Re: [Scilab-users] ipcv vs scicv
Hi all,
Stephane, Thanks for the quick examples to illustrate this.
Claus, as shown by Stephane, this could be now by ourselves now from our
own modules. Stephane has illustrated a quick demo on this, while the one I
was testing with is the copy of toolbox skeleton inside the Scilab contrib,
modified to 2 modules, installed with atoms and let it load at start,
results are the same.
There are 2 potential issues which we should take notes:
1. The module which is loaded last, will have the "dominant" in the base
function. for e.g.: the example that Stephane showed, we could call
"lib1.foo" and "lib2.foo", and calling just "foo" will be the same as
"lib2.foo".
2. Only macros function could be called this way, if a module has a scilab
gateway that called direct from Scilab, (such as scicv), you could not call
scicvlib.imread as the imread is directly expose from the C lib to Scilab.
So during the design of the module, you could have a macros wrapper to call
the gateway functions so that they could be listed in the library's macros
list.
Thanks.
rgds,
CL
On Fri, 4 Dec 2020 at 00:36, Stéphane Mottelet
wrote:
> This is the way it works in scilab for libraries:
>
> mkdir lib1mputl("function
> y=foo(x);y=x;end","lib1/foo.sci")genlib("lib1","lib1")mkdir
> lib2mputl("function
> y=foo(x);y=2*x;end","lib2/foo.sci")genlib("lib2","lib2")lib("lib1")lib("lib2")
> -> lib1.foo(1)
> ans =
>
>1.
> --> lib2.foo(1)
> ans =
>
>2.
>
>
> S.
>
> Le 03/12/2020 à 17:28, Antoine Monmayrant a écrit :
>
>
> On 03/12/2020 02:10, TanCL wrote:
>
> Hi, my previous message seems like not able to get posted due to some reason,
> just to add some comment on this, I think the functions with same name under
> 2 different modules could be possibly called by:
>
> That is exactly how things are working in Ada:
>
> myfunc : local definition in the current file
> somepackage.myfunc : definition one can find in somepackage that is
> included in the current file by "with somepackage;"
>
> And one can rename the package with "package S renames somepackage" which
> allow writting "s.myfunc"
> I assume it's similar in Python too.
>
> Antoine
>
> moduleA.myfunc
>
> moduleB.myfunc
>
>
> __
>
> Start Toolbox A
> Load macros
>
> Start Toolbox B
> Load macros
>
>
> --> toolboxAlib.scilab_sum(2,1)
>
> "This is function from ToolboxA"
> ans =
>
>3.
>
> --> toolboxBlib.scilab_sum(2,1)
>
> "This is function from ToolboxB"
> ans =
>
>3.
>
>
>
>
>
> --
> Sent from:
> http://mailinglists.scilab.org/Scilab-users-Mailing-Lists-Archives-f2602246.html
>
>
> ___
> users mailing
> listusers@lists.scilab.orghttp://lists.scilab.org/mailman/listinfo/users
>
>
>
> ___
> users mailing
> listusers@lists.scilab.orghttps://antispam.utc.fr/proxy/1/c3RlcGhhbmUubW90dGVsZXRAdXRjLmZy/lists.scilab.org/mailman/listinfo/users
>
> --
> Stéphane Mottelet
> Ingénieur de recherche
> EA 4297 Transformations Intégrées de la Matière Renouvelable
> Département Génie des Procédés Industriels
> Sorbonne Universités - Université de Technologie de Compiègne
> CS 60319, 60203 Compiègne cedex
> Tel : +33(0)344234688http://www.utc.fr/~mottelet
>
> ___
> users mailing list
> users@lists.scilab.org
> http://lists.scilab.org/mailman/listinfo/users
>
___
users mailing list
users@lists.scilab.org
http://lists.scilab.org/mailman/listinfo/users
Re: [Scilab-users] ipcv vs scicv
This is the way it works in scilab for libraries:
mkdir lib1
mputl("function y=foo(x);y=x;end","lib1/foo.sci")
genlib("lib1","lib1")
mkdir lib2
mputl("function y=foo(x);y=2*x;end","lib2/foo.sci")
genlib("lib2","lib2")
lib("lib1")
lib("lib2")
-> lib1.foo(1)
ans =
1.
--> lib2.foo(1)
ans =
2.
S.
