-
Notifications
You must be signed in to change notification settings - Fork 5
/
ChangeLog
1183 lines (895 loc) · 42 KB
/
ChangeLog
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
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
StarSpan ChangeLog
$Id: ChangeLog,v 1.74 2008-05-08 01:08:12 crueda Exp $
2008-05-07 (1.0.08)
- starspan_create_strip(): Explicitly assigning the geoTransform to the
generated strips. It uses the geoTransform from the first miniraster but
sets the origin point to be (0,0).
In particular, for the --mini_raster_strip command with shapefile creation,
this adjustment makes the generated datasets to match properly.
- Since starspan now checks that the projection of the input datasets be
the same, I converted the polygon shapefile for the current tests to
match the projection of the raster files:
cd tests/data/vector
ogr2ogr -t_srs "+proj=utm +zone=10 +ellps=WGS84 +datum=WGS84 +units=m +no_defs" \
ply/ply.shp starspan_testply.shp
Also converted the other shapefiles for future use.
(cd tests && make) # -> OK
2008-05-01,02
- Refactoring miniraster strip creation to allow duplicate pixel handling:
- MiniRasterObserver: no Traverser as a member any more.
- IntersectionInfo: now includes the traverser.
- MiniRasterStripObserver: setOutVector() and setOutVectorDirectly()
to allow association of output vector either with no ownership or with
ownership.
- MiniRasterStripObserver.create_strip() removed. Instead, this functionality
was moved to new utility starspan_create_strip().
- New starspan_getMiniRasterStripObserver2() and starspan_minirasterstrip2()
to handle the generation of the strip with duplicate_pixel modes.
- TODO: complete preliminary tests.
2008-04-24
- GRASS interface documentation updated in web site:
https://starspan.casil.ucdavis.edu/doku/doku.php?id=grass
- Rasterization process documented in:
https://starspan.casil.ucdavis.edu/doku/doku.php?id=rasterization
--rasterize option still internal for testing purposes.
Also added rasterize command to GRASS interface to facilitate this
testing, although the GDAL GRASS driver cannot create new GRASS rasters;
using ENVI for output rasters for now.
2008-04-22
- Implemented grass interface commands:
report
csv
- Each command is GRASS-configured individually so it is better dispatched
in terms of pertinent options. Also, the command can be dispached from
the GRASS GUI, including selection according to input map: vector or raster.
NOTE: starspan_dup_pixel() was segfaulting with versions geos 2.2.3 and
gdal 1.5.0. Just updated geos to 3.0.0 (on pegasus) and problem fixed.
2008-04-21
- New starspan_grass.cc: preliminaries for providing a GRASS interface.
2008-04-18
- New starspan_rasterize.cc: goal: do some basic comparisons with the
output generated by GRASS' v.to.rast.
2008-04-17
- Code cleanup:
- Removed (not used):
- starspan_csv_dup_pixel.cc
- starspan_db.cc
- starspan_tuct1.cc
- starspan_update_dbf.cc
- function starspan_myErrorHandler()
- Improved comments to some sources
- Moved src/rasterizers/agg_*.h to new dir src/rasterizers/agg/
(agg is used only for line rasterization but other types of
rasterization may eventually be located in src/rasterizers/.)
2008-04-11
- new IntersectionInfo struct to add more information in intersectionFound()
and intersectionEnd() traversal observer interface.
- Shapefile creation for mini_raster_strip now working including for --box
option. Tested as follows (under tests/):
* Mini raster strip with accompanying shapefile and using the –in option:
../starspan2 \
--vector data/vector/starspan_testply.shp \
--raster data/raster/starspan2raster.img \
--mini_raster_strip generated/mrstrip/myoutput_ \
generated/mrstrip/myoutput.shp \
--in \
--separation 10
* Mini raster strip example with accompanying shapefile and using the –box option:
../starspan2 \
--vector data/vector/starspan_testply.shp \
--raster data/raster/starspan2raster.img \
--mini_raster_strip generated/mrstrip_box/box_100_ \
generated/mrstrip_box/box_100_.shp \
--box 100 \
--separation 10
See: https://starspan.casil.ucdavis.edu/doku/doku.php?id=minirasterstrip
2008-04-10
- Some code cleanup:
Removed vector/{Makefile,Vector_shapelib.cc}
- Adding shapefile creation for mini_raster_strip
2008-04-05 (1.0.08)
- tests/Makefile: use zcat for comparison instead of gunzip
(so the .gz, which is on CVS, is preserved).
