 40025
NTD_MinGrayValue = 40026
NTD_MaxGrayValue = 40027
NTD_Gradient = 40028
NTD_AxisIntercept = 40029
NTD_ColorTable = 40030
NTD_Description = 40031
NTD_OverflightDirection = 40032
NTD_GeoLatitude = 40033
NTD_GeoLongitude = 40034
NTD_Altitude = 40035
NTD_AOSAsimuth = 40036
NTD_LOSAsimuth = 40037
NTD_MaxElevation = 40038
NTD_OverflightTime = 40039
NTD_IsBlackLineCorrection = 40040
NTD_IsAtmosphereCorrected = 40041
NTD_IsCalibrated = 40042
NTD_IsNormalized = 40043
NTD_OriginalHeader = 40044
NTD_IsValueTableAvailable = 40045
NTD_ValueTableStringField = 40046
NTD_ValueTableFloatField = 40047
NTD_TransparentPixel = 50000
#
# model_pixel_scale_tag_count ? ...
# Sometimes DWD product defines an array of length 2 (instead of 3 (as in geotiff)).
#
MODEL_PIXEL_SCALE_COUNT = int(os.environ.get("GEOTIFF_MODEL_PIXEL_SCALE_COUNT", 3))
ninjo_tags_dict = {
# Geotiff tags
GTF_ModelPixelScale:
TIFFFieldInfo(GTF_ModelPixelScale, MODEL_PIXEL_SCALE_COUNT,
MODEL_PIXEL_SCALE_COUNT, TIFFDataType.TIFF_DOUBLE,
FIELD_CUSTOM, True, False, "ModelPixelScale" ),
GTF_ModelTiepoint:
TIFFFieldInfo(GTF_ModelTiepoint, 6, 6, TIFFDataType.TIFF_DOUBLE,
FIELD_CUSTOM, True, False, "ModelTiePoint" ),
# DWD tags
NTD_Magic:
TIFFFieldInfo(NTD_Magic, -1, -1, TIFFDataType.TIFF_ASCII,
FIELD_CUSTOM, True, False, "Magic" ),
NTD_SatelliteNameID:
TIFFFieldInfo(NTD_SatelliteNameID, 1, 1, TIFFDataType.TIFF_LONG,
FIELD_CUSTOM, True, False, "SatelliteNameID" ),
NTD_DateID:
TIFFFieldInfo(NTD_DateID, 1, 1, TIFFDataType.TIFF_LONG,
FIELD_CUSTOM, True, False, "DateID" ),
NTD_CreationDateID:
TIFFFieldInfo(NTD_CreationDateID, 1, 1, TIFFDataType.TIFF_LONG,
FIELD_CUSTOM, True, False, "CreationDateID" ),
NTD_ChannelID:
TIFFFieldInfo(NTD_ChannelID, 1, 1, TIFFDataType.TIFF_LONG,
FIELD_CUSTOM, True, False, "ChannelID" ),
NTD_HeaderVersion:
TIFFFieldInfo(NTD_HeaderVersion, 1, 1, TIFFDataType.TIFF_SLONG,
FIELD_CUSTOM, True, False, "HeaderVersion" ),
NTD_FileName:
TIFFFieldInfo(NTD_FileName, -1, -1, TIFFDataType.TIFF_ASCII,
FIELD_CUSTOM, True, False, "FileName" ),
NTD_DataType:
TIFFFieldInfo(NTD_DataType, 5, 5, TIFFDataType.TIFF_ASCII,
FIELD_CUSTOM, True, False, "DataType" ), # 4 chars + NUL character
NTD_SatelliteNumber:
TIFFFieldInfo(NTD_SatelliteNumber, -1, -1, TIFFDataType.TIFF_ASCII,
FIELD_CUSTOM, True, False, "SatelliteNumber" ),
NTD_ColorDepth:
TIFFFieldInfo(NTD_ColorDepth, 1, 1, TIFFDataType.TIFF_SLONG,
FIELD_CUSTOM, True, False, "ColorDepth" ),
NTD_DataSource:
TIFFFieldInfo(NTD_DataSource, -1, -1, TIFFDataType.TIFF_ASCII,
FIELD_CUSTOM, True, False, "DataSource" ),
NTD_XMinimum:
TIFFFieldInfo(NTD_XMinimum, 1, 1, TIFFDataType.TIFF_SLONG,
FIELD_CUSTOM, True, False, "XMinimum" ),
NTD_XMaximum:
TIFFFieldInfo(NTD_XMaximum, 1, 1, TIFFDataType.TIFF_SLONG,
