July 2013 - cgma-dev - mailman3.cels.anl.gov

Re: [cgma-dev] CGMA with Cubit not reading ACIS files
by David Thompson 31 Jul '13

31 Jul '13

Hi Iulian, > We are building usually on linux only; ... The software we're working on needs to support the big 3 (Linux, OS X, Windows). > Rajeev is building on 64 bit OSx, but only with occ. > Does it pass make check for you or not? Yes. I have to fiddle with DYLD_LIBRARY_PATH because of the way Cubit is prepared on OS X, but make check passes. Thanks, David > > You cannot configure with cubit and occ at the same time. > > I did not. > > > What is your configuration ? > > CXX=clang++ CC=clang CXXFLAGS="-arch i386" CFLAGS="-arch i386" /path/to/cgm/configure --prefix=/path/to/cgm/x86_install --enable-shared --with-cubit=/Applications/Cubit-13.1/Cubit.app/Contents/MacOS > > > Thanks, > David > > > > Hi all, > > > > I have a version of CGMA working with OCC but when I build with cubit support, some simple test programs I wrote don't work any longer. Specifically, calling > > > > const char* filename = "/path/to/cgm/test/model1.sat"; > > ModelImportOptions opts; > > DLIList<RefEntity*> entities; > > stat = GeometryQueryTool::instance()->import_solid_model( > > filename, ACIS_SAT_TYPE, opts, entities); > > > > returns CUBIT_SUCCESS and prints a non-zero number of "ACIS Entities" have been read but then > > > > 1. entities is an empty list and > > 2. calling GeometryQueryTool::instance()->bodies() also returns an empty list. > > > > Has anyone else seen this problem? > > > > Thanks, > > David

1 0

Re: [cgma-dev] CGMA with Cubit not reading ACIS files
by Iulian Grindeanu 31 Jul '13

31 Jul '13

1 0

Re: [cgma-dev] CGMA with Cubit not reading ACIS files
by David Thompson 31 Jul '13

31 Jul '13

Hi Iulian, > You cannot configure with cubit and occ at the same time. I did not. > What is your configuration ? CXX=clang++ CC=clang CXXFLAGS="-arch i386" CFLAGS="-arch i386" /path/to/cgm/configure --prefix=/path/to/cgm/x86_install --enable-shared --with-cubit=/Applications/Cubit-13.1/Cubit.app/Contents/MacOS Thanks, David > Hi all, > > I have a version of CGMA working with OCC but when I build with cubit support, some simple test programs I wrote don't work any longer. Specifically, calling > > const char* filename = "/path/to/cgm/test/model1.sat"; > ModelImportOptions opts; > DLIList<RefEntity*> entities; > stat = GeometryQueryTool::instance()->import_solid_model( > filename, ACIS_SAT_TYPE, opts, entities); > > returns CUBIT_SUCCESS and prints a non-zero number of "ACIS Entities" have been read but then > > 1. entities is an empty list and > 2. calling GeometryQueryTool::instance()->bodies() also returns an empty list. > > Has anyone else seen this problem? > > Thanks, > David

1 0

Re: [cgma-dev] CGMA with Cubit not reading ACIS files
by Iulian Grindeanu 31 Jul '13

31 Jul '13

1 0

CGMA with Cubit not reading ACIS files
by David Thompson 31 Jul '13

31 Jul '13

Hi all, I have a version of CGMA working with OCC but when I build with cubit support, some simple test programs I wrote don't work any longer. Specifically, calling const char* filename = "/path/to/cgm/test/model1.sat"; ModelImportOptions opts; DLIList<RefEntity*> entities; stat = GeometryQueryTool::instance()->import_solid_model( filename, ACIS_SAT_TYPE, opts, entities); returns CUBIT_SUCCESS and prints a non-zero number of "ACIS Entities" have been read but then 1. entities is an empty list and 2. calling GeometryQueryTool::instance()->bodies() also returns an empty list. Has anyone else seen this problem? Thanks, David

1 0

Re: [cgma-dev] Observer question?
by Tim Tautges 22 Jul '13

22 Jul '13

On 07/22/2013 04:20 PM, David Thompson wrote: > Hi Tim, > >>> ... for tracking persistent geometry across operations ... Would you advise adding UUIDs as attributes ...? >> >> No; you'd get exactly the same results. We hang an attribute on the engine entities pointing to the CGM entities, then after the operation, see what's left, and remove the CGM entities that are no longer referenced. So in a sense, we only get "before" and "after" pictures, not "during" pictures. > > > I think we're in violent agreement; if there is no "during" picture, then I should just fetch entities "after" which have the same entityId as they did "before." Is that what you meant? > Yep. And, CGM should do that for you already. That is, the CGM entities that survive will persist, and the others will have their observer functions called before they get deleted. - tim > Thanks, > David > -- ================================================================ "You will keep in perfect peace him whose mind is steadfast, because he trusts in you." Isaiah 26:3 Tim Tautges Argonne National Laboratory (tautges(a)mcs.anl.gov) (telecommuting from UW-Madison) phone (gvoice): (608) 354-1459 1500 Engineering Dr. fax: (608) 263-4499 Madison, WI 53706

