optical properties

[amihashemi]

Hi,

I'm trying to use Freecad to create a GDML for optical simulation using geant4. I just wonder if there is anyway to assign the optical properties. I can easily create a geometry and export it to geant4 but I'm stuck on defining the optical properties, I would appreciate any help..

Best regards,
Ami

[KeithSloan]

Thanks for reaching out, Currently optical properties is not supported, interested in adding support.

I see https://geant4-forum.web.cern.ch/t/define-optical-properties-of-materials-in-gdml-files/1822 and this points to a G01 file that has the following definitions is solids

    <opticalsurface name="surf1" model="glisur" finish="polished" type="dielectric_dielectric" value="1.0"/>
    <opticalsurface name="surf2" model="glisur" finish="polished" type="dielectric_metal" value="1.0">
    <property name="REFLECTIVITY" ref="REFLECTIVITY" />
    </opticalsurface>

And

      <skinsurface name="skinsrf1" surfaceproperty="surf1" >
      <volumeref ref="Detector"/>
      </skinsurface>

   <bordersurface name="bordersrf1" surfaceproperty="surf2" >
      <physvolref ref="pv1"/>
      <physvolref ref="pv2"/>
   </bordersurface>

Do you know where I could find the full description of the options available?

Would I be correct in that skinsurface applies to all surfaces of volume Detector?

Unclear to me what a bordersurface is
given in the example

     <position name="detinWorldp" unit="mm" x="0" y="0" z="0" />
     <position name="pos2" unit="mm" x="2000" y="0" z="0" /> 

Would the border surface be the y, z plane ?

[amihashemi]

thanks for the response, that would be great if it can be added to the GDML extension for freecad use.
There are some description here about the optical properties in gean4:
https://geant4-userdoc.web.cern.ch/UsersGuides/ForApplicationDeveloper/html/TrackingAndPhysics/physicsProcess.html#optical-photon-processes

You can check this for some example implementation of optical properties in geant4, GDML file is provided there:
https://gitlab.cern.ch/geant4/geant4/-/tree/master/examples/extended/optical

I think bordersurface is for two dimensional geometry where the reflective/absorption properties of a border is defined between two surfaces.

[KeithSloan]

(Ignoring bordersurface for now)

Okay I have pushed a branch called Optical, it parses(imports) the G01 example, setting up the related optical values
see the group 'Opticals' after import.

Not sure if I have the correct types, like for some I have assumed to be enumeration with just values as per the example.

Next stage will to support export and then allocation of surface to volume, finally creation of values via Gui

Do you have any sample GDML with optical properties you could test if they import okay with this branch.

[amihashemi]

Thank you very much. It would be great to have the optical examples in geant4 (OpNovice and OpNovice2: https://gitlab.cern.ch/geant4/geant4/-/tree/master/examples/extended/optical), I think those examples cover many of optical properties.. there is a sample GDML file in OpNovice

[KeithSloan]

Have not checked the GDML by loading into Geant4, just eye balled.

[KeithSloan]

Also there are some GDML files in (https://github.com/hanswenzel/CaTS/tree/main/gdml) with optical properties

The latest push fixes a number of barfs with these files. They use a lot of bordersurface and I have yet to work out a way of handling in FreeCAD.

[KeithSloan]

bordersurface specifies a surface and two physical volumes, if for each of the FreeCAD App::Parts ( That get created ) I set a property which is the paired volume and another which is. the surface, would that work? i.e. On export whem I scan the file and App::Parts I create a bordersurface from the PV of the Part and the PV from the referenced part, specify the surface which they should both have as a property, do you think that would work?

[amihashemi]

I'm not sure if this works, we may need to try it to see if the GDML can be read by geant4. In my understanding, geant4 shows error with the geometries have overlap, I think we should try to see how geant4 works with the surface...

[amihashemi]

does changing the new optical version to /FreeCAD/Mod add the additional optical properties to my freecad? or the GUI still needs more work?

[KeithSloan]

Not sure which OS you are using but to pick up the changes you need to switch to the branch Optical.

If you go to the GDML directory in ...FreeCAD/Mod/GDML and run git. i.e. git status.

To fetch the branch you should be able to do something like git fetch origin Optical, git checkout Optical.

The Addon_Manager tool in FreeCAD 0.20 has the ability to switch branches, but when I tried the latest Release Candidate there were some regressions, due I think to Python 3.9.

The current code will import, there is not any facility yet to create new definitions from scratch, so I would create the optical definitions in an XML file that can be imported i.e. on top of any existing definitions.

You should be able to change things via the GUI.

[amihashemi]

Thank you! I use Ubuntu 20.04. I used Freecad 0.20 earlier but had some issues with Addon_Manager, so I had to switch to the 0.19 version, perhaps the problem was within Python...
Would you please share some of your XML files, I just want to have an example before trying mine..
In my understanding, I can use Freecad to create any geometry and then import an XML file that contain the optical properties of the materials and surfaces, but still I need to have an XML file that have detailed information corresponding to the created geometry in Freecad. Optical version of GDML basically then stack the data from XML and freecad and can export them as a new GDML file. Am I correct?

[KeithSloan]

Thanks - I tested things and found some bugs, so please fetch the latest version.

Some example xml files are now in the Resources directory e.g. opticals.xml

You always have to setup the GDML structure first i.e. in the GDML workbench, create a new file.
You could create GDML object geometry and then import say the opticals.xml file or the other way round i.e. import the opticals first

After you imported the opticals XML file you can gthen go to a GDML Volume i.e. FreeCAD Part, use the properties Window to change the SkinSurface property. Well you should be able to but there is a bug I need to fix as the enumeration is not correctly set for imported XML files, it works if the GDML file is opened.

[KeithSloan]

Okay with latest version of Optical branch, Parts created by commands or icons should now have skinSurface property which can be changed.

[amihashemi]

ok maybe I need to fetch the branch again, there might be an issue...

[amihashemi]

Also I noticed that Opticals:Surface:finish is only set for "polished", I think other options available for it, so it would be great if we can change it not being set as "polished"...

[KeithSloan]

Could you let me know what the options are? I would prefer to keep as an enumerated list.

[amihashemi]

sure,
check these two materials: https://indico.cern.ch/event/789510/contributions/3279418/attachments/1818134/2972494/AH_OpticalPhotons_slides.pdf (page 11) and https://geant4.web.cern.ch/sites/default/files/geant4/collaboration/workshops/users2002/talks/lectures/OpticalPhoton.pdf (page 15).