FIELD_CUSTOM, True, False, "XMaximum" ),
NTD_YMinimum:
TIFFFieldInfo(NTD_YMinimum, 1, 1, TIFFDataType.TIFF_SLONG,
FIELD_CUSTOM, True, False, "YMinimum" ),
NTD_YMaximum:
TIFFFieldInfo(NTD_YMaximum, 1, 1, TIFFDataType.TIFF_SLONG,
FIELD_CUSTOM, True, False, "YMaximum" ),
NTD_Projection:
TIFFFieldInfo(NTD_Projection, 5, 5, TIFFDataType.TIFF_ASCII,
FIELD_CUSTOM, True, False, "Projection" ), # 4 chars + NUL character
NTD_MeridianWest:
TIFFFieldInfo(NTD_MeridianWest, 1, 1, TIFFDataType.TIFF_FLOAT,
FIELD_CUSTOM, True, False, "MeridianWest" ),
NTD_MeridianEast:
TIFFFieldInfo(NTD_MeridianEast, 1, 1, TIFFDataType.TIFF_FLOAT,
FIELD_CUSTOM, True, False, "MeridianEast" ),
NTD_EarthRadiusLarge:
TIFFFieldInfo(NTD_EarthRadiusLarge, 1, 1, TIFFDataType.TIFF_FLOAT,
FIELD_CUSTOM, True, False, "EarthRadiusLarge" ),
NTD_EarthRadiusSmall:
TIFFFieldInfo(NTD_EarthRadiusSmall, 1, 1, TIFFDataType.TIFF_FLOAT,
FIELD_CUSTOM, True, False, "EarthRadiusSmall" ),
NTD_GeodeticDate:
TIFFFieldInfo(NTD_GeodeticDate, -1, -1, TIFFDataType.TIFF_ASCII,
FIELD_CUSTOM, True, False, "GeodeticDate" ), # Max 20
NTD_ReferenceLatitude1:
TIFFFieldInfo(NTD_ReferenceLatitude1, 1, 1, TIFFDataType.TIFF_FLOAT,
FIELD_CUSTOM, True, False, "ReferenceLatitude1" ),
NTD_ReferenceLatitude2:
TIFFFieldInfo(NTD_ReferenceLatitude2, 1, 1, TIFFDataType.TIFF_FLOAT,
FIELD_CUSTOM, True, False, "ReferenceLatitude2" ),
NTD_CentralMeridian:
TIFFFieldInfo(NTD_CentralMeridian, 1, 1, TIFFDataType.TIFF_FLOAT,
FIELD_CUSTOM, True, False, "CentralMeridian" ),
NTD_PhysicValue:
TIFFFieldInfo(NTD_PhysicValue, -1, -1, TIFFDataType.TIFF_ASCII,
FIELD_CUSTOM, True, False, "PhysicValue" ), # Max 10
NTD_PhysicUnit:
TIFFFieldInfo(NTD_PhysicUnit, -1, -1, TIFFDataType.TIFF_ASCII,
FIELD_CUSTOM, True, False, "PhysicUnit" ), # Max 10
NTD_MinGrayValue:
TIFFFieldInfo(NTD_MinGrayValue, 1, 1, TIFFDataType.TIFF_SLONG,
FIELD_CUSTOM, True, False, "MinGrayValue" ),
NTD_MaxGrayValue:
TIFFFieldInfo(NTD_MaxGrayValue, 1, 1, TIFFDataType.TIFF_SLONG,
FIELD_CUSTOM, True, False, "MaxGrayValue" ),
NTD_Gradient:
TIFFFieldInfo(NTD_Gradient, 1, 1, TIFFDataType.TIFF_FLOAT,
FIELD_CUSTOM, True, False, "Gradient" ),
NTD_AxisIntercept:
TIFFFieldInfo(NTD_AxisIntercept, 1, 1, TIFFDataType.TIFF_FLOAT,
FIELD_CUSTOM, True, False, "AxisIntercept" ),
NTD_ColorTable:
TIFFFieldInfo(NTD_ColorTable, -1, -1, TIFFDataType.TIFF_ASCII,
FIELD_CUSTOM, True, False, "ColorTable" ),
NTD_Description:
TIFFFieldInfo(NTD_Description, -1, -1, TIFFDataType.TIFF_ASCII,
FIELD_CUSTOM, True, False, "Description" ),
NTD_OverflightDirection:
TIFFFieldInfo(NTD_OverflightDirection, -1, -1, TIFFDataType.TIFF_ASCII,
FIELD_CUSTOM, True, False, "OverflightDirection" ),
NTD_GeoLatitude:
TIFFFieldInfo(NTD_GeoLatitude, 1, 1, TIFFDataType.TIFF_FLOAT,
FIELD_CUSTOM, True, False, "GeoLatitude" ),
