[Dcmlib] Re: Private Dicom Dict.

[Jean-Pierre ROUX]

At 13:25 -0400 31/08/05, Mathieu Malaterre wrote:
>Creatis CVS User wrote:
>>CVSROOT:	/cvs/public
>>Module name:	gdcm
>>Changes by:	jpr	05/08/31 18:32:37
>>
>>Modified files:
>>	.              : TODO
>>Log message:
>>Comments: It shouldn't be too difficult to 'manualy' ask memory merging
>>of a Private Dict into the public one (I can do it soon).
>
>Before you do so, can you explain your approach ?

I was thinking on something very simple :

User that *absolutely* wants to use one of the Private Dicom 
Dictionaries you supplied in gdcm/Dicts uses a method like :
AddDictionnary( privateDictionnaryName);
or, if he wants to use a private dictionary of his own :
AddDictionnary( fullPathPrivateDictionnaryName);

(He can use both, if he knows more about the private fields than you 
did when you extracted the private dictionaries from pdf documents).

Method AddDictionnary() just merges (in memory) the contents of the 
private dictionary into the public one (same effect when user merged 
public and private dictionary source files)

Plus a method like
RemoveDictionary(dictionayName);

Till, say a fortnight ago, some users *had* to add some entries in 
the Public Dict, to be able to parse some files, containing oddities 
like :

- Implicit VR Private entries within an Explicit Transfer Syntax 
image (solved a long time ago)
- Endianness switching inside the SQItems (solved a few months ago)
- 'True length' SQItems within 'No length' Sequences, or 'No length' 
SQItems within 'True length' Sequences (solved a few weeks ago)
- Implicit VR Private 'No length' Sequences (solved a few days ago, 
using Peter' s images)

Since this last (and final ?) oddity solving gdcm user doesn't have 
to modify the Dicom Dictionary (Implicit VR private 'True length' 
Sequences  are just skipped, being considered as a hudge single 
entry, Implicit VR private 'No length' Sequences are parsed correctly)

--> The only case using or not Private Dictionary now makes a 
difference, is for some Implicit VR entries, that can switch from 
BinEntry to ValEntry.
To get the value, user will have to modify his code !

(Probabely, we shall have to remove stupid difference between 
BinEntry and ValEntry -on disc, every entry is binary. The difference 
is in the time consuming extra processing we do on what we decided to 
be a ValEntry ...)
 
Any comment ?
JP

>
>Thanks
>Mathieu
>_______________________________________________
>Dcmlib mailing list
>Dcmlib at creatis.insa-lyon.fr
>http://www.creatis.insa-lyon.fr/mailman/listinfo/dcmlib

   Jean-Pierre ROUX
   CREATIS - CNRS UMR 5515, INSERM U 630
   Laboratoire de Radiologie Experimentale
   Hopital Cardiologique
   28 Avenue du Doyen LEPINE
   B.P. Lyon-Montchat
   69394 Lyon Cedex 03
 
   Tel      : (+33) 04 72 35 74 12
   Fax      : (+33) 04 72 68 49 16
   URL      : http://www.creatis.univ-lyon1.fr
   e-mail   : jpr at creatis.univ-lyon1.fr