Le 03/12/2020 à 17:28, Antoine Monmayrant a écrit :
On 03/12/2020 02:10, TanCL wrote:
Hi, my previous message seems like not able to get posted due to some reason,
just to add some comment on this, I think the functions with same name under
2 different modules could be possibly called by:
That is exactly how things are working in Ada:
myfunc : local definition in the current file
somepackage.myfunc : definition one can find in somepackage that
is included in the current file by "with somepackage;"
And one can rename the package with "package S renames somepackage"
which allow writting "s.myfunc"
I assume it's similar in Python too.
Antoine
moduleA.myfunc
moduleB.myfunc
__
Start Toolbox A
Load macros
Start Toolbox B
Load macros
--> toolboxAlib.scilab_sum(2,1)
"This is function from ToolboxA"
ans =
3.
--> toolboxBlib.scilab_sum(2,1)
"This is function from ToolboxB"
ans =
3.
--
Sent
from:http://mailinglists.scilab.org/Scilab-users-Mailing-Lists-Archives-f2602246.html
___
users mailing list
users@lists.scilab.org
http://lists.scilab.org/mailman/listinfo/users
___
users mailing list
users@lists.scilab.org
https://antispam.utc.fr/proxy/1/c3RlcGhhbmUubW90dGVsZXRAdXRjLmZy/lists.scilab.org/mailman/listinfo/users
--
Stéphane Mottelet
Ingénieur de recherche
EA 4297 Transformations Intégrées de la Matière Renouvelable
Département Génie des Procédés Industriels
Sorbonne Universités - Université de Technologie de Compiègne
CS 60319, 60203 Compiègne cedex
Tel : +33(0)344234688
http://www.utc.fr/~mottelet
___
users mailing list
users@lists.scilab.org
http://lists.scilab.org/mailman/listinfo/users
Re: [Scilab-users] ipcv vs scicv
On 03/12/2020 02:10, TanCL wrote: Hi, my previous message seems like not able to get posted due to some reason, just to add some comment on this, I think the functions with same name under 2 different modules could be possibly called by: That is exactly how things are working in Ada: myfunc : local definition in the current file somepackage.myfunc : definition one can find in somepackage that is included in the current file by "with somepackage;" And one can rename the package with "package S renames somepackage" which allow writting "s.myfunc" I assume it's similar in Python too. Antoine moduleA.myfunc moduleB.myfunc __ Start Toolbox A Load macros Start Toolbox B Load macros --> toolboxAlib.scilab_sum(2,1) "This is function from ToolboxA" ans = 3. --> toolboxBlib.scilab_sum(2,1) "This is function from ToolboxB" ans = 3. -- Sent from: http://mailinglists.scilab.org/Scilab-users-Mailing-Lists-Archives-f2602246.html ___ users mailing list users@lists.scilab.org http://lists.scilab.org/mailman/listinfo/users ___ users mailing list users@lists.scilab.org http://lists.scilab.org/mailman/listinfo/users
Re: [Scilab-users] ipcv vs scicv
Hi Tan I agree. The proposal, is it something we can do ourselves now, or will it require implementation into Scilab first? Thank you for resending. /Claus On 03-12-2020 02:10, TanCL wrote: Hi, my previous message seems like not able to get posted due to some reason, just to add some comment on this, I think the functions with same name under 2 different modules could be possibly called by: moduleA.myfunc moduleB.myfunc __ Start Toolbox A Load macros Start Toolbox B Load macros --> toolboxAlib.scilab_sum(2,1) "This is function from ToolboxA" ans = 3. --> toolboxBlib.scilab_sum(2,1) "This is function from ToolboxB" ans = 3. -- Sent from: http://mailinglists.scilab.org/Scilab-users-Mailing-Lists-Archives-f2602246.html ___ users mailing list users@lists.scilab.org http://lists.scilab.org/mailman/listinfo/users ___ users mailing list users@lists.scilab.org http://lists.scilab.org/mailman/listinfo/users
Re: [Scilab-users] Headless graphics in scilab: time & memory issues
The xmllab is actually still working, try e.g. http://marguerite.lmac.utc.fr/~xmllab/cgi-bin/XMLlabServer.cgi?timing=no=english=/home/xmllab/scilab-5.5.2/share/scilab/contrib/xmllab/1.7.6//examples/Image-Processing/clown.xml=launch S. Le 03/12/2020 à 11:20, Stéphane Mottelet a écrit : Le 03/12/2020 à 11:01, Antoine Monmayrant a écrit : On 03/12/2020 10:46, Stéphane Mottelet wrote: In headless mode the rendering cannot use joGL and there is the same overhead as when you try to export a figure to a vectorial format (svg,pdf,eps) : OK thanks for the clarification. That's a weird and unexpected limitation, no? No. In fact fast offscreen rendering is possible in joGL. In my previous answer I should have said "In headless mode the rendering *does not* use joGL". The pity is that the main author of Scilab Graphics Java backend has left the team long time ago. It would have been an easy task