NTD_GeoLongitude:
TIFFFieldInfo(NTD_GeoLongitude, 1, 1, TIFFDataType.TIFF_FLOAT,
FIELD_CUSTOM, True, False, "GeoLongitude" ),
NTD_Altitude:
TIFFFieldInfo(NTD_Altitude, 1, 1, TIFFDataType.TIFF_FLOAT,
FIELD_CUSTOM, True, False, "Altitude" ),
NTD_AOSAsimuth:
TIFFFieldInfo(NTD_AOSAsimuth, 1, 1, TIFFDataType.TIFF_FLOAT,
FIELD_CUSTOM, True, False, "AOSAsimuth" ),
NTD_LOSAsimuth:
TIFFFieldInfo(NTD_LOSAsimuth, 1, 1, TIFFDataType.TIFF_FLOAT,
FIELD_CUSTOM, True, False, "LOSAsimuth" ),
NTD_MaxElevation:
TIFFFieldInfo(NTD_MaxElevation, 1, 1, TIFFDataType.TIFF_FLOAT,
FIELD_CUSTOM, True, False, "MaxElevation" ),
NTD_OverflightTime:
TIFFFieldInfo(NTD_OverflightTime, 1, 1, TIFFDataType.TIFF_FLOAT,
FIELD_CUSTOM, True, False, "OverflightTime" ),
NTD_IsBlackLineCorrection:
TIFFFieldInfo(NTD_IsBlackLineCorrection, 1, 1, TIFFDataType.TIFF_SLONG,
FIELD_CUSTOM, True, False, "IsBlackLineCorrection" ),
NTD_IsAtmosphereCorrected:
TIFFFieldInfo(NTD_IsAtmosphereCorrected, 1, 1, TIFFDataType.TIFF_SLONG,
FIELD_CUSTOM, True, False, "IsAtmosphereCorrected" ),
NTD_IsCalibrated:
TIFFFieldInfo(NTD_IsCalibrated, 1, 1, TIFFDataType.TIFF_SLONG,
FIELD_CUSTOM, True, False, "IsCalibrated" ),
NTD_IsNormalized:
TIFFFieldInfo(NTD_IsNormalized, 1, 1, TIFFDataType.TIFF_SLONG,
FIELD_CUSTOM, True, False, "IsNormalized" ),
NTD_OriginalHeader:
TIFFFieldInfo(NTD_OriginalHeader, -1, -1, TIFFDataType.TIFF_ASCII,
FIELD_CUSTOM, True, False, "OriginalHeader" ),
NTD_IsValueTableAvailable:
TIFFFieldInfo(NTD_IsValueTableAvailable, 1, 1, TIFFDataType.TIFF_SLONG,
FIELD_CUSTOM, True, False, "IsValueTableAvailable" ),
NTD_ValueTableStringField:
TIFFFieldInfo(NTD_ValueTableStringField, -1, -1, TIFFDataType.TIFF_ASCII,
FIELD_CUSTOM, True, False, "ValueTableStringField" ),
NTD_ValueTableFloatField:
TIFFFieldInfo(NTD_ValueTableFloatField, 1, 1, TIFFDataType.TIFF_FLOAT,
FIELD_CUSTOM, True, False, "ValueTableFloatField" ),
NTD_TransparentPixel:
TIFFFieldInfo(NTD_TransparentPixel, 1, 1, TIFFDataType.TIFF_SLONG,
FIELD_CUSTOM, True, False, "TransparentPixel" ),
}
# Add Ninjo tags to the libtiff library
_ninjo_tags_extender = libtiff.add_tags(ninjo_tags_dict.values())
ninjo_tags = sorted(ninjo_tags_dict.keys())
#-------------------------------------------------------------------------------
#
# Read Ninjo products config file.
#
#-------------------------------------------------------------------------------
def get_product_config(product_name, force_read=False):
"""Read Ninjo configuration entry for a given product name.
:Parameters:
product_name : str
Name of Ninjo product.
:Arguments:
force_read : Boolean
Force re-reading config file.
**Notes**:
* It will look for a *ninjotiff_products.cfg* in MPOP's
configuration directory defined by *PPP_CONFIG_DIR*.