- Removed FR 200347: --mini_raster_box option
- Added FR 200348: extract a centered box wrt feature's bounding box.
This is implemented with the --box option, which behaves as a variant
of the --buffer operation, so it's handled by traverser.
Only one of --buffer or --box may be specified.
- Removed starspan_minirasterstrip.cc (function starspan_miniraster_strip())
Was never implemented.
2008-03-27 (1.0.08)
- Fresh release to facilitate creation of windows binaries.
2008-03-03 (1.0.07)
- Adding options --sql, --where, and --dialect à la ogrinfo.
Preliminary tests OK.
More details about the syntax for --sql: https://gdal.org/ogr/ogr_sql.html
The "where" section corresponds to the --where option.
In starspan you will write something like for example:
starspan --vector myvector.shp \
--where "(SITE_ID >= 10) and (SITE_ID <= 20)" ...other options...
Although more tests are pending, I think these new options should work as
expected.
- removed unneeded starspan_csv_dup_pixel.cc from Makefile.am
- configure.in, starspan.in: using program_prefix and program_suffix to
refer to binary executable, so the corresponding ./configure options
are honored.
- Adjustments to GEOS include files according to deprecated warning
issued by GEOS 3.0.0
- Since GDAL and GRASS are working nicely together on pegasus, I just made a
basic test to confirm that starspan can access vector and raster GRASS
datasets, and everything seems to work fine. For example, according the
the naming of the datasources explained in https://gdal.org/ogr/drv_grass.html
and https://gdal.org/frmt_grass.html:
starspan --raster /proj/delta/gis/grass/delta/lidar_shadows/cellhd/200506_p1_31_beam_irradiance --report
starspan --vector /proj/delta/gis/grass/delta/carueda/vector/starspan_testply/head --report
Note that you don't have to launch these commands from within GRASS to access
the datasets. However, more information is reported for rasters
when the command is launched from within GRASS.
Note also that accepting the inputs according to the GRASS syntax, for example:
starspan-grass raster=00506_p1_31_beam_irradiance@lidar_shadows vector=somevector@somewhere ...etc...
is another story; this would required a concrete module in starspan especially
designed to use the GRASS APIs.
2008-02-05
- Included --gui extended option to launch the GUI
2008-02-02
- starspan_csv_dup_pixel.cc: Minor changes in log messages for
easier inspection.
2008-01-28
- Added processing of rasters according to associated masks.
See starspan_csv_dup_pixel.cc
Tests done with:
VECTOR=/proj/delta/gis/vector_data/phenology/200406/all_emergent_polyg_2004.shp
RASTERS=/proj/delta/raster/hymap/classification/version_4/200406/inputs/indexes/sdw_p1_*_indx
MASKS=/proj/delta/raster/hymap/base/ref1_geo3/200406/mask_edge/200406_p1_*_ref1_mask.img
cd /proj/delta/raster/hymap/classification/version_4/200406/results
starspan --duplicate_pixel direction 90 --vector $VECTOR \
--raster $RASTERS --mask $MASKS \
--csv test_dupixel.csv --verbose > test_dupixel.log
- TODO: mask handling on other options/commands -- now, only done for the
--duplicate_pixel and --csv commands
2008-01-24 (1.0.06)
- First tests on --duplicate_pixel
- Dev note: layer->ResetReading() call in traverse()
New static flag: bool _resetReading in class Traverser:
By default, traverse() starts by calling ResetReading on the layer
(_resetReading == true).
However, this is in general not appropriate because it precludes the
traversal from being used as a subtask.
For example, in starspan_csv_dup_pixel, all features are processed
but the extraction from each one is done with a corresponding traversal.
As a quick fix to allow for this behavior at least for specific commands,
this global (static) flag can be used to instruct the traversal not
to call ResetReading on the layer.
- Fixed some memory errors and leaks -- thanks valgrind, again!
2008-01-16 (1.0.06)
- FR #350112: Duplicate pixel handling -- starting implementation.
See discussion: https://starspan.casil.ucdavis.edu/doku/doku.php?id=fr_350112
2007-12-14 (1.0.05)
- FR #200332: Added --delimiter option, now used in:
function command
------------------- ------------
starspan_csv --csv
starspan_stats --stats
starspan_update_csv --update-csv
starspan_getCountByClassObserver --count-by-class
In --update-csv, the given delimiter is used for reading in the given
input CSV file and to generate the new CSV file.