There are possibilities within polished (smooth surface) and ground (rough surface) and some in between. They are indicated in page 15 of the second afformentioned links.

[KeithSloan]

Great - I will look tomorrow.

[KeithSloan]

"Is there any way that Freecad recognize the surfaces from a geometry that was created in part? like can we get six surfaces defined when we have a box?"

Not easily, what I am working on at the moment is for bordersurface determine a common face, create a new GDML object, like GDMLBorderSurface, which has details and exports as bordersurface.

Then plan is to offer an icon|command border surface that lets you select two GDMLObjects that it performs a common face for etc.

At the moment finding a common face is quite complex for some shapes and I have to debug things, see post in FreeCAD forum https://forum.freecadweb.org/viewtopic.php?f=22&t=69382

Anyway enough developing for one day, maybe tomorrow.

When you say created in part? do you mean Part or PartDesign

[amihashemi]

okay, great. Thanks!
I meant part, I never used PartDesign yet...

[KeithSloan]

Okay latest version has updated enumeration for Surface Finish.~

FreeCAD 0.20 supported more sophisticated enumerations so this should be revisited and improved when it is finally released.

Let me know if you get/see any problems

[amihashemi]

I tried new branch and X icon basically shows me new window to choose a skinSurface. But same as before, the options to choose are not available except None (there might be an issue with the Enum definition there).

[KeithSloan]

Works for me, are you sure you have imported the surface definitions.
If you look at Tree view, you should see all the groups one of which should be Opticals, under Opticals you should find Surfaces and then the actual surfaces under that see the jpg

t

[amihashemi]

yes, that works, I was pointing to the skin Surface underneath of GDML here:

[KeithSloan]

I see the problem Opticals should be at the same group level as Constants, Variables, Quantities etc.

I will correct the code that does the import.

[amihashemi]

I dragged the Opticals into the LV_Box since it is easier to deal with the properties if we have many parts. The point that I was making is that "Skin Surface" option is set for None in GDML when we click on the "LV_Box"...

[KeithSloan]

The code for setting Skin Surface currently assumes that Opticals is at the base level, I will look to change.

[KeithSloan]

I am getting confused. Skin Surface should default to None, you either have to change it via the property enumeration or use the new icon.

[KeithSloan]

Try latest version, should work if opticals has been moved.

[amihashemi]

when I select the new icon and set the surface, I get the error like:
Traceback (most recent call last):
14:10:14 File "/home/ami/.local/share/FreeCAD/Mod/GDML/./freecad/gdml/GDMLCommands.py", line 228, in onSet
14:10:14 obj.SkinSurface = surf
14:10:14 ValueError: {'sclassname': 'N4Base10ValueErrorE', 'sErrMsg': "'surf2' is not part of the enumeration in test2_worldVOL#LV_Box.SkinSurface", 'sfile': '/home/conda/feedstock_root/build_artifacts/freecad_1653826019775/work/src/App/PropertyStandard.cpp', 'iline': 494, 'sfunction': 'virtual void App::PropertyEnumeration::setPyObject(PyObject*)', 'swhat': "'surf2' is not part of the enumeration in test2_worldVOL#LV_Box.SkinSurface", 'btranslatable': False, 'breported': False}

I will try the new version

[amihashemi]

One note about the surface properties between two media, I think there should be a way to activate and deactivate the surface properties while material properties always needed for optical simulations, here is a quote from geant4 documents:
"First of all, there is no need to define a surface for polished interfaces between two media. As long as the two media have an index of refraction stored in their respective G4MaterialPropertiesTable, the [G4OpBoundaryProcess][]::[PostStepDoIt][] will handle the refraction and reflection correctly."

[KeithSloan]

Noted:

Are you still getting errors trying to set SkinSurface?

[amihashemi]

I got the error when I tried the last version. I should note when I export the gdml file I get the correct gdml file while the errors are there in freecad, there might be an issue with the GUI only?!

[amihashemi]

Hi, I just wonder if there has been any success in adding bordersurface properties to the GDML file. Thank you for your support

[KeithSloan]

If you check the link to the FreeCAD forum posthttps://forum.freecadweb.org/viewtopic.php?f=22&t=69382

You will see some suggestions of using Vertexes to check for common face, also to use a facility in Realthunders Linkstage3 version of FreeCAD which would require use of that version rather than the main supported version of FreeCAD 0.20 that has just been released.

I made a start on using Vertexes but have been busy with other commitments.

But I don't think checking vertexes would work in ALL cases for example where the common face was the top/bottom of a cylinder,
Or even boxes of different sizes.

I should have some free time from Tuesday so will try and make some progress on checking for vertexes with the same coordinates and create a couple of test cases i.e. cylinder with top/bottom circle and boxes of different sizes with a view to updating the forum post asking for a better solution.

Would a version that checked for equal vertexes have any value for you?

[amihashemi]

Great, thanks for sharing the forum discussion, I will look at the vertexes but I doubt we can define optical properties between two materials. I think I have the new version of FreeCad 0.20 and I will check on vertexes to see if I can generate workable GDML to some extent that apply for optical simulation.

For GDML with optical properties, we would generally need to have 1) optical properties of material (which contains getting few tabulated data such as reflection, absorption length, and scintillation), 2) surface properties that are set in "solids" section, 3) auxiliary for Detector (optical photon detector) and scintillator which would be defined in "structure" section, and 4) bordersurface that connect the auxiliary of detector and scintilator which is defined in "structure" section. (you may check the simpleLArTPC.gdml)

I think all of them can be done without having the surface defined in Freecad... please let me know if I can give any help

[KeithSloan]

Okay you may want to try a new branch borderSurfaceIcon.

It adds another icon ( currently just an X)

If you select, one surface and two parts and then click on it, it will create a GDMLbordersurface definition at the base Document level.
Currently NO CHECK is made that the two parts have a common Face.

Minimum checking of subsequently exported GDML.

If you think this is useful despite the lack of checking common face, then I will add support for importing existing GDML files with bordersurface definitions.

By the way when looking at the files you have suggested, I find the cycle facility (first icon) useful, the cycle being between solid, wireframe hidden

BorderSurface.FCStd.txt
OpticalSurfaces-World.gdml.txt

[KeithSloan]

Okay - there is now another new branch - bordersurfaceIconCF

It has some code to try and find a common face i.e. Distance between centre of mass of faces is zero.
It works for OpticalSurfaces.gdml but there is a problem with simpleLarTPC.gdml in that it does not find a common face.