for him. If your project is hosted on a Linux server, I would advise you to redirect the display of your Scilab instance to a virtual frame buffer by using any flavor of VNC (TurboVNC is nice as the server can use the hardware GPU capabilities of the server). In this case you can stick to the Rec driver and j xs2png (or any bitmap variant) will be fast. That's what I used with xmllab many years ago... S. I would not have guessed that going headless and using a bitmap format would result in facing this limitation. Moreover, there is still a memory issue as I can crash scilab when pushing things a bit further. Antoine ___ users mailing list users@lists.scilab.org https://antispam.utc.fr/proxy/2/c3RlcGhhbmUubW90dGVsZXRAdXRjLmZy/antispam.utc.fr/proxy/1/c3RlcGhhbmUubW90dGVsZXRAdXRjLmZy/lists.scilab.org/mailman/listinfo/users -- Stéphane Mottelet Ingénieur de recherche EA 4297 Transformations Intégrées de la Matière Renouvelable Département Génie des Procédés Industriels Sorbonne Universités - Université de Technologie de Compiègne CS 60319, 60203 Compiègne cedex Tel : +33(0)344234688 http://www.utc.fr/~mottelet ___ users mailing list users@lists.scilab.org http://lists.scilab.org/mailman/listinfo/users
Re: [Scilab-users] Headless graphics in scilab: time & memory issues
Le 03/12/2020 à 11:01, Antoine Monmayrant a écrit : On 03/12/2020 10:46, Stéphane Mottelet wrote: In headless mode the rendering cannot use joGL and there is the same overhead as when you try to export a figure to a vectorial format (svg,pdf,eps) : OK thanks for the clarification. That's a weird and unexpected limitation, no? No. In fact fast offscreen rendering is possible in joGL. In my previous answer I should have said "In headless mode the rendering *does not* use joGL". The pity is that the main author of Scilab Graphics Java backend has left the team long time ago. It would have been an easy task for him. If your project is hosted on a Linux server, I would advise you to redirect the display of your Scilab instance to a virtual frame buffer by using any flavor of VNC (TurboVNC is nice as the server can use the hardware GPU capabilities of the server). In this case you can stick to the Rec driver and j xs2png (or any bitmap variant) will be fast. That's what I used with xmllab many years ago... S. I would not have guessed that going headless and using a bitmap format would result in facing this limitation. Moreover, there is still a memory issue as I can crash scilab when pushing things a bit further. Antoine ___ users mailing list users@lists.scilab.org https://antispam.utc.fr/proxy/1/c3RlcGhhbmUubW90dGVsZXRAdXRjLmZy/lists.scilab.org/mailman/listinfo/users -- Stéphane Mottelet Ingénieur de recherche EA 4297 Transformations Intégrées de la Matière Renouvelable Département Génie des Procédés Industriels Sorbonne Universités - Université de Technologie de Compiègne CS 60319, 60203 Compiègne cedex Tel : +33(0)344234688 http://www.utc.fr/~mottelet ___ users mailing list users@lists.scilab.org http://lists.scilab.org/mailman/listinfo/users
Re: [Scilab-users] Headless graphics in scilab: time & memory issues
Hello Antoine, Exporting in headless mode (all drivers but the default "Rec") is fundamentally different from the default "Rec" mode where graphics are actually rendered on screen on a joGL Canvas. In this case there is virtually no overhead when exporting this Canvas to a bitmap image format (png,jpeg,gif,...). In headless mode the rendering cannot use joGL and there is the same overhead as when you try to export a figure to a vectorial format (svg,pdf,eps) : nx=4000; ny=300; nc=100; x=1:nx; y=[1:ny].'; z=(y*x); driver Rec h=scf(); h.color_map=hotcolormap(nc); Sgrayplot(x,y,z.'); xs2pdf(h,"screen.pdf") at line 13 of executed file /Users/mottelet/Desktop/xend_test.sce xs2pdf: Unable to create export file, not enough memory. Decreasing the number of elements or the size of the figure should fix this error. S. Le 03/12/2020 à 09:48, Antoine Monmayrant a écrit : Hello all, In these days of remote working, I need to use again scilab in headless mode to generate graphics. In other words, I connect in command line only to a remote server where scilab is installed and I want it to generate graphics using a combination of driver, xinit and xend. There seems to be some issues with driver/xinit/xend: huge memory overhead and long computation time that makes it impossible to generate some plots in headless mode. For example, a simple Sgrayplot that takes ~1 second to plot and save to a png in normal mode takes more than 200 seconds and huge amount of ram in headless mode (see attached test script). I also have some plots that