2007-12-10 (1.0.04)
- Changes to compile StarSpan against GEOS 3.x.x
See GEOS_VERSION_MAJOR-conditioned code in:
traverser.h, traverser.cc
Other various changes consisted in dropping the geos::* namespace,
which is now handled in traverser.h
NOTE: Successfully tested against 3.0.0rc3 on various machines but
was unsuccessful against 3.0.0rc4 (latest at this date): an error
at link time relate with not finding destructor ~Coordinate or
something liket that.
2007-12-03 (1.0.03) Performance inprovement with spatial filter.
- traverser.cc: Spatial filtering is now used by default. In case no
spatial filtering is desired, define the environment variable
STARSPAN_NO_SPATIAL_FILTER.
Performance comparison with the --stats command on the following inputs:
vector: /proj/delta/gis/field_work/field_data/200706/data3/all_teams_nudged.shp
64 rasters: /proj/delta/raster/hymap/base/ref1_geo3/200706/*_ort
Rasters with intersecting pixels: 47
Rasters with No intersecting pixels: 17
'time' reported: real user system
With spatial filter: 133.63 76.34 16:35.81
No spatial filter: 20m10.507s 5m55.868s 1m23.537s
In this case the improvement gives a ~10 times speedup!
2007-12-01
- Testing the use of spatial filter to get features:
STARSPAN_USE_SPATIAL_FILTER=1 starspan ...
under tests/, for example:
make STARSPAN=starspan STARSPAN_USE_SPATIAL_FILTER=1 test_csv
2007-11-30 (1.0.02)
- --stats now accepts multiple rasters; each one processed at a time
- NOTE about the call observer->intersectionFound(OGRFeature* feature)
in Traverser::process_feature:
Particularly in the case of a GeometryCollection, it might be the
case that this feature is found to be an intersecting one even though
that no pixel will be found to be intersecting. Observers may in general
want to confirm the actual intersection by looking at the number of pixels
that are reported to be intersected.
2007-11-19
- Option --fields now also accepts "none"
- Fixed argument check for --RID
2007-11-16: Carlos Rueda
** Starting version 1.0.xx **
- configure.in to version 1.0.01
- --report too restrictive about inputs: fixed.
2006-3-25: perrygeo
- configure.in to version 0.998
- src/tranverser/polyqt.cc : relative epsilon instead of constant
2006-03-24: perrygeo
- Exposed the SUM statistic
- Removed NoData/Null Values from the stats calculations
- Calculate two new stats: MEDIAN and NULLS (ie number of nodata
values)
2005-11-17
tests/misc/LFS/: new code to test LFS
2005-08-18 (0.997)
- 0.997 created to synchronize the whole thing.
2005-08-17 (0.996)
- mksrcdist.sh: set DISTDIR to /tmp/DISTDIR-starspan/
The problem was that ./reconf on the exported contents was not
creating the files depcomp, missing, and install-sh, required
for a successfull ./configure; make; ... on the user machine.
It seems the reason is that automake finds those files in ../ or
../../ and then doesn't create new copies. I looked for an
option to force automake to create those files but didn't find
any; so a simple solution is to create the distribution in a
different directory not descending from this one.
2005-07-04 (0.996)
Refactoring on management of set of visited pixels (pixset) with
the goal of optimizing memory usage in the case of big features:
Traverser.h declares new classes PixSet and PixSet::Iterator
New source file: pixset.cc
PixSet abstracts the operations on a set of pixel locations and
PixSet::Iterator provides a means to scan the contained locations
in the set (I have followed the Java style for the iterator
instead of the C++ one).
The first implementation of PixSet and PixSet::Iterator is based
on the set<EPixel> structure as in the previous version; this is
a preliminary test before using a different approach.
2005-07-01 (0.995)
- Traverser: Put some variables as instance member to avoid
recomputation: pix_abs_area, pixelProportion_times_pix_abs_area.
- Traverser::rasterize_poly_QT revisited:
- all-discard case is now under the condition that pixelProportion > 0.0.
- Fixed comparison in condition to include the whole envelope
- Traverser.h _Rect.lowerRight() fixed bug which generated the
same rectangle representing a single pixel
These changes also fix bug #200097.