If it does not find a common face it prompts with a dialog that allows you to override or cancel

[amihashemi]

okay, thank you very much, I will try the new branch to see how it works on my linux machine.

[KeithSloan]

Question on gdml file simpleLarTPC.gdml

The box is 1000 x 1000 x 2000

So given origin at 0,0,0 the box coordinates on x axis are -500, 500

Now the various positions of the tubes are x = 475
The Tube is radius 3 and z of 0.5

So how can tubes and box have a bordersurface when they don't touch, have a common face?

[amihashemi]

There are five tube detectors and a box scintillator, the detectors are located at the surface and attached to the scintillator to gather the optical photons upon simulation of light transport through the scintillator. All fives have common faces that are next to box with the size of 1000-1000-2000. bordersurface can be simply defined by choosing two components that we know that are attached to each other for example Det0 and box or Det1 and box. Here is how the GDML look like

[amihashemi]

Yes, I think that is set for other extended package in respect to Geant4. I think the best way is to have the options that cover the Geant4 enums and additional options depend upon the use for other extended packages (e.g. 0-9 numbers?).

[KeithSloan]

You might try branch Optical

It will now parse GDML for optical properties, simpleLarTPC.gdml will set finish to 'extended | 0' and type to 'extended | 0'

[KeithSloan]

If select either of the Groups Materials or Optical and export type xml, then an xml file with just the current Materials or Opticals is created, you can then start a new GDML file, new from GDML workbench and then import the xml file.

That is if you can use an existing GDML file as a starting point by exporting just the Materials or Opticals and then import the xml files into a GDML structure.

[KeithSloan]

Opticals branch now merged with Main branch, which is the branch that should be used with FreeCAD 0.20.

If you are still using branch master with FreeCAD 0.20 please remove and reinstall the GDML workbench via the Addon Manager

[amihashemi]

ok sounds great, I will check the main branch and do few examples with it, thank you very much for the support and time.. really appreciate it.

[amihashemi]

Hi Keith,

I came up with an issue regarding the bordersurface definitions, it appears that "set border surface" does not work properly. I put the example here that I'm working on right now, I try to define ScintWrap between every Obj# and Det# that are defined, I'm just not able to select those objects to assign them neither using the icon nor just copying ScintWrap# and trying to change it...

I would appreciate if you can take a look ...

small_pet_6.71cm_8by8MPPC.zip

Thank you!

[KeithSloan]

I find it is best to use Tree View

I had to click on"Obj" to expand it to see "Det"

Then I was able to select a Surface ScintWrap, Obj and Det, then clicking on the "Set Border Surface" icon worked for me.
You should then be able do the same for Obj001 etc

On my Mac I use the 'Command' key when selecting more than one object.

[amihashemi]

ok great thanks for the help! for some reasons, "ScintWrap" underneath of "Surfaces" was gone for mine. I fixed it it now it works, thank you!

[mhindi2]

Hi Keith, I will try to look at these issues after I get caught up on things; being away for almost a month sets you back quite a bit! Congratulations on the Masters degree! Enjoy time with the famiy. Munther

…

On 7/8/22 07:24, Keith Sloan wrote: Hi @mhindi2 <https://github.com/mhindi2>, Hope you had a good trip to Europe visiting family etc. Sorry for landing you with this on your return I am away this coming week and Masters graduation and visiting Friends and Family the following week, so wondering if you can look at this problem. I am not seeing any unknown floating variable error ( FreeCAD 0.20 and branch Main ) However I am seeing a number of errors on trying to load the exported created GDML. load_gdml.txt <https://github.com/KeithSloan/GDML/files/9072749/load_gdml.txt> Specifically |G4GDML: VALIDATION ERROR! ID value 'Obj007' is not unique at line: 87 G4GDML: VALIDATION ERROR! ID value 'Obj006' is not unique at line: 102 G4GDML: VALIDATION ERROR! ID value 'Obj005' is not unique at line: 117 G4GDML: VALIDATION ERROR! ID value 'Obj004' is not unique at line: 132 G4GDML: VALIDATION ERROR! ID value 'Obj003' is not unique at line: 147 G4GDML: VALIDATION ERROR! ID value 'Obj002' is not unique at line: 162 G4GDML: VALIDATION ERROR! ID value 'Obj001' is not unique at line: 177 G4GDML: VALIDATION ERROR! ID value 'Obj' is not unique at line: 192 G4GDML: VALIDATION ERROR! element 'physvol' is not allowed for content model '(materialref,solidref,(physvol+|divisionvol|replicavol|paramvol),loop*,auxiliary*)' at line: 245 G4GDML: VALIDATION ERROR! ID value 'skinNone' is not unique at line: 249 G4GDML: VALIDATION ERROR! ID value 'skinNone' is not unique at line: 252 G4GDML: VALIDATION ERROR! ID value 'skinNone' is not unique at line: 255 G4GDML: VALIDATION ERROR! ID value 'skinNone' is not unique at line: 258 G4GDML: VALIDATION ERROR! ID value 'skinNone' is not unique at line: 261 G4GDML: VALIDATION ERROR! ID value 'skinNone' is not unique at line: 264 G4GDML: VALIDATION ERROR! ID value 'skinNone' is not unique at line: 267 G4GDML: VALIDATION ERROR! ID attribute 'G4_CESIUM_IODIDE' is referenced but was never declared at line: 278 G4GDML: VALIDATION ERROR! ID attribute 'G4_AIR' is referenced but was never declared at line: 278 G4GDML: VALIDATION ERROR! ID attribute 'G4_SILICON_DIOXIDE' is referenced but was never declared at line: 278 G4GDML: VALIDATION ERROR! ID attribute 'None' is referenced but was never declared at line: 278 | Not sure why my Geant4 installation is not finding G4_CESIUM_IODIDE, G4_AIR, G4_SILICON_DIOXIDE. It also goes on to Barf at a number of overlapping volumes. Thanks — Reply to this email directly, view it on GitHub <#93 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AWV3NCBDKMJJYJMREN3NPIDVTA2ZHANCNFSM5YLCMEYA>. You are receiving this because you were mentioned.Message ID: ***@***.***>

[KeithSloan]

Okay I fixed one bug - If SkinSurface property is None do not export, see latest Main.

load_gdml is barfing with Volume name same as Solid name.
My install of Geant is not finding G4_CESIUM, G4_SILICON_DIOXIDE, G4_AIR
Think this is an error in my installation of Geant4 rather then a real problem with export.