* As an example, see *ninjotiff_products.cfg.template* in
MPOP's *etc* directory.
"""
return ProductConfigs()(product_name, force_read)
class _Singleton(type):
def __init__(cls, name_, bases_, dict_):
super(_Singleton, cls).__init__(name_, bases_, dict_)
cls.instance = None
def __call__(cls, *args, **kwargs):
if cls.instance is None:
cls.instance = super(_Singleton, cls).__call__(*args, **kwargs)
return cls.instance
class ProductConfigs(object):
__metaclass__ = _Singleton
def __init__(self):
self.read_config()
def __call__(self, product_name, force_read=False):
if force_read:
self.read_config()
return self._products[product_name]
@property
def product_names(self):
return sorted(self._products.keys())
def read_config(self):
from ConfigParser import ConfigParser
def _eval(val):
try:
return eval(val)
except:
return str(val)
filename = self._find_a_config_file()
#print "Reading Ninjo config file: '%s'" % filename
log.info("Reading Ninjo config file: '%s'" % filename)
cfg = ConfigParser()
cfg.read(filename)
products = {}
for sec in cfg.sections():
prd = {}
for key, val in cfg.items(sec):
prd[key] = _eval(val)
products[sec] = prd
self._products = products
@staticmethod
def _find_a_config_file():
name_ = 'ninjotiff_products.cfg'
home_ = os.path.dirname(os.path.abspath(__file__))
penv_ = os.environ.get('PPP_CONFIG_DIR', '')
for fname_ in [os.path.join(x, name_) for x in (home_, penv_)]:
if os.path.isfile(fname_):
return fname_
raise ValueError("Could not find a Ninjo tiff config file")
#-------------------------------------------------------------------------------
#
# Read tiff file.
#
#-------------------------------------------------------------------------------
class _TIFF(object):
""" Just an context wrapper around an libtiff.TIFF instance.
"""
def __init__(self, filename, mode='r'):
"""Open a tiff file.
see: libtiff.TIFF.open()
"""
self.tiff = TIFF.open(filename, mode)
self.tiff.ninjo_tags_dict = ninjo_tags_dict
self.tiff.ninjo_tags = ninjo_tags
def __enter__(self):
return self.tiff
def __exit__(self, type_, value, traceback):
self.tiff.close()
def _read_directories(self):
"""Iterate over directories in a tiff file.
:Parameters:
self : libtiff.TIFF
A TIFF instance.
:Returns:
tiff_directory : Tiff object
A Tiff directory instance.
"""
yield self
while not self.LastDirectory():
self.ReadDirectory()
yield self
self.SetDirectory(0)
def info(filename):
"""Read metadata from Tiff file.
:Parameters:
filename : str
Name of Tiff file.
:Returns:
iterator : a Python generator iterator
A "list" of tiff metadata.
**Usage**::
for inf in info(filename):
print inf, '\n'
"""
with _TIFF(filename) as self:
for d in _read_directories(self):
l = []
for item in d.info().split('\n'):
k, v = item.split(':', 1)
if (k.endswith('OffSets') or
k.endswith('ByteCounts') or
k == 'FileName' or
k == 'DataType'):
continue
l.append(item)
for tag in d.ninjo_tags:
value = d.GetField(tag)
name = d.ninjo_tags_dict[tag].field_name
if value is None:
continue
l.append('%s: %s' % (name, str(value)))
yield '\n'.join(l)
def image_data(filename):
"""Read image data from Tiff file.
:Parameters:
filename : str
Name of Tiff file.
**Usage**::
for img in image_data(filename):
print img
"""
with _TIFF(filename) as self:
for d in _read_directories(self):
yield d.read_tiles()
def colortable(filename):
"""Read colortables from Tiff file.
:Parameters:
filename : str
Name of Tiff file.
**Usage**::
for clt in colortable(filename):
print clt
"""
with _TIFF(filename) as self:
return self.GetField('ColorMap')
#-------------------------------------------------------------------------------
#
# Write Ninjo Products
#
#-------------------------------------------------------------------------------
def _get_physic_value(physic_unit):
# return Ninjo's physics unit and value.
if physic_unit.upper() in ('K', 'KELVIN'):
return 'Kelvin', 'T'
elif physic_unit.upper() in ('C', 'CELSIUS'):
return 'Celsius', 'T'
elif physic_unit == '%':
return physic_unit, 'Reflectance'
elif physic_unit.upper() in ('MW M-2 SR-1 (CM-1)-1',):
return physic_unit, 'Radiance'
else:
return physic_unit, 'Unknown'
def _get_projection_name(area_def):
# return Ninjo's projection name.
proj_name = area_def.proj_dict['proj']
if proj_name in ('eqc',):
return 'PLAT'
elif proj_name in ('stere',):
lat_0 = area_def.proj_dict['lat_0']
if lat_0 < 0:
return 'SPOL'
else:
return 'NPOL'
return None
def _finalize(geo_image):
"""Finalize a mpop GeoImage for Ninjo. Specialy take care of phycical scale
and offset.