2005-06-18 (0.995)
STARSPAN_VERSION only needs to be defined in one place: configure.in
2005-06-18 (0.9941 - interim)
- CVS: removed generated files; new reconf script;
mksrcdist.sh.in updated to create public distribution.
- some refactoring
2005-06-06 (0.994)
- Added option --show-fields
- Elapsed time reported only if --verbose is given
2005-06-05 (0.994)
starspan.in: Fixed calls using the list of passed arguments; now
"$@" is used instead $*, which was wrong particularly to deal
with arguments containing spaces. Note also that just $@
(without the double quotes) won't work either.
2005-05-26 (0.994)
Version set to 0.994
Some minor documentation updates.
Removed: --RID_as_given
Added: --RID {file | path | none}
Style for RID values in output.
file : the simple filename without path
path : exactly as obtained from command line or field in vector
none : no RID is included in output
Updated: --csv now processing --RID
Check --raster_directory is only given with --raster @field scheme.
2005-05-21 (0.994)
PATH in starspan{.in} was not set to include GDAL and GEOS's bin/
directories. Fixed.
todo: check proper environment to run tests
2005-04-28 (0.993)
Hack in starspan_myErrorHandler (starspan_util.cc) as a workaround
in current GDAL code which doesn't implement SetScale/SetOffset
methods for some raster band types.
Traverser code cleanup (traverser.h, traverser.cc):
Removed old methods for rasterization of polygons:
processValidPolygon_FF(geos_poly)
processValidPolygon_BB(geos_poly)
Versions containing those methods are:
traverser.h,v 1.9
traverser.cc,v 1.10
starspan_countbyclass.cc: verbose messages and flush for each
finished feature.
Note: configure doesn't set execution permission to script
starspan. This flag is set by `make install' on destination.
For running the tests under tests/, I added a `chmod +x ../starspan'
to tests/Makefile. I discarded adding a `chmod +x starspan' to
configure.in (in AC_OUTPUT), as I'm not sure if that's portable.
2005-04-28 (0.99 -released)
Quadtree algorithm implemented for polygon rasterization:
New traverser/polyqt.cc.
Preliminary tests OK. The speedup over the flood-fill algorithm
with a polygon that contains 22860 pixels is:
time(flood-fill) / time(quadtree) = 98s/4s = 24.5 (!)
TODO: More tests; check the other goemetry types resulting from
intersections that are ignored.
starspan2.cc: shutdown calls to release resources: input data
objects and calls to CPLPopErrorHandler() and geos::Unload::Release().
mksrcdist.sh: using my developer account to export from cvs
(currently anonymous access to casil is not working.)
Version released today: 0.99
2005-04-26 (0.98beta)
Changes committed.
Removed temporary NEW-SOURCES-PENDING-MERGE-COMMIT/.
Pending: configure regeneration.
2005-04-21 (***** to-be 0.98beta )
Traverser: Finally implemented a flood-filling algorithm to
rasterize polygons. Temporary source files:(*)
NEW-SOURCES-PENDING-MERGE-COMMIT/traverser/traverser.h
NEW-SOURCES-PENDING-MERGE-COMMIT/traverser/traverser.cc
Other minor changes also pending for merge and commit:
NEW-SOURCES-PENDING-MERGE-COMMIT/starspan_countbyclass.cc
(*) cvs repository (on casil) is down now, so I'll wait
until it's up again to update and commit actual sources...
2005-04-08 (0.97beta)
Setting up GNU build scheme:
File renamings:
DEVNOTES.txt --> ChangeLog
README.txt --> README
INSTALL.txt --> INSTALL
starspan --> starspan.in
New files:
configure.in
Makefile.am
AUTHORS
COPYING
NEWS
THANKS
Removed files:
Makefile.props.templ
Makefile.props.fink
2005-04-04
New BUILD.windows.txt with notes about compiling the whole
system in windows
2005-03-23 (0.963beta)
Makefile.props.templ: GDAL_SRC_DIR removed: we don't
need this any more as GDAL (from 1.2.6) already installs
vrtdataset.h in GDAL's PREFIX/include/.
traverser.cc: Fixed a bug related to rerunning a traversal
on new rasters. In particular starspan_csv.cc works on a
raster at a time. The symptom was that the first specified
raster was processed OK but not the others in a command line
like this:
starspan --raster R1 R2 ... --vector V --csv my.csv
The fix was in Traverser.removeRasters to add a call to
globalInfo.rasterPoly.empty(), which empties the contents
of the rasterPoly, so when addRaster is called, it gets
the new ring from the first given raster.