Array is generating overlapping volumes :-(
load_gdml2.txt

[KeithSloan]

Okay code fixed for requirement that bordersurface needs all physvols to have a name.

For some reason World Volume given the name 'V-World', did not see why so had to hand edit the gdml file.

After that load_gdml was able to produce a display, still lots of Barfs for things it is not happy with.

load_gdml3.txt

[mhindi2]

Hi Ami, I have not had a chance yet to look at the full thread between you and Keith. Is it possible to send me the FreeCAD file that is giving you problems with arrays? At the time I developed arrays there was no problem in duplicate names, so I would like to test your file against the branch that I have. However, the branch I have does not contain Keith's developments on optical surfaces, so if it is possible to have something that gives problems with arrays without optical surfaces, I'd like to have that, so I am not dealing with two developments simultaneously (optical surfaces and arrays). Thanks, Munther

…

On 7/8/22 22:13, amihashemi wrote: Hi Munther, Thanks very much for the suggestion, I think the problem in my case is not exactly properties but the array of detectors that are located on top of the crystals. Beside the array of detector, optical properties definition for Freecad/GDML works based on the few cases that I tried. The problem here is the detector array which can not be exported to GDML and that sort of difficult to manually change in the GDML file or .cc code... Thanks for your help! Ami — Reply to this email directly, view it on GitHub <#93 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AWV3NCETVGJFYCLZVHFH42LVTEDABANCNFSM5YLCMEYA>. You are receiving this because you were mentioned.Message ID: ***@***.***>

[mhindi2]

OK, I am seeing the 'float' object has no attribute 'x' error on export. I will try to look at that. This is with my latest version, not Kieth's Main. BTW, I am not sure who this reply is going to? Can I have private addresses instead of the github ones please? Munther

…

On 7/8/22 22:13, amihashemi wrote: Hi Munther, Thanks very much for the suggestion, I think the problem in my case is not exactly properties but the array of detectors that are located on top of the crystals. Beside the array of detector, optical properties definition for Freecad/GDML works based on the few cases that I tried. The problem here is the detector array which can not be exported to GDML and that sort of difficult to manually change in the GDML file or .cc code... Thanks for your help! Ami — Reply to this email directly, view it on GitHub <#93 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AWV3NCETVGJFYCLZVHFH42LVTEDABANCNFSM5YLCMEYA>. You are receiving this because you were mentioned.Message ID: ***@***.***>

[amihashemi]

Hi Munther,

I shared the example case that I shared earlier here,
small_pet_6.71cm_8by8MPPC_array.zip

The problem is when I do the array the GDML doesn't export the file in any case. If you find the previous case of a similar crystal/detector case without the array of detectors on top of crystals the exporting to GDML works just fine...

I get your replies, and here is my personal email: [email protected]

Thank you very much!
Ami

[amihashemi]

Hi Keith and Munther,

I've been working with the array functions in a last couple of days, and applied it to many different structures and problems. It works primarily well with Geant4 application except the renaming of the objects (I found that the rename of the object, detector or scintillator would not appear in the ScintWrap even when the bordersurface is defined by selection of renamed objects), this is a small bug but I thought worth reporting here.

The bigger issue that I encounter is that the multiunion structure does not universally working in Geant4- and it appears that the multiUnion as a way of defining the structure has not been fully functional in all aspect of Geant4 particularly in new development regarding the optical simulation-- so I just wonder if I can change the array function definition in FreeCAD/GDML to a series of structural boxes and units rather than multiUnion definition, for e.g. if I have a panel array of 2*2 where the copy object is a box, then the output includes 4 box definition with transferred data points. I just wonder if this change is possible with the current GDML version where the output of the array function from GDML dump the series of copied/rotated objects.

Thank you very much!

Kind regards,
Ami

[mhindi2]

Hi Ami, Could you please provide an example file illustrating the renaming problem, or else a screenshot showing the problem. Again, for the multiUnion, could you provide an example where it does not work. Is this a FreeCAD/GDML Workbench problem or a Geant4 problem? If it is the latter then this should be reported on the Geant4 forum. I suppose we could in principle offer a preference setting that allows the export of an array as separate volumes. I think, but I don't know for sure, that Geant simulations with multiUnions run faster than the equivalent simulations with individual placements, so I would hesitate to change the export universally to be individual placements. Thanks, Munther

…

On 7/27/22 11:56, amihashemi wrote: Hi Keith and Munther, I've been working with the array functions in last a couple of days, and applied it to many different structure and problems. It works primarily well with Geant4 application except the renaming of the objects (I found that the rename of the object, detector or scintillator would not appear in the ScintWrap even when the bordersurface is defined by selection of renamed objects), this is a small bug but I thought worth reporting here. The bigger issue that I encounter is that the multiunion structure does not universally working in Geant4- and it appears that the multiUnion as a way of defining the structure has not been fully functional in all aspect of Geant4 particularly in new development regarding the optical simulation-- so I just wonder if I can change the array function definition in FreeCAD/GDML to a series of structural boxes and units rather than multiUnion definition, for e.g. if I have a panel array of 2*2 where the copy object is a box, then the output includes 4 box definition with transferred data points. I just wonder if a change like that is possible with the current GDML version where the output of the array function from GDML dump the series of copied/rotated objects. Thank you very much! Kind regards, Ami — Reply to this email directly, view it on GitHub <#93 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AWV3NCHD4EYGWG4UVHUOPO3VWGA5LANCNFSM5YLCMEYA>. You are receiving this because you were mentioned.Message ID: ***@***.***>

[amihashemi]

Hi Munther,

Here are some small snapshots in regard to the first issue:

1. snapshot shows that I selected bordersurface between the Det and Scint

2. shows that the defined ScintWrap2 still contains the names that were initially created (Part and Part001)

[amihashemi]

Munther,

Providing an example for the second issue is a bit difficult, recently Geant4 has added a package named as G4Opticks where the structures are converted to a mesh components in order to run the optical simulations on high performance computing platforms. The issue is the function that convert the structure from an object to a mesh is not yet developed for MultiUnion object, as a mater of the fact, any MultiUnion component will result in the error. The error in Geant4 looks like this (This is from using G4Opticks--):
void X4SolidBase::convertMultiUnion(): Assertion `0' failed.
Aborted (core dumped)

I still think the MultiUnion structure that you already added to GDML package is extremely useful and I used it for several structures and did use Geant4 for optical simulation, so thank you very much for adding this capability. Having said that, the recent/new capabilities of Geant4 have not included the MultiUnion feather yet, it would be nice if we can get translation of MultiUnion array to a raw series of objected box/shape definitions. This will be useful for the Geant4 feathers where the translation of structure to a mesh is targeted, but I'm not sure how difficult is to implement with the GDML package.