:Parameters:
geo_image : mpop.imageo.geo_image.GeoImage
See MPOP's documentation.
:Returns:
image : numpy.array
Final image.
scale : float
Scale for transform pixel value to physical value.
offset : float
Offset for transform pixel value to physical value.
fill_value : int
Value for used masked out pixels.
**Notes**:
physic_val = image*scale + offset
"""
if geo_image.mode == 'L':
# PFE: mpop.satout.cfscene
dtype = np.uint8
data = geo_image.channels[0]
fill_value = geo_image.fill_value or 0
if np.ma.count_masked(data) == data.size:
# All data is masked
data = np.ones(data.shape, dtype=dtype) * fill_value
scale = 1
offset = 0
else:
chn_max = data.max()
chn_min = data.min()
scale = ((chn_max - chn_min) /
(2**np.iinfo(dtype).bits - 1.0))
# Handle the case where all data has the same value.
scale = scale or 1
offset = chn_min
mask = data.mask
data = ((data.data - offset) / scale).astype(dtype)
data[mask] = fill_value
return data, scale, offset, fill_value
elif geo_image.mode == 'RGB':
channels, fill_value = geo_image._finalize()
fill_value = fill_value or (0, 0, 0)
data = np.dstack((channels[0].filled(fill_value[0]),
channels[1].filled(fill_value[1]),
channels[2].filled(fill_value[2])))
return data, 1.0, 0.0, fill_value[0]
else:
raise ValueError("Don't known how til handle image mode '%s'" %
str(geo_image.mode))
def save(geo_image, filename, ninjo_product_name=None, **kwargs):
"""MPOP's interface to Ninjo TIFF writer.
:Parameters:
geo_image : mpop.imageo.geo_image.GeoImage
See MPOP's documentation.
filename : str
The name of the TIFF file to be created
:Keywords:
ninjo_product_name : str
Optional index to Ninjo configuration file.
kwargs : dict
See _write
"""
data, scale, offset, fill_value = _finalize(geo_image)
area_def = geo_image.area
time_slot = geo_image.time_slot
# Some Ninjo tiff names
kwargs['image_dt'] = time_slot
kwargs['transparent_pix'] = fill_value
kwargs['gradient'] = scale
kwargs['axis_intercept'] = offset
kwargs['is_calibrated'] = True
write(data, filename, area_def, ninjo_product_name, **kwargs)
def write(image_data, output_fn, area_def, product_name=None, **kwargs):
"""Generic Ninjo TIFF writer.
If 'prodcut_name' is given, it will load corresponding Ninjo tiff metadata
from '${PPP_CONFIG_DIR}/ninjotiff.cfg'. Else, all Ninjo tiff metadata should
be passed by '**kwargs'. A mixture is allowed, where passed arguments
overwrite config file.
:Parameters:
image_data : 2D numpy array
Satellite image data to be put into the NinJo compatible tiff
output_fn : str
The name of the TIFF file to be created
area_def: pyresample.geometry.AreaDefinition
Defintion of area
product_name : str
Optional index to Ninjo configuration file.
:Keywords:
kwargs : dict
See _write
"""
upper_left = area_def.get_lonlat(0, 0)
lower_right = area_def.get_lonlat(area_def.shape[0], area_def.shape[1])
scale = abs(lower_right[0] - upper_left[0])/area_def.shape[1],\
abs(upper_left[1] - lower_right[1])/area_def.shape[0]
if len(image_data.shape) == 3:
shape = (area_def.y_size, area_def.x_size, 3)
write_rgb = True
log.info("Will generate RGB product '%s'" % product_name)
else:
shape = (area_def.y_size, area_def.x_size)
write_rgb = False
log.info("Will generate product '%s'" % product_name)
if image_data.shape != shape:
raise ValueError, "Raster shape %s does not correspond to expected shape %s" % (
str(image_data.shape), str(shape))