The news for end users is to say something like:
Fixed a bug in --csv command on multiple rasters which
made only the first given raster to be properly processed.
Hmm, while trying unsuccessfully my 'make test' target, I
noticed that GDAL 1.2.6 has a bug related with CreateCopy
which call new methods SetOffset/SetScale; these methods
are not implemented for envi raster bands so they inherited
from GDALRasterBand wich fail with a CPLError; That explains
this new behaviour. I have notified the GDAL dev newsgroup
about this situation to see what Frank says about that.
For now, the mini raster options ARE broken due to this
GDAL bug!
I just checkout the last gdal cvs snapshot and it seems
that the problem is still there (by looking at the code).
2005-03-23 (0.962beta)
Usage message reorganized: testing and other developer
oriented options are shown with --help-dev.
New method traverser.getPixelSetSize(), which is now
used in MiniRasterObserver.intersectionEnd().
New static method Stats.computeCounts() to count the number of
occurrences of each value in a vector<int>.
Method Traverser.getPixelSet() removed. Other minor changes
to documentation in traverser.h.
Added methods in Traverser:
getPixelIntegerValuesInBand(unsigned band_index, vector<int>& list);
getPixelDoubleValuesInBand(unsigned band_index, vector<double>& list);
which work on set of visited pixels in current feature.
Removed methods in Traverser:
getPixelIntegerValuesInBand(unsigned band_index, vector<CRPixel>* colrows, vector<int>& list);
getPixelDoubleValuesInBand(unsigned band_index, vector<CRPixel>* colrows, vector<double>& list);
StatsObserver: eliminated list of pixels in current feature. This
functionality is already provided by the traverser.
New option --count-by-class implemented in new observer
CountByClassObserver (starspan_countbyclass.cc). See starspan.h
for more information.
2005-03-04 (0.961beta)
starspan_update_csv.cc: simple printf-to-cout changes
Minor text updates in starspan script.
2005-02-26 (0.961beta)
Added option --separation <num-pixels> to include a separation
between minirasters in strip. Feature request #200074.
starspan_miniraster.cc: Since ENVI complains about "Unknown" projections
(which us the map info created by GDAL by default), I decided to set
this projection explicitly by using the projection from the (arbitrarely
chosen) first miniraster. Now ENVI opens the generated strips without
any complain. This fixes bug #200073
--report now also shows projection info for rasters.
starspan_miniraster.cc: workaround to GDAL bug which
generates repeated header items when FlushCache is called
multiple time on an ENVI dataset.
2005-02-19 (0.96beta)
--buffer <distance> [<quadrantSegments>]
If the parameter starts with '@', then it will be interpreted as the
name of the attribute from which the value will be obtained. Example:
--buffer @buf_distance 21
indicates that the buffer distance is to be taken from the field
buf_distance in the corresponding geometry feature; the quadrantSegments
is fixed to 21. This fulfills request #200066
--mini_raster_parity {even | odd | @<field>}
Implemented. This fulfills request #200069
--mr renamed to --mini_rasters
Tests pending for --buffer and --mini_raster_parity.
2005-02-19 (0.952beta)
- Making a source distribution. Example:
make dist VERSION=0.1
which is the same as:
./mksrcdist.sh 0.1
- polygon/ moved to libtests/ and NOT added to cvs (this is old stuff).
- java/ moved to libtests/ and ADDED to cvs.
2005-02-19 (0.951beta)
- new scheme to make source distribution (see mksrcdist.sh)
- source restructuring. These dirs were moved to src/:
csv jts raster rasterizers stats traverser util vector
2005-02-18 (0.95beta)
- Top-level sources moved to src/
2005-02-16 (0.95beta)
- 0.94beta released.
- starting 0.95beta
2005-02-16 (0.94beta)
- minirasterstrip:
- fid strip now also created. I tried to use RasterIO
documented capability of replicating a value; however it seems
it's buggy so as a workaround I'm just making the buffer the same
size of the region I want to fill with the replicated value.
"Background" value for the fid strip is -1 as normally FIDs start
from zero.
- loc strip now also created. (background values x=y=0 --arbitrarely)
- also set background value 0 for raster strip in all bands.
