malaterre [Wed, 11 May 2005 15:01:46 +0000 (15:01 +0000)]
BUG: Fix previously instroduce cmake bug, now should compile fine even without VTK at all. Fix float/double in VTK4.2 and finally backport a patch from ITK
malaterre [Wed, 11 May 2005 14:40:56 +0000 (14:40 +0000)]
ENH: Some minor enhance:
1. Support compilation of gdcm against VTK 4.2 but not testing
2. Support compilation of gdcm agains VTK <= 4.4.2 but not testing
3. Complain if the version is not matched
4. minor compilation glitch
jpr [Tue, 3 May 2005 11:06:22 +0000 (11:06 +0000)]
" This program allows to see at a glance"
" if the gdcm::File remains unimpaired after a Write"
" In a future step, we could move it to"
" gdcm Testing, for a systematic checking"
" of the entire dataset"
" Later ..."
jpr [Tue, 3 May 2005 10:57:27 +0000 (10:57 +0000)]
gdcm::File::IsReadable() is no usable here, because we deal with
any kind of gdcm::Readable *document*
not only gdcm::File (as opposed to gdcm::DicomDir)
BTH : how can you call a method belonging to parent class;
when overloaded in the child class, with VS60?
Aware user (who *does* know all the files whose names
are in InternalFileNameList are 'coherent', that is to say :
- exist
- may be open,
- are gdcm-readable
- have the same sizes
- have the same 'pixel' type
- are single frame
- have the same color convention
- anything else ?
is now allowed to ask for a 'light' checking (actualy : initialization, from the
first file characteristics), using vtkGdcmReader::SetCheckFileCoherenceLight()
Add a 'heavy documented' example on the way to extract 'overlays'
when they are stored in the 'ACR fashion'
(that's NOT to say in ACR-NEMA images; it still works an DICOM V3 images)
CONP: Alrite seems like that before gcc 2.96 ios::failure was not defined. I could only find ref to ios_base::failure. This thing is really poorly documented my guess is that this is not really in c++ but just a stupid extension
Now, user can ask gdcm::DicomDir to skip SeqEntries or shadow groups, when
possible :
// Old style (still available) :
dcmdir = new gdcm::DicomDir(dirName, true);
// new style (user is allowed no to load Sequences an/or Shadow Groups)
dcmdir = new gdcm::DicomDir( );
dcmdir->SetParseDir(true);
dcmdir->SetLoadMode(NO_SEQ | NO_SHADOW);
dcmdir->Load(dirName);
* Dicts/DictGroupName.dic : Remove all entries defined by "???"
* gdcmPython/gdcm.i : add equivalent to GDCM_UNFOUND and other
string constants defined in gdcmCommon.h.
-- BeNours
regrain [Thu, 31 Mar 2005 09:46:51 +0000 (09:46 +0000)]
* src/gdcmPixelReadConvert.cxx : problems of around of double to uint8, when
conversion of YBR to RGB pixels. So, now, the conversion is made using
integers only
-- BeNours
regrain [Wed, 30 Mar 2005 15:30:32 +0000 (15:30 +0000)]
* Testing/TestAllReadCompareDicom.cxx : now use test files (invented file
format) to replace 'at-run' dicom files generated by the
TestAllReadCompareDicom test
-- BeNours
jpr [Wed, 23 Mar 2005 16:54:40 +0000 (16:54 +0000)]
Temporary black lis of LibIDO rectangular images ...
Probabely no human beeing in the world cares about LibIDO rectangular images
but Creatis users -that still don't use gdcm ?-
(It's too sad to see failed tests
I'll fix the bug tomorow)
JPRx
jpr [Tue, 22 Mar 2005 11:29:37 +0000 (11:29 +0000)]
* FIX : element 0x0000 of shadow groups is now force to a ValEntry, and it's VR
is forced to UL
* ENH : user id now allowed to tell he doesn't want to deal with Sequences
and/or with Shadow groups
use SetLoadMode (NO_SEQ | NO-SHADOW);