# Ninjo's physical units and value.
# If just a physical unit (e.g. 'C') is passed, it will then be
# translated into Ninjo's unit and value (e.q 'CELCIUS' and 'T').
physic_unit = kwargs.get('physic_unit', None)
if physic_unit and not kwargs.get('physic_value', None):
kwargs['physic_unit'], kwargs['physic_value'] = \
_get_physic_value(physic_unit)
# Ninjo's projection name.
kwargs['projection'] = kwargs.pop('projection', None) or \
_get_projection_name(area_def) or \
area_def.proj_id.split('_')[-1]
if product_name:
options = get_product_config(product_name)
else:
options = {}
options['meridian_west'] = upper_left[0]
options['meridian_east'] = lower_right[0]
options['pixel_xres'] = scale[0]
options['pixel_yres'] = scale[1]
options['origin_lon'] = upper_left[0]
options['origin_lat'] = upper_left[1]
options['min_gray_val'] = image_data.min()
options['max_gray_val'] = image_data.max()
options.update(kwargs) # Update/overwrite with passed arguments
_write(image_data, output_fn, write_rgb=write_rgb, **options)
#-------------------------------------------------------------------------------
#
# Write tiff file.
#
#-------------------------------------------------------------------------------
def _write(image_data, output_fn, write_rgb=False, **kwargs):
"""Proudly Found Elsewhere (PFE) https://github.com/davidh-ssec/polar2grid
by David Hoese.
Create a NinJo compatible TIFF file with the tags used
by the DWD's version of NinJo. Also stores the image as tiles on disk
and creates a multi-resolution/pyramid/overview set of images
(deresolution: 2,4,8,16).
:Parameters:
image_data : 2D or 3D numpy array
Satellite image data to be put into the NinJo compatible tiff
An 3D array (HxWx3) is expected for a RGB image.
filename : str
The name of the TIFF file to be created
:Keywords:
cmap : tuple/list of 3 lists of uint16's
Individual RGB arrays describing the color value for the
corresponding data value. For example, image data with a data
type of unsigned 8-bit integers have 256 possible values (0-255).
So each list in cmap will have 256 values ranging from 0 to
65535 (2**16 - 1). (default linear B&W colormap)
sat_id : int
DWD NinJo Satellite ID number
chan_id : int
DWD NinJo Satellite Channel ID number
data_source : str
String describing where the data came from (SSEC, EUMCAST)
tile_width : int
Width of tiles on disk (default 512)
tile_length : int
Length of tiles on disk (default 512)
data_cat : str
NinJo specific data category
- data_cat[0] = P (polar) or G (geostat)
- data_cat[1] = O (original) or P (product)
- data_cat[2:4] = RN or RB or RA or RN or AN (Raster, Bufr, ASCII, NIL)
Example: 'PORN' or 'GORN' or 'GPRN' or 'PPRN'
pixel_xres : float
Nadir view pixel resolution in degrees longitude
pixel_yres : float
Nadir view pixel resolution in degrees latitude
origin_lat : float
Top left corner latitude
origin_lon : float
Top left corner longitude
image_dt : datetime object
Python datetime object describing the date and time of the image
data provided in UTC
projection : str
NinJo compatible projection name (NPOL,PLAT,etc.)