Visual tests in Envi: OK
Added test case in tests/Makefile with comparison of data files.
For some reason .hdr files are created with mapinfo repeated several
times (gdal bug?)
- Fixed a minor formatting in creation of filenames for minirasters to
keep compatibility with the way this used to be done with standard C:
Now: ... << setfill('0') << setw(4) << FID ...
2005-02-15 (0.94beta)
- mini raster strip command: implementation restarted based on
MiniRasterObserver. A subclass MiniRasterStripObserver uses a
list of basic info about each generated miniraster (mrbi_list)
to merge the minirasters into the final strip.
The basic functionality is now working (tested visually in envi).
Next:
- create the accompanying .fid and .glt rasters.
- delete temporary minirasters
(Note: starspan_miniraster_strip.cc no longer used; all this is
implemented in starspan_minirasters.cc)
- Comments about --O and --outformat proposed variations in:
https://casil.ucdavis.edu/tracker/?func=detail&aid=200072&group_id=48&atid=350112
2005-02-12 (0.94beta)
- starting 0.94beta
- I just grabbed the last GDAL cvs snapshot and build a fresh gdal. I see
that vrtdataset.h (among others that used to be under frmts) are now siblings
of gdal.h under GDAL_PREFIX/include/. So a future version of our Makefile
will use only the GDAL_PREFIX. (No more GDAL_SRC_DIR).
2005-02-11 (0.93beta)
- starting 0.93beta.
- New Makefile.props.fink (jgrn307)
- added INCLUDE += -I$(GDAL_SRC_DIR) in Makefile (jgrn307)
- More documentation in Makefile
- copyright.txt added
- README.txt now with a general description of the tool
- new doc/srccontents.txt (with old readme contents)
- TODO: clean up starspan script from particular settings!
2005-02-10 (0.92beta)
- Makefile changes and new target `dist' to create a source distribution
from current CVS repository. E.g:
make dist VERSION=0.92beta
created:
DISTDIR/starspan-0.92beta.tar.gz
2005-02-07 (0.92beta)
- Starting to work on new --mini_raster_strip option.
See starspan_miniraster_strip.cc (partially implemented)
- INSTALL.txt created (based on https://starspan.casil.ucdavis.edu/?Build)
2005-02-04 (0.92beta)
- setting up a test scheme: new directory tests/ and
tests/Makefile to apply tests.
New target `test' in master Makefile.
Structure under tests/ is as follows:
data/ Contains all required raster, vector, and other data
on which tests are to be performed.
expected/ Contains data used as expected outputs from starspan.
Currently it contains data generated by starspan itself
but that have been checked externally with Envi.
generated/ This is created and populated by running the
test cases in Makefile.
Currently the validation scheme is based on cmp-ing (binary comparing)
the outputs in generated/ with corresponding expected outputs in
expected/.
- new --buffer option. Also included a basic test case.
2005-02-02 (0.92beta)
- new traverser services to get access to set of visited pixel during
current intersecting feature.
- miniraster using nodata value to nullify pixels outside geometries.
2005-02-01 (0.92beta)
--nodata added
2005-01-24 (0.92beta)
All options with double dash.
New option --version.
2005-01-24 (0.92beta)
starspan_minirasters.cc revisited: Now using traverser scheme
(first tests..)
2005-01-23 (0.91beta)
Simplified dependencies scheme:
- No more external dependency on AGG. Now the following AGG headers
(from the agg22.tar.gz distribution) are kept under rasterizers/
agg_basics.h
agg_dda_line.h
(even these headers will eventually be eliminated after a pending
revision of the line rasterizer.)
- DBF format support dropped: no more shapelib dependency.
The following sources are no longer maintained:
starspan_db.cc
starspan_update_dbf.cc
In summary, the only dependencies are now GEOS and GDAL.
2004-11-13
New options: -raster_field, -raster_directory
starspan_csv_raster_field.cc
Working only for points.
MyPolygonizer eliminated; now geos::Polygonizer makes proper destruction.
2004-11-12 ****
Reverting to version at 2004-11-09 -- cvs -q update -D "2004-11-09"
P starspan.h
P starspan2.cc
P starspan_csv.cc
P starspan_jtstest.cc
P raster/Raster.h
P raster/Raster_gdal.cc
P traverser/traverser.cc
P traverser/traverser.h
I'm keeping the previous scheme for multiple raster in traverser,
but merging other changes I made:
Ready.