works without a problem in normal mode, but fail in headless mode because of some memory limitation: xend: An error occurred: Unable to create export file, not enough memory. Decreasing the number of elements or the size of the figure should fix this error. Is there a way to get around these problems? Thanks for your help, Antoine ___ users mailing list users@lists.scilab.org https://antispam.utc.fr/proxy/1/c3RlcGhhbmUubW90dGVsZXRAdXRjLmZy/lists.scilab.org/mailman/listinfo/users -- Stéphane Mottelet Ingénieur de recherche EA 4297 Transformations Intégrées de la Matière Renouvelable Département Génie des Procédés Industriels Sorbonne Universités - Université de Technologie de Compiègne CS 60319, 60203 Compiègne cedex Tel : +33(0)344234688 http://www.utc.fr/~mottelet ___ users mailing list users@lists.scilab.org http://lists.scilab.org/mailman/listinfo/users
Re: [Scilab-users] {EXT} Re: Headless graphics in scilab: time & memory issues
Hello Antoine, > De : Antoine Monmayrant > Envoyé : jeudi 3 décembre 2020 10:00 > > Can any of you confirm this bug? I don't know what you mean by headless mode but I get the same error as you error when I run your script (except it is in Franch): "xs2png : Impossible de créer le fichier d'export, pas assez de mémoire disponible. Diminuer le nombre d'éléments ou la taille de la figure devrait corriger cette erreur." Regards -- Christophe Dang Ngoc Chan Mechanical calculation engineer General This e-mail may contain confidential and/or privileged information. If you are not the intended recipient (or have received this e-mail in error), please notify the sender immediately and destroy this e-mail. Any unauthorized copying, disclosure or distribution of the material in this e-mail is strictly forbidden. ___ users mailing list users@lists.scilab.org http://lists.scilab.org/mailman/listinfo/users
Re: [Scilab-users] Headless graphics in scilab: time & memory issues
Hello again, Just to add to my previous message: increasing nx or ny in my script can crash scilab during the call to xend(). On my small local machine, this occurs for: nx=4000; ny=3000; for which scilab takes ~7 seconds to create the png in "normal" mode... Can any of you confirm this bug? Antoine On 03/12/2020 09:48, Antoine Monmayrant wrote: Hello all, In these days of remote working, I need to use again scilab in headless mode to generate graphics. In other words, I connect in command line only to a remote server where scilab is installed and I want it to generate graphics using a combination of driver, xinit and xend. There seems to be some issues with driver/xinit/xend: huge memory overhead and long computation time that makes it impossible to generate some plots in headless mode. For example, a simple Sgrayplot that takes ~1 second to plot and save to a png in normal mode takes more than 200 seconds and huge amount of ram in headless mode (see attached test script). I also have some plots that works without a problem in normal mode, but fail in headless mode because of some memory limitation: xend: An error occurred: Unable to create export file, not enough memory. Decreasing the number of elements or the size of the figure should fix this error. Is there a way to get around these problems? Thanks for your help, Antoine ___ users mailing list users@lists.scilab.org http://lists.scilab.org/mailman/listinfo/users ___ users mailing list users@lists.scilab.org http://lists.scilab.org/mailman/listinfo/users
[Scilab-users] Headless graphics in scilab: time & memory issues
Hello all, In these days of remote working, I need to use again scilab in headless mode to generate graphics. In other words, I connect in command line only to a remote server where scilab is installed and I want it to generate graphics using a combination of driver, xinit and xend. There seems to be some issues with driver/xinit/xend: huge memory overhead and long computation time that makes it impossible to generate some plots in headless mode. For example, a simple Sgrayplot that takes ~1 second to plot and save to a png in normal mode takes more than 200 seconds and huge amount of ram in headless mode (see attached test script). I also have some plots that works without a problem in normal mode, but fail in headless mode because of some memory limitation: xend: An error occurred: Unable to create export file, not enough memory. Decreasing the number of elements or the size of the figure should fix this error. Is there a way to get around these problems? Thanks for your help, Antoine xend_test.sce Description: application/scilab-sce ___ users mailing list users@lists.scilab.org http://lists.scilab.org/mailman/listinfo/users