If the example explanation is not clear, I will try to come up with few slides to explain this better.

I'm just looking for a path to resolve this issue whether it included in the GDML package or not, I would certainly appreciate any insight.

Best regards,
Ami

[amihashemi]

example could be: 1) the structure that is created with the set of arrays/MultiUnion (GDML1) 2) the same structure only using the rotation and translation/position reference (GDML2). So the idea is can we come up with a way to convert the GDML 1 to GDML2 or only output GDML2 instead of 1?

Ami

[KeithSloan]

"I found that the rename of the object, detector or scintillator would not appear in the ScintWrap even when the bordersurface is defined by selection of renamed objects"

Thanks for reporting. The bordersurface is currently saving the info on the objects as App::PropertyString so they are not updated if the object name is changed. I will change the code to save the info as a link to the object so that if the object name is changed it should be picked up, just need to also make the appropriate change for export.

I would also like to change where the bordersurface objects 'ScintWrap(x)' are stored, namely the 'Opticals' group, say an additional group in Opticals, so Opticals becomes Matrix | Surface | Borders but would appreciate views on this - Thanks.

Munther if you could concentrate on the multiUnion - Thanks

[KeithSloan]

ASIDE

"Geant4 has added a package named as G4Opticks where the structures are converted to a mesh components in order to run the optical simulations on high performance computing platforms."

Interesting - there are lots of different ways objects can be meshed.

Not optics related but numerous people have been using FreeCAD to import STEP files from various CAD software and meshing.
Now my very limited understanding of MonteCarlo simulation is that you want the minimum mesh definition which accurately describes the shape, where as with say FEM analysis etc you want as fine mesh as possible.

Now I got Christophe Geuzaine one of the professors that developed Gmsh to add a function in Gmsh to create a minimum mesh from a CAD/STEP file. I have yet to add it to the GDML workbench but I think the function may now be in the released version of Gmsh rather than just a development snapshot.

[amihashemi]

Hi Keith, Gmsh is useful for some FEM analysis at the current stage, but the major optical simulation with Geant4 initiates a ray tracing problem. In this context, the meshing of the structure means that proxy objects are constructed based on the set of indexed triangles that photon ray tracing will be performed on them. The mesh itself is transferable to computational threads for high-performance computing as the number of generated photons is substantially high (you may check this https://developer.nvidia.com/blog/creating-optimal-meshes-for-ray-tracing/). There are many reasons for the current limitations, but at this point, we don't have the tool in Geant4 to convert the MultiUnion arrays (for only these particular simulations). Combining Gmsh mesh with Geant4 is a kind of far fetch idea to me but it is possible.

[mhindi2]

Hi Ami, I understand what you'd like to have, but it is not something that can be implemented  easily or quickly; exporting a separate solid instead of a multiUnion node is easy; what is more tricky is now exporting for each solid a separate volume and then making sure that, if the array was part of a bordersurface, that also for each node a separate bordersurface is also exported. The resulting file is going to be much larger. Re meshing: in geant simulation tests that I have conducted using meshes of otherwise simple geometrical solids the running times with meshes were 2-5 times slower than running with the equivalent simple solid. Running with clusters and/or GPUs would not reduce that time differential. Did you check with the developers of G4Opticks to see when they might support multiUnions? I'd rather not send a lot of time producing code that is then obviated when G4Opticks starts to support multiUnions. As a stop gap, may I suggest first trying to convert the array into a mesh in FreeCAD (I believe both FreeCAD and the GDML workbench have commands/icons to convert a shape into a mesh) and then exporting that as a gdml file and see what G4Opticks does. (I don't have G4Opticks installed and have never used it, so can't try the tests myself.) Munther

…

On 7/27/22 13:24, amihashemi wrote: example could be: 1) the structure that is created with the set of arrays/MultiUnion (GDML1) 2) the same structure only using the rotation and translation/position reference (GDML2). So the idea is can we come up with a way to convert the GDML 1 to GDML2? Ami — Reply to this email directly, view it on GitHub <#93 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AWV3NCGT6BVZQJCFOTCP3HDVWGLIHANCNFSM5YLCMEYA>. You are receiving this because you were mentioned.Message ID: ***@***.***>

[amihashemi]

Hi Munther,

Thank you very much for the response.

Yes, I have contacted with G4Opticks, and appears that the "MultiUnion" ability may not be added in the near future, but probably they will add it later. I think their idea is to focus more on getting the high-performance simulation right, rather than generalizing the input geometry.

Using mesh from FreeCAD is nice but not sure it feeds into a G4Opticks, but I will give it a shot.

I totally understand it might be difficult to add it to the current version of GDML. I think I would need to write my own c++ input code for this set of problems or maybe write a wrapper to convert MultiUnion from FreeCAD/GDML output to a new GDML file.

Thank you very much!