2004-11-12 (0.9beta)
starspan_csv: simplified scheme: each raster processed independently.
Output scheme:
FID, {vect-attrs}, RID, [col,row,] [x,y,] {rast-bands}
2004-11-10 (0.9beta)
Quote fields in CSV if necessary.
Changes to allow the input rasters to be at different locations.
Do tests!
2004-10-21 (0.89beta)
Some cleanup: Raster, Traverser, Stats.
To improve performance on computation of integers:
Traverser#getPixelIntegerValuesInBand(..)
Stats::compute(vector<int>& values)
StatsObserver now uses integer extraction if no band is of
floating point type.
In particular, calculation of MODE should now be more efficient
because a map<int,int> is used.
Traverser::summary with some counters for a summary report.
2004-10-20
More checks for malformed (WKT) geometries.
Such a problem generates a GEOS exception but starspan was not
catching it.
2004-10-18
Starting to use geos::Polygonizer as a means to fix bad polys.
(geos::Polygonizer is not yet in the 2.0.1 official release, I'm using
the CVS snapshop.)
The idea is to node the segments defining a "bad" (self-intersecting)
polygon, and then polygonize the resulting multilinestring.
See Traverser#processPolygon().
Now working. But there are some memory leaks associated with
Polygonizer..
(my first test of Polygonizer was in java using JTS directly:
See java/src/starspan/tools/PolygonizerTest.java.)
2004-10-14 (0.88beta)
New option -RID_as_given: by default, RIDs are the filenames without
any path. With this option RIDs will take the filenames exactly as
given in the command line.
2004-10-12 (0.87beta)
Some cleanup and checks for memory leaks:
- StatsObserver.computeResults(): vector of values released.
- Traverser::processPolygon: geos_poly released.
- Traverser: pixset : Duplication of pixel is always checked
for all intersections types except for polygons.
pixset is reset on a per feature basis, so potential duplication
of pixels accross features won't be detected.
- There is a minor memory leak in GDAL 1.2.3 (geotiff.cpp:3787)
that is already fixed in CVS. I will wait for next GDAL release.
Option -skip_invalid_polys. By default all polygons will be processed.
2004-10-11 (0.86beta)
Traverser::processPolygon: now catches GEOS exceptions which
may occur when a pixel is intersected with a self-intersecting
polygon; to be more specific it seems the problem arises when
the self-intersection happens to be in the area involved in the
pixel/poly intersection.
Currently the exceptions are just ignored so the process can
continue.
An environmental variable STARSPAN_DUMP_POLYS_ON_EXCEPTION controls
if the involved polygons should be WKT-printed to stderr.
An appropiated solution to this case is PENDING.
See https://geos.refractions.net/pipermail/geos-devel/2004-September/001045.html
2004-10-08 (0.86beta)
new util/Progress class.
Traverser: options for progress and verbose output.
added Traverser::processMultiPolygon
added Traverser::processGeometryCollection
Traverser::setVerbose
Traverser::setProgress
Traverser::setLog
2004-10-07 (0.85beta)
traverser::processPolygon_point removed. Now intersection
of pixels and polygons are only done according to proportion of
intersected area (0.5 by default).
(previous revisions: traverser.cc,v 1.34, traverser.h,v 1.25)
GDAL note: Use -D_FILE_OFFSET_BITS=64 for C flags before
./configure'ing GDAL in case large file support (LFS) is not working
properly. (I fixed this kind of problem on pavo machine today.)
2004-10-06 (0.84beta)
New scheme for -calbase command:
-calbase linkname output.csv stat1 stat2 ...
-vector V : with field linkname (eg. Cal_ID)
-speclib F : csv format: linkname,1,2,...,m
-raster R1 ... : images
stat1 stat2 ... are optional desired statistics from rasters; by
default AVG is calculated
This is based on the following additions/changes:
New method Traverser#setDesiredFeatureByField(fieldname, fieldvalue).
New starspan_getFeatureStatsByField
starspan_tuct2.cc: entry point: starspan_tuct_2(...)
First tests: OK
New scheme for stats calculation: now based on a band-oriented fashion,
which should be more efficient according to GDAL model although efficiency
also depends on the specific interleave of raster datasets; in particular
BSQ is the preferred interleave for band-wise calculations.
Included "mode" statistic.