meridian_west : float
Western image border (default 0.0)
meridian_east : float
Eastern image border (default 0.0)
radius_a : float
Large/equatorial radius of the earth (default <not written>)
radius_b : float
Small/polar radius of the earth (default <not written>)
ref_lat1 : float
Reference latitude 1 (default <not written>)
ref_lat2 : float
Reference latitude 2 (default <not written>)
central_meridian : float
Central Meridian (default <not written>)
physic_value : str
Physical value type. Examples:
- Temperature = 'T'
- Albedo = 'ALBEDO'
physic_unit : str
Physical value units. Examples:
- 'CELSIUS'
- '%'
min_gray_val : int
Minimum gray value (default 0)
max_gray_val : int
Maximum gray value (default 255)
gradient : float
Gradient/Slope
axis_intercept : float
Axis Intercept
altitude : float
Altitude of the data provided (default 0.0)
is_atmo_corrected : bool
Is the data atmosphere corrected? (True/1 for yes) (default False/0)
is_calibrated : bool
Is the data calibrated? (True/1 for yes) (default False/0)
is_normalized : bool
Is the data normalized (True/1 for yes) (default False/0)
description : str
Description string to be placed in the output TIFF (optional)
transparent_pix : int
Transparent pixel value (default -1)
:Raises:
KeyError :
if required keyword is not provided
"""
def _raise_value_error(text):
log.error(text)
raise ValueError(text)
def _default_colormap(reverse=False):
# Basic B&W colormap
if reverse:
return [[ x*256 for x in range(255, -1, -1) ]]*3
return [[ x*256 for x in range(256) ]]*3
def _eval_or_none(key, eval_func):
try:
return eval_func(kwargs[key])
except KeyError:
return None
log.info("Creating output file '%s'" % (output_fn,))
tiff = TIFF.open(output_fn, "w")
# Extract keyword arguments
cmap = kwargs.pop("cmap", None)
sat_id = int(kwargs.pop("sat_id"))
chan_id = int(kwargs.pop("chan_id"))
data_source = str(kwargs.pop("data_source"))
tile_width = int(kwargs.pop("tile_width", 512))
tile_length = int(kwargs.pop("tile_length", 512))
data_cat = str(kwargs.pop("data_cat"))
pixel_xres = float(kwargs.pop("pixel_xres"))
pixel_yres = float(kwargs.pop("pixel_yres"))
origin_lat = float(kwargs.pop("origin_lat"))
origin_lon = float(kwargs.pop("origin_lon"))
image_dt = kwargs.pop("image_dt")
projection = str(kwargs.pop("projection"))
meridian_west = float(kwargs.pop("meridian_west", 0.0))
meridian_east = float(kwargs.pop("meridian_east", 0.0))
radius_a = _eval_or_none("radius_a", float)
radius_b = _eval_or_none("radius_b", float)
ref_lat1 = _eval_or_none("ref_lat1", float)
ref_lat2 = _eval_or_none("ref_lat2", float)
central_meridian = _eval_or_none("central_meridian", float)
min_gray_val = int(kwargs.pop("min_gray_val", 0))
max_gray_val = int(kwargs.pop("max_gray_val", 255))
altitude = float(kwargs.pop("altitude", 0.0))
is_blac_corrected = int(bool(kwargs.pop("is_blac_corrected", 0)))
is_atmo_corrected = int(bool(kwargs.pop("is_atmo_corrected", 0)))
is_calibrated = int(bool(kwargs.pop("is_calibrated", 0)))
is_normalized = int(bool(kwargs.pop("is_normalized", 0)))
description = _eval_or_none("description", str)
physic_value = str(kwargs.pop("physic_value", 'None'))
physic_unit = str(kwargs.pop("physic_unit", 'None'))
gradient = float(kwargs.pop("gradient", 1.0))
axis_intercept = float(kwargs.pop("axis_intercept", 0.0))
transparent_pix = int(kwargs.pop("transparent_pix", -1))
# Keyword checks / verification
if not cmap:
if physic_value == 'T':
reverse = True
else:
reverse = False
cmap = _default_colormap(reverse)
if len(cmap) != 3:
_raise_value_error("Colormap (cmap) must be a list of 3 lists (RGB), not %d" %
len(cmap))
if len(data_cat) != 4:
_raise_value_error("NinJo data type must be 4 characters")
if data_cat[0] not in ["P", "G"]:
_raise_value_error("NinJo data type's first character must be 'P' or 'G' not '%s'" %
data_cat[0])
if data_cat[1] not in ["O", "P"]:
_raise_value_error("NinJo data type's second character must be 'O' or 'P' not '%s'" %
data_cat[1])
if data_cat[2:4] not in ["RN","RB","RA","BN","AN"]:
_raise_value_error("NinJo data type's last 2 characters must be one of %s not '%s'" %
("['RN','RB','RA','BN','AN']", data_cat[2:4]))