[mhindi2]

Ami, Please see the response I just posted. Munther

…

On 7/27/22 15:54, amihashemi wrote: Hi Keith, Gmsh is useful for some FEM analysis at the current stage, but the major optical simulation with Geant4 initiates a ray tracing problem. In this context, the meshing of the structure means that proxy objects are constructed based on the set of indexed triangles that photon ray tracing will be performed on them. The mesh itself is transferable to computational threads for high-performance computing as the number of generated photons is substantially high (you may check this https://developer.nvidia.com/blog/creating-optimal-meshes-for-ray-tracing/). There are many reasons for the current limitations, but at this point, we don't have the tool in Geant4 to convert the MultiUnion arrays. Combining Gmsh mesh with Geant4 is a kind of far fetch idea to me but it is possible. — Reply to this email directly, view it on GitHub <#93 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AWV3NCH4GDBV2BHK7Z73IETVWG42BANCNFSM5YLCMEYA>. You are receiving this because you were mentioned.Message ID: ***@***.***>

[mhindi2]

Ok, I think I've got it working. Am not sure I can attach the files to this email. I will also send them to your gmail. Munther

…

On 7/27/22 16:10, amihashemi wrote: Hi Munther, Thank you very much for the response. Yes, I have contacted with G4Opticks, and appears that the "MultiUnion" ability may not be added in the near future, but probably they will add it later. I think their idea is to focus more on getting the high-performance simulation right, rather than generalizing the input geometry. Using mesh from FreeCAD is nice but not sure it feeds into a G4Opticks, but I will give it a shot. I totally understand it might be difficult to add it to the current version of GDML. I think I would need to write my own c++ input code for this set of problems or maybe write a wrapper to convert MultiUnion from FreeCAD/GDML output to a new GDML file. Thank you very much! — Reply to this email directly, view it on GitHub <#93 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AWV3NCGDP7O4TPCMPJ2GLLDVWG6X5ANCNFSM5YLCMEYA>. You are receiving this because you were mentioned.Message ID: ***@***.***>

[porosev]

Some comments about #93..
...

Evaluating the CaTS example \Geant4-11.0.2\examples\advanced\CaTS\gdml\simpleLArTPC.gdml I have found that they keep all optical properties inside of GDML file. Anyway it has some drawback.... in some situations it requires many hand-work, because you should define wavelength dependent reflectivity, refraction, emission spectra of scintillators etc...

In my simulations (with using standard Box, Tubes etc.. solids to define a setup geometry) I try to use the following approach:

I define detector as:
...
class DetectorConstruction : public G4VUserDetectorConstruction, public DetectorSurfaces
...

this additional service class "DetectorSurfaces" is used for surfaces definitions from hand-typed data and reading from files:
/////////////////////////////////// Surfaces_h /////////////////////////////////////////////////
#ifndef Surfaces_h
#define Surfaces_h 1

#include <globals.hh>

class G4OpticalSurface;

class DetectorSurfaces
{
protected:
// standard surfaces

G4OpticalSurface *polishedAir;              // polished crystal surface, not wrapped
G4OpticalSurface* polishedWhitePainted;     // polished crystal surface painted white
G4OpticalSurface *PMTCathodeMaterial;       // surface between window and cathode

public:
DetectorSurfaces();
};

#endif
/////////////////////////////////// Surfaces_cc /////////////////////////////////////////////////
...
DetectorSurfaces::DetectorSurfaces()
{
polishedAir = NULL;
polishedWhitePainted = NULL;
PMTCathodeMaterial = NULL;

//Get ENVIROMENT VARIABLE TO MATERIAL DATA
char* path = getenv("MY_G4DATA");
if (!path)
{
G4Exception("DetectorConstruction::defineSurfaces()", "", FatalException, "MY_G4DATA environment variable not set.");
return;
}

////////////////////////////////////////////////////////////////////
//
polishedAir = new G4OpticalSurface("polishedAir", unified);
polishedAir->SetType(dielectric_dielectric);
polishedAir->SetModel(unified);
polishedAir->SetFinish(ground); // necessary even for polished surfaces to enable UNIFIED code
polishedAir->SetSigmaAlpha(1.3 * degree); // Janecek2010

//////////////////////////////////////////////////////////////////
//
// COMMON PROPERTIES
G4double ener[2] = { .1*eV, 10.*eV };
G4double paintspecularlobe[2] = { 1.0, 1.0 };
G4double paintspecularspike[2] = { 0.0, 0.0 };
G4double paintbackscatter[2] = { 0.0, 0.0 };

//////////////////////////////////////////////////////////////////
// EJ-510
// https://eljentechnology.com/products/accessories/ej-510-ej-520
//
G4double paintrindex[2] = { 1.62, 1.62 };

polishedWhitePainted = new G4OpticalSurface("polishedWhitePainted");
polishedWhitePainted->SetType(dielectric_dielectric);
polishedWhitePainted->SetModel(unified);
polishedWhitePainted->SetFinish(polishedbackpainted);
polishedWhitePainted->SetSigmaAlpha(1.3 * degree);

G4MaterialPropertiesTable* polishedWhitePaintedProperty = new G4MaterialPropertiesTable();
polishedWhitePaintedProperty->AddProperty("RINDEX", ener, paintrindex, 2);
polishedWhitePaintedProperty->AddProperty("SPECULARLOBECONSTANT", ener, paintspecularlobe, 2);
polishedWhitePaintedProperty->AddProperty("SPECULARSPIKECONSTANT", ener, paintspecularspike, 2);
polishedWhitePaintedProperty->AddProperty("BACKSCATTERCONSTANT", ener, paintbackscatter, 2);

std::vector dEJE;
std::vector dEJR;
OptService::ReadParameters(path, "EJ-510_Reflectance.txt", dEJE, dEJR, true, CLHEP::eV);
polishedWhitePaintedProperty->AddProperty("REFLECTIVITY", dEJE, dEJR);
polishedWhitePainted->SetMaterialPropertiesTable(polishedWhitePaintedProperty);

//////////////////////////////////////////////////////////////////
// R10233
PMTCathodeMaterial = new G4OpticalSurface("CathodeMaterial", unified);
PMTCathodeMaterial->SetType(dielectric_metal);
PMTCathodeMaterial->SetModel(unified);
PMTCathodeMaterial->SetFinish(polished);
PMTCathodeMaterial->SetSigmaAlpha(0.);
G4MaterialPropertiesTable *PMTCathodeMaterialProperty = new G4MaterialPropertiesTable();

std::vector Cathode_energy;
std::vector Cathode_data;

//I use home=made file reader for data in the following format: E(eV) Parameter(A.U.)
OptService::ReadParameters(path, "R10233_TrueReflectanceWithoutGlass.txt", Cathode_energy, Cathode_data, true, CLHEP::eV);
PMTCathodeMaterialProperty->AddProperty("REFLECTIVITY", Cathode_energy, Cathode_data);

std::vector vEngEff;
std::vector vEff;
OptService::ReadParameters(path, "R10233_QE.txt", vEngEff, vEff, true, CLHEP::eV);
PMTCathodeMaterialProperty->AddProperty("EFFICIENCY", vEngEff, vEff);