1 0

Re: [cgma-dev] Observer question?
by David Thompson 22 Jul '13

22 Jul '13

Hi Tim, >> ... for tracking persistent geometry across operations ... Would you advise adding UUIDs as attributes ...? > > No; you'd get exactly the same results. We hang an attribute on the engine entities pointing to the CGM entities, then after the operation, see what's left, and remove the CGM entities that are no longer referenced. So in a sense, we only get "before" and "after" pictures, not "during" pictures. I think we're in violent agreement; if there is no "during" picture, then I should just fetch entities "after" which have the same entityId as they did "before." Is that what you meant? Thanks, David

1 0

Re: [cgma-dev] Observer question?
by David Thompson 22 Jul '13

22 Jul '13

Hi Jane, > As Tim said, after boolean operation, the engine determines which geometry get changed. And if you see GEOMETRY_TOPOLOGY_MODIFIED, you definitely need to call OCCBody::update_OCC_entity to get them synched. Thanks, that's very helpful to know. I was hoping that CGM provided some functionality similar to this across all the geometry engines, but this is the next best thing. David

1 0

Re: [cgma-dev] Observer question?
by Tim Tautges 22 Jul '13

22 Jul '13

On 07/22/2013 02:51 PM, David Thompson wrote: >>> ... >>> The notify_observer() method called on observers doesn't appear to provide information about the entity(-ies) replacing those being modified. Is there a way to obtain the resulting split or merged entities? If I create an attribute to hold an ID on faces or edges of interest, how fast will lookup be? >> >> If it's anything like ACIS, you generally don't know the correspondence between an entity that gets deleted and the ones used to replace it. IIRC you don't get many notifications of split or merge type, only of delete type (and implicitly create type, which show up as not having associated CGM entities yet). > > > Hmm. If different engines behave differently, then I guess we shouldn't rely on it for tracking persistent geometry across operations. Would you advise adding UUIDs as attributes for that kind of tracking? > No; you'd get exactly the same results. We hang an attribute on the engine entities pointing to the CGM entities, then after the operation, see what's left, and remove the CGM entities that are no longer referenced. So in a sense, we only get "before" and "after" pictures, not "during" pictures. - tim > Thanks, > David > -- ================================================================ "You will keep in perfect peace him whose mind is steadfast, because he trusts in you." Isaiah 26:3 Tim Tautges Argonne National Laboratory (tautges(a)mcs.anl.gov) (telecommuting from UW-Madison) phone (gvoice): (608) 354-1459 1500 Engineering Dr. fax: (608) 263-4499 Madison, WI 53706

1 0

Re: [cgma-dev] Observer question?
by Jiangtao Hu 22 Jul '13

22 Jul '13

Hi, David: As Tim said, after boolean operation, the engine determines which geometry get changed. And if you see GEOMETRY_TOPOLOGY_MODIFIED, you definitely need to call OCCBody::update_OCC_entity to get them synched. Or if you have BRepBuilderAPI_MakeShape *op or LocOpe_SplitShape* sp, you can call op->Modified(old) or op->Generated(old) or sp->DescendantShapes(old) to see what the engine returns for the changed geometry. Again, the boolean operation for OCC engine is not perfect itself, especially dealing with failure case, for issues that you think the operation should or shouldn't work, you can report them at http://www.opencascade.org/org/forum/ . Best luck for you! Jane Hu Asst. Researcher Dept. of Engineering Physics UW @ Madison "And we know that for those who love God, that is, for those who are called according to his purpose, all things are working together for good." (Romans 8:28) ________________________________ From: Tim Tautges <tautges(a)mcs.anl.gov> To: David Thompson <david.thompson(a)kitware.com> Cc: cgma-dev(a)mcs.anl.gov; Jiangtao Hu <jiangtao_ma(a)yahoo.com> Sent: Monday, July 22, 2013 12:45 PM Subject: Re: [cgma-dev] Observer question? On 07/22/2013 11:01 AM, David Thompson wrote: > Hi Tim, > >> Good hearing from you, when did you leave Sandia? > > It's been about a year. > >> On the observer question, entities that are not intersected will persist, including the RefEntity's, though whether >> something isn't modified depends on what the engine does (it's keyed off whether the underlying engine changes its >> topological entity). With Cubit, we've had mostly expected results that way, so no big worries. > > Any idea about the OCC backend? I ask because the attached program calls webcut on one of the example models > (62_shaver1.brep) and faces to both sides of the cut sheet (as well as one that intersects the sheet) all report > GEOMETRY_TOPOLOGY_MODIFIED. > Hmm, I haven't worked much with the OCC backend, so I couldn't say. I'm cc'ing Jane Hu, who developed that backend, she might know. > The notify_observer() method called on observers doesn't appear to provide information about the entity(-ies) > replacing those being modified. Is there a way to obtain the resulting split or merged entities? If I create an > attribute to hold an ID on faces or edges of interest, how fast will lookup be? > If it's anything like ACIS, you generally don't know the correspondence between an entity that gets deleted and the ones used to replace it. IIRC you don't get many notifications of split or merge type, only of delete type (and implicitly create type, which show up as not having associated CGM entities yet). - tim > Thanks, David > -- ================================================================ "You will keep in perfect peace him whose mind is steadfast, because he trusts in you." Isaiah 26:3 Tim Tautges Argonne National Laboratory (tautges(a)mcs.anl.gov) (telecommuting from UW-Madison) phone (gvoice): (608) 354-1459 1500 Engineering Dr. fax: (608) 263-4499 Madison, WI 53706

1 0