if description is not None and len(description) >= 1000:
log.error("NinJo description must be less than 1000 characters")
raise ValueError("NinJo description must be less than 1000 characters")
file_dt = datetime.utcnow()
file_epoch = calendar.timegm(file_dt.timetuple())
image_epoch = calendar.timegm(image_dt.timetuple())
def _write_oneres(image_data, pixel_xres, pixel_yres):
log.info("Writing tags and data for a resolution %dx%d" % image_data.shape[:2])
# Write Tag Data
# Built ins
tiff.SetField("ImageWidth", image_data.shape[1])
tiff.SetField("ImageLength", image_data.shape[0])
tiff.SetField("BitsPerSample", 8)
tiff.SetField("Compression", libtiff.COMPRESSION_DEFLATE)
if write_rgb:
tiff.SetField("Photometric", libtiff.PHOTOMETRIC_RGB)
tiff.SetField("SamplesPerPixel", 3)
else:
tiff.SetField("Photometric", libtiff.PHOTOMETRIC_PALETTE)
tiff.SetField("SamplesPerPixel", 1)
tiff.SetField("ColorMap", cmap)
tiff.SetField("Orientation", libtiff.ORIENTATION_TOPLEFT)
tiff.SetField("SMinSampleValue", 0)
tiff.SetField("SMaxsampleValue", 255)
tiff.SetField("PlanarConfig", libtiff.PLANARCONFIG_CONTIG)
tiff.SetField("TileWidth", tile_width)
tiff.SetField("TileLength", tile_length)
tiff.SetField("SampleFormat", libtiff.SAMPLEFORMAT_UINT)
# NinJo specific tags
if description is not None:
tiff.SetField("Description", description)
if MODEL_PIXEL_SCALE_COUNT == 3:
tiff.SetField("ModelPixelScale", [pixel_xres, pixel_yres, 0.0])
else:
tiff.SetField("ModelPixelScale", [pixel_xres, pixel_yres])
tiff.SetField("ModelTiePoint", [0.0, 0.0, 0.0, origin_lon, origin_lat, 0.0])
tiff.SetField("Magic", "NINJO")
tiff.SetField("SatelliteNameID", sat_id)
tiff.SetField("DateID", image_epoch)
tiff.SetField("CreationDateID", file_epoch)
tiff.SetField("ChannelID", chan_id)
tiff.SetField("HeaderVersion", 2)
tiff.SetField("FileName", output_fn)
tiff.SetField("DataType", data_cat)
tiff.SetField("SatelliteNumber", "\x00") # Hardcoded to 0
if write_rgb:
tiff.SetField("ColorDepth", 24)
elif cmap:
tiff.SetField("ColorDepth", 16)
else:
tiff.SetField("ColorDepth", 8)
tiff.SetField("DataSource", data_source)
tiff.SetField("XMinimum", 1)
tiff.SetField("XMaximum", image_data.shape[1])
tiff.SetField("YMinimum", 1)
tiff.SetField("YMaximum", image_data.shape[0])
tiff.SetField("Projection", projection)
tiff.SetField("MeridianWest", meridian_west)
tiff.SetField("MeridianEast", meridian_east)
if radius_a is not None:
tiff.SetField("EarthRadiusLarge", float(radius_a))
if radius_b is not None:
tiff.SetField("EarthRadiusSmall", float(radius_b))
#tiff.SetField("GeodeticDate", "\x00") # ---?
if ref_lat1 is not None:
tiff.SetField("ReferenceLatitude1", ref_lat1)
if ref_lat2 is not None:
tiff.SetField("ReferenceLatitude2", ref_lat2)
if central_meridian is not None:
tiff.SetField("CentralMeridian", central_meridian)
tiff.SetField("PhysicValue", physic_value)
tiff.SetField("PhysicUnit", physic_unit)
tiff.SetField("MinGrayValue", min_gray_val)
tiff.SetField("MaxGrayValue", max_gray_val)
tiff.SetField("Gradient", gradient)
tiff.SetField("AxisIntercept", axis_intercept)
tiff.SetField("Altitude", altitude)
tiff.SetField("IsBlackLineCorrection", is_blac_corrected)
tiff.SetField("IsAtmosphereCorrected", is_atmo_corrected)
tiff.SetField("IsCalibrated", is_calibrated)
tiff.SetField("IsNormalized", is_normalized)
tiff.SetField("TransparentPixel", transparent_pix)
# Write Base Data Image
tiff.write_tiles(image_data)
tiff.WriteDirectory()
# Write multi-resolution overviews (or not)
tiff.SetDirectory(0)
_write_oneres(image_data, pixel_xres, pixel_yres)
for index, scale in enumerate((2, 4, 8, 16)):
shape = (image_data.shape[0]/scale,
image_data.shape[1]/scale)
if shape[0] > tile_width and shape[1] > tile_length:
tiff.SetDirectory(index + 1)
_write_oneres(image_data[::scale,::scale], pixel_xres*scale, pixel_yres*scale)
tiff.close()
log.info("Successfully created a NinJo tiff file: '%s'" % (output_fn,))
if __name__ == '__main__':
import sys
try:
filename = sys.argv[1]
except IndexError:
print >> sys.stderr, "usage: python ninjotiff.py <ninjotiff-filename>"
sys.exit(2)
for inf in info(filename):
print inf, '\n'
|