2004-10-04 (0.83beta)
starspan_stats.cc: new starspan_getFeatureStats(FID,vect,rast,stats)
to get statistics for FID in raster rast. StatsObserver was adjusted
to support the new service.
starspan_tuct1.cc: code for new ''-calbase outfile'' command.
This command expects inputs -vector V, -raster R1 R2 ...,
and (new) -speclib F.
starspan create outfile in CSV format with the following scheme:
FID,RID,BandNumber,FieldBandValue,ImageBandValue
2004-09-30 (0.82beta)
starspan_stats: bug fixed: I was calling pixels.empty() instead
of pixels.clear()!
select_fields is now a pointer to a vector.
-stats: now it also process selection of fields (-fields ...)
2004-09-29 (0.81beta)
-dump_geometries: new testing operation (implemented in
starspan_dump.cc) to generate plot of geometries and raster extent.
This provides a better control on visualization.
Output format is the textual form accepted by the Ptolemy Plot
library, https://ptolemy.eecs.berkeley.edu/java/ptplot/ .
A new extended usage option --view_geometries launches the
ptplotapplication tool.
typical usage:
starspan -vector V -raster R -dump_geometries mygeometries.csv
starspan --view_geometries mygeometries.csv
LineRasterizer: Now uses origin (x0,y0) to make a proper traslation
before calling the bresenham algorithm. Preliminary tests (via
--view_geometries) give now accurate results.
Makefile.props.templ: New scheme for Makefile.
This template is to create Makefile.props according to settings
in the local system. Now Makefile includes Makefile.props.
A new property STARSPAN_PREFIX specifies where the starspan tool
will be installed.
starspan shell script updated with better documentation about
Java archive dependencies (extended usage).
2004-09-28 (0.8beta)
StatsObserver: vector<Pixel> pixels (was vector<Pixel*>).
getBandValues renamed to getBandValuesForPixel in both Raster and Traveser
classes.
2004-09-27 (0.8beta)
-update_{csv,dbf} options: fixed bug: col,row were not made 0-based for
proper extraction.
2004-09-26 (0.7beta)
pixel duplicates revisited:
LineRasterizer#line now includes flag 'last' to indicate if ending
pixel should be reported.
LineRasterizer: more tests to verify intersections by using
JTS test builder and Envi ...
2004-09-25
stats: fixed extraction bug: (col,row) were 1-based via TraversalEvent's,
but then used as 0-based.
As a general revision, all (col,row) locations are now 0-based in traverser
module, and clients have been adjusted to report them as 1-based.
building with GDAL 1.2.3 .. ok
2004-09-24
traverser:
new Observer#end() to notify observer when process
has finished.
now list of observers: addObserver instead of setObserver
StatsObserver: new observer to compute basic statistics.
command-line scheme:
-stats [avg | stdev | min | max ]... file.csv
2004-09-16
starspan2.cc: collect args before creating main objects.
(arg processing might be based on getopt or similar in a future version.)
New module csv to read CSV files--adapted from TPOP source code
https://cm.bell-labs.com/cm/cs/tpop/code.html
New operation starspan_update_csv
New operation starspan_update_dbf
(I started to use GDAL 1.2.2 which was just realeased today.)
-envisl: name for spectra: FID:col:row:x:y
New options:
-noColRow: do not include col,row fields in -csv, -dbf
-noXY: do not include x,y fields in -csv, -dbf
2004-09-13
General refactoring to allow the processing of multiple input rasters.
Traverser has now methods to specify data input objects; all object
interfaces updated.
All output formats updated for multi raster processing:
-envi, -envisl: A check is done to verify that all bands are
the same type.
-jtstest: checks only one raster is specified.
-mr: checks only one raster is specified; should be fixed for
multiple rasters but this is NOT going to be done for now.
starspan_gen_envisl:
Data file is basename.img for Envi image, and basename.sld for
spectral library.
In header for .img: samples = 1; lines = #pixels; bands = #bands
In header for .sld: samples = #bands; lines = #pixels; bands = 1
Note: data file is written out in exactly the same way (bip interleave).
2004-09-11
Problems with large file support; this is happening on ceres but
works fine on scorpio and px (my laptop).
I'm also having a strange behavior with gdal not being able to
recognize an envi image on scorpio, though it just reads fine on
ceres and px (?).
2004-09-10
Band# names now starting from 1.
-envi: creates envi image
-envisl: creates envi spectral library