PMTCathodeMaterial->SetMaterialPropertiesTable(PMTCathodeMaterialProperty);
}

///////////////////////////////////////////////////////////////////////////////////////////////////

after that I can easily define any surface in my code (during detector construction) between any physical volumes:

G4VPhysicalVolume * DetectorConstruction::constructDetector()
{
....
//-------------------------------------------------
// setup surfaces
//-------------------------------------------------
new G4LogicalBorderSurface("scint2AirFaceBorderSurface", physiScint, physiWorld, polishedWhitePainted);

new G4LogicalBorderSurface("grease2AirFaceBorderSurface", physiGrease, physiWorld, polishedAir);

new G4LogicalBorderSurface("glass2CathodeBorderSurface", physiPMT, physiCathode, PMTCathodeMaterial);

return physiWorld;
}

So, I plan to test how it works with GDML because GDMLParser lets to get pointer to physical volume using its name...

class G4GDMLParser
{
....
GetPhysVolume(const G4String& name)
..
}

So, theoretically it should be possible.... for some projects ;))

[mhindi2]

@porosov If I understand correctly, what you are suggesting is an approach where the geometry is defined through FreeCAD/GDML, but the optical properties added in C++. This is a good approach, especially that currently FreeCAD/GDML does not provide any means of editing the Matrices that define the energy/wavelength tables. At present we assume these have to read in from a GDML file, or cut and paste from some external file. To the extent that one has to process detector response in C++ at any rate, I see no harm in your approach. One can concentrate in FreeCAD on constructing and visualizing the geometry, something that is not easy (for complicated geometries, at any rate) to do in C++. FreeCAD has some limited support for Spreadsheets. In the future I can envision an addition to the GDML Workbench that takes a spreadsheet containing the optical properties and adding it to the Optical properties. FreeCAD can import excel spreadsheets. Munther Hindi

…

On 8/8/22 10:22, porosev wrote: Some comments about #93 <#93>.. ... Evaluating the CaTS example \Geant4-11.0.2\examples\advanced\CaTS\gdml\simpleLArTPC.gdml I have found that they keep all optical properties inside of GDML file. Anyway it has some drawback.... in some situations it requires many hand-work, because you should define wavelength dependent reflectivity, refraction, emission spectra of scintillators etc... In my simulations (with using standard Box, Tubes etc.. solids to define a setup geometry) I try to use the following approach: I define detector as: ... class DetectorConstruction : public G4VUserDetectorConstruction, public DetectorSurfaces ... this additional service class "DetectorSurfaces" is used for surfaces definitions from hand-typed data and reading from files: /////////////////////////////////// Surfaces_h ///////////////////////////////////////////////// #ifndef Surfaces_h #define Surfaces_h 1 #include <globals.hh> class G4OpticalSurface; class DetectorSurfaces { protected: // standard surfaces |G4OpticalSurface *polishedAir; // polished crystal surface, not wrapped G4OpticalSurface* polishedWhitePainted; // polished crystal surface painted white G4OpticalSurface *PMTCathodeMaterial; // surface between window and cathode | public: DetectorSurfaces(); }; #endif /////////////////////////////////// Surfaces_cc ///////////////////////////////////////////////// ... DetectorSurfaces::DetectorSurfaces() { polishedAir = NULL; polishedWhitePainted = NULL; PMTCathodeMaterial = NULL; //Get ENVIROMENT VARIABLE TO MATERIAL DATA char* path = getenv("MY_G4DATA"); if (!path) { G4Exception("DetectorConstruction::defineSurfaces()", "", FatalException, "MY_G4DATA environment variable not set."); return; } //////////////////////////////////////////////////////////////////// // polishedAir = new G4OpticalSurface("polishedAir", unified); polishedAir->SetType(dielectric_dielectric); polishedAir->SetModel(unified); polishedAir->SetFinish(ground); // necessary even for polished surfaces to enable UNIFIED code polishedAir->SetSigmaAlpha(1.3 * degree); // Janecek2010 ////////////////////////////////////////////////////////////////// // // COMMON PROPERTIES G4double ener[2] = { .1*eV, 10.*eV }; G4double paintspecularlobe[2] = { 1.0, 1.0 }; G4double paintspecularspike[2] = { 0.0, 0.0 }; G4double paintbackscatter[2] = { 0.0, 0.0 }; ////////////////////////////////////////////////////////////////// // EJ-510 // https://eljentechnology.com/products/accessories/ej-510-ej-520 // G4double paintrindex[2] = { 1.62, 1.62 }; polishedWhitePainted = new G4OpticalSurface("polishedWhitePainted"); polishedWhitePainted->SetType(dielectric_dielectric); polishedWhitePainted->SetModel(unified); polishedWhitePainted->SetFinish(polishedbackpainted); polishedWhitePainted->SetSigmaAlpha(1.3 * degree); G4MaterialPropertiesTable* polishedWhitePaintedProperty = new G4MaterialPropertiesTable(); polishedWhitePaintedProperty->AddProperty("RINDEX", ener, paintrindex, 2); polishedWhitePaintedProperty->AddProperty("SPECULARLOBECONSTANT", ener, paintspecularlobe, 2); polishedWhitePaintedProperty->AddProperty("SPECULARSPIKECONSTANT", ener, paintspecularspike, 2); polishedWhitePaintedProperty->AddProperty("BACKSCATTERCONSTANT", ener, paintbackscatter, 2); std::vector dEJE; std::vector dEJR; OptService::ReadParameters(path, "EJ-510_Reflectance.txt", dEJE, dEJR, true, CLHEP::eV); polishedWhitePaintedProperty->AddProperty("REFLECTIVITY", dEJE, dEJR); polishedWhitePainted->SetMaterialPropertiesTable(polishedWhitePaintedProperty); ////////////////////////////////////////////////////////////////// // R10233 PMTCathodeMaterial = new G4OpticalSurface("CathodeMaterial", unified); PMTCathodeMaterial->SetType(dielectric_metal); PMTCathodeMaterial->SetModel(unified); PMTCathodeMaterial->SetFinish(polished); PMTCathodeMaterial->SetSigmaAlpha(0.); G4MaterialPropertiesTable *PMTCathodeMaterialProperty = new G4MaterialPropertiesTable(); std::vector Cathode_energy; std::vector Cathode_data; //I use home=made file reader for data in the following format: E(eV) Parameter(A.U.) OptService::ReadParameters(path, "R10233_TrueReflectanceWithoutGlass.txt", Cathode_energy, Cathode_data, true, CLHEP::eV); PMTCathodeMaterialProperty->AddProperty("REFLECTIVITY", Cathode_energy, Cathode_data); std::vector vEngEff; std::vector vEff; OptService::ReadParameters(path, "R10233_QE.txt", vEngEff, vEff, true, CLHEP::eV); PMTCathodeMaterialProperty->AddProperty("EFFICIENCY", vEngEff, vEff); PMTCathodeMaterial->SetMaterialPropertiesTable(PMTCathodeMaterialProperty); } /////////////////////////////////////////////////////////////////////////////////////////////////// after that I can easily define any surface in my code (during detector construction) between any physical volumes: G4VPhysicalVolume * DetectorConstruction::constructDetector() { .... //------------------------------------------------- // setup surfaces //------------------------------------------------- new G4LogicalBorderSurface("scint2AirFaceBorderSurface", physiScint, physiWorld, polishedWhitePainted); new G4LogicalBorderSurface("grease2AirFaceBorderSurface", physiGrease, physiWorld, polishedAir); new G4LogicalBorderSurface("glass2CathodeBorderSurface", physiPMT, physiCathode, PMTCathodeMaterial); return physiWorld; } So, I plan to test how it works with GDML because GDMLParser lets to get pointer to physical volume using its name... class G4GDMLParser { .... GetPhysVolume(const G4String& name) .. } So, theoretically it should be possible.... for some projects ;)) — Reply to this email directly, view it on GitHub <#93 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AWV3NCDNQORIPJLCLSZES23VYE667ANCNFSM5YLCMEYA>. You are receiving this because you were mentioned.Message ID: ***@***.***>

[amihashemi]

In my understanding optical properties can be imported into FreeCAD already (optical.xml file is one example file where previously used to import the optical properties) but not able to import the bordersurface, Keith may comment on this...
Ami

[KeithSloan]

That is correct you can define optical properties in an xml file. Importing xml files adds them to the current document/design.

[KeithSloan]

Okay I found there was a bug on import of bordersurface, also duplicate code in GDMLCommands.py, have updated branches Main and forAMI

[amihashemi]

Great, thank you very much Keith.
Would it be possible to add copynumber on the to the current version of GDML? I remember in one of the previous versions, we had this copynumber assignment. It would be great to add it to the current version, for example in definition of array, physicalvolume would enumerate with the same order as their located in the structure tree. Here is an example where crystal indices where assigned for crystals (https://github.com/Geant4/geant4/blob/master/examples/advanced/doiPET/src/doiPETDetectorConstruction.cc) if you search crystalIndex you would find in G4PVPlacement.

Thank you!

[mhindi2]

I am playing around with converting the arrays to a tessellation. It is a two-step process, first the array needs to be converted to a mesh (the Create Mesh from Shape icon) and then convert the mesh to a Tessellation (the Mesh 2 Tess icon). The material is NOT automatically copied, so it has to be set AFTER the conversion. I am still having some problems with assigning the materials correctly. Once I resolve that I will send you the files to try with G4Opticks. Munther

…

On 7/27/22 16:10, amihashemi wrote: Hi Munther, Thank you very much for the response. Yes, I have contacted with G4Opticks, and appears that the "MultiUnion" ability may not be added in the near future, but probably they will add it later. I think their idea is to focus more on getting the high-performance simulation right, rather than generalizing the input geometry. Using mesh from FreeCAD is nice but not sure it feeds into a G4Opticks, but I will give it a shot. I totally understand it might be difficult to add it to the current version of GDML. I think I would need to write my own c++ input code for this set of problems or maybe write a wrapper to convert MultiUnion from FreeCAD/GDML output to a new GDML file. Thank you very much! — Reply to this email directly, view it on GitHub <#93 (reply in thread)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AWV3NCGDP7O4TPCMPJ2GLLDVWG6X5ANCNFSM5YLCMEYA>. You are receiving this because you were mentioned.Message ID: ***@***.***>

[KeithSloan]

There are ways to Tessellate in one step that also copy the material see first two options on this wiki page https://github.com/KeithSloan/GDML/wiki/Step2Tessellate

Using Gmsh option should allow fine control over the mesh.

@porosev Has experimented with using different options seehttps://github.com/KeithSloan/GDML/issues/97

[amihashemi]

Hi Munther and Keith, I look forward to checking the GDML with tessellation. It will be a great step if that works for optical properties. G4Opticks has recently been updated to G4CXOpticks and some of the structure definitions in GDML have been added to the new version. I will check with the new and old versions. Thank you very much, Best regards, Ami On Tue, Oct 11, 2022 at 3:57 AM Munther Hindi ***@***.***> wrote:

…

I am playing around with converting the arrays to a tessellation. It is a two-step process, first the array needs to be converted to a mesh (the Create Mesh from Shape icon) and then convert the mesh to a Tessellation (the Mesh 2 Tess icon). The material is NOT automatically copied, so it has to be set AFTER the conversion. I am still having some problems with assigning the materials correctly. Once I resolve that I will send you the files to try with G4Opticks. Munther On 7/27/22 16:10, amihashemi wrote: > > Hi Munther, > > Thank you very much for the response. > > Yes, I have contacted with G4Opticks, and appears that the > "MultiUnion" ability may not be added in the near future, but probably > they will add it later. I think their idea is to focus more on getting > the high-performance simulation right, rather than generalizing the > input geometry. > > Using mesh from FreeCAD is nice but not sure it feeds into a > G4Opticks, but I will give it a shot. > > I totally understand it might be difficult to add it to the current > version of GDML. I think I would need to write my own c++ input code > for this set of problems or maybe write a wrapper to convert > MultiUnion from FreeCAD/GDML output to a new GDML file. > > Thank you very much! > > — > Reply to this email directly, view it on GitHub > < #93 (reply in thread)>, > or unsubscribe > < https://github.com/notifications/unsubscribe-auth/AWV3NCGDP7O4TPCMPJ2GLLDVWG6X5ANCNFSM5YLCMEYA >. > You are receiving this because you were mentioned.Message ID: > ***@***.***> > — Reply to this email directly, view it on GitHub <#93 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AH6QDXT4OCJBGIPCCGH6NO3WCUMX3ANCNFSM5YLCMEYA> . You are receiving this because you were mentioned.Message ID: ***@***.***>