sed-ml-discuss — Migrating to google groups: https://groups.google.com/d/forum/sed-ml-discuss

Re: [SED-ML-discuss] Soliciting ideas for HARMONY topics

[Dagmar W. <dag...@un...>]

Dear all,

is anyone of you planning to attend HARMONY this year *and* interested
in a discussion of the proposal to extend SED-ML to cover SBML Qual models?

https://dx.doi.org/10.6084/m9.figshare.2075143.v1

Do you have any other suggestions what to work on during HARMONY wrt
SED-ML?

Dagmar



On 03/20/2016 11:29 PM, David Nickerson wrote:
> Hi all,
> 
> HARMONY 2016 is coming up (and registration is open:
> http://co.mbine.org/events/HARMONY_2016) and we are starting to
> collect topics for possible SED-ML related breakouts/sessions during
> the meeting.
> 
> If you have any ideas for topics or burning issues that you would like
> to discuss, please let me know or raise them in on this list.
> 
> 
> Cheers,
> David.
> 

Re: [SED-ML-discuss] Exists a tool which can handle the L1V2 specification example?

[Matthias K. <kon...@go...>]

Hi Ilya,
thanks a lot for the workflow.
I could get everything running and create the plots with the fixed SEDML
archive.

In addition, the fixed example can also be reproduced with SEDML WebTools
(the max function also created problems here).
See figures attached.

@all
Is there any other tool besides tellurium, SEDML Webtools & BioUML which is
able to create the outputs?

Matthias


On Sun, Mar 27, 2016 at 9:17 PM, Ilya Kiselev <ax...@de...
> wrote:

>
> Would it make sense to restrict xpath expressions so that we don't run
>> into these sorts of problems?  This is only going to get worse, I think.
>> We don't even have recommended guidelines for how to construct XPaths.
>>
>
> I tend to agree with David on this. If we use standard (XPath) then it is
> tool responsibility to support it. if different tools will support
> different subsets (e.g. some tools will consider only to support
> recommeneded subset) - this will result in lack of interoperability.
>
>
>>
>> 1. curves must have id
>>>
>>
>> Why do curves have to have an id?  Literally nothing in SED-ML can
>> reference them, right?  Can we drop this requirement and make them optional
>> for the next level/version?
>>
>
> You are right.
> Currently we use ids to display to user in legend. Probably "name"
> attribute will suit better although it is optional and not unique so we
> will be forced to autogenerate something.
>
> I can see that Matthias use name for data generator in yDataReference
> which may cofuse user because different curves can have different
> xDataReferences (as in his plot_2).
>
>
>> 2. *max *functions should be used as special csymbol:
>>>     <apply>
>>>         <csymbol encoding="text" definitionURL="http://sed-ml.org/#max">
>>>             max
>>>         </csymbol>
>>>         <ci> task1_____PZ </ci>
>>>     </apply>
>>>
>>
>> Is there software that will parse 'max' in infix and create this output?
>> Or something that will take libsbml ASTNodes and convert them to this?
>>
>
> BioUML currently uses jlibsedml to read and write sedml. It produces
> correct form for max function, but can not read this shortened version.
>
> I am not sure what do you mean here. I think it is easy to write simple
> script which will take sedml and fix these functions.
>
> --
> Best regards, Ilya
>
>
> <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=oa-2115-v2-a> This
> email has been sent from a virus-free computer protected by Avast.
> www.avast.com
> <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=oa-2115-v2-a>
> <#m_-4851681771776372597_DDB4FAA8-2DD7-40BB-A1B8-4E2AA1F9FDF2>
>
>
> ------------------------------------------------------------------------------
> Transform Data into Opportunity.
> Accelerate data analysis in your applications with
> Intel Data Analytics Acceleration Library.
> Click to learn more.
> http://pubads.g.doubleclick.net/gampad/clk?id=278785351&iu=/4140
> _______________________________________________
> SED-ML-discuss mailing list
> SED...@li...
> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>
>


-- 
Matthias König
Junior Group Leader LiSym - Systems Medicine of the Liver
Humboldt-University Berlin, Institute for Theoretical Biology
  https://www.livermetabolism.com
kon...@go...
Tel: +49 30 20938450
Tel: +49 176 81168480

Re: [SED-ML-discuss] Exists a tool which can handle the L1V2 specification example?

[Ilya K. <ax...@de...>]

> Would it make sense to restrict xpath expressions so that we don't run
> into these sorts of problems?  This is only going to get worse, I think.
> We don't even have recommended guidelines for how to construct XPaths.
>

I tend to agree with David on this. If we use standard (XPath) then it is
tool responsibility to support it. if different tools will support
different subsets (e.g. some tools will consider only to support
recommeneded subset) - this will result in lack of interoperability.


>
> 1. curves must have id
>>
>
> Why do curves have to have an id?  Literally nothing in SED-ML can
> reference them, right?  Can we drop this requirement and make them optional
> for the next level/version?
>

You are right.
Currently we use ids to display to user in legend. Probably "name"
attribute will suit better although it is optional and not unique so we
will be forced to autogenerate something.

I can see that Matthias use name for data generator in yDataReference which
may cofuse user because different curves can have different xDataReferences
(as in his plot_2).


> 2. *max *functions should be used as special csymbol:
>>     <apply>
>>         <csymbol encoding="text" definitionURL="http://sed-ml.org/#max">
>>             max
>>         </csymbol>
>>         <ci> task1_____PZ </ci>
>>     </apply>
>>
>
> Is there software that will parse 'max' in infix and create this output?
> Or something that will take libsbml ASTNodes and convert them to this?
>

BioUML currently uses jlibsedml to read and write sedml. It produces
correct form for max function, but can not read this shortened version.

I am not sure what do you mean here. I think it is easy to write simple
script which will take sedml and fix these functions.

-- 
Best regards, Ilya

<https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=oa-2115-v2-a>
This
email has been sent from a virus-free computer protected by Avast.
www.avast.com
<https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail&utm_term=oa-2115-v2-a>
<#DDB4FAA8-2DD7-40BB-A1B8-4E2AA1F9FDF2>

Re: [SED-ML-discuss] Exists a tool which can handle the L1V2 specification example?

[Ilya K. <ax...@de...>]

Hi Matthias.

First of all you will need to download BioUML 0.9.10 again and reinstall it
(sorry about that).

I've made a little guide
<https://docs.google.com/document/d/1g4FvsetAE8Ui95ZCjftkJlDko20fCC_-UMQbrL7xf6M/edit?usp=sharing>
on how to import sedml and run it in BioUML.
Currently BioUML can not export .sedx archive, only .sedml.

Hope it will help. If you'll have any problems, please ask me.

Best regards, Ilya.

Re: [SED-ML-discuss] Exists a tool which can handle the L1V2 specification example?

[Jonathan C. <jon...@cs...>]

+1 from me. Modellers shouldn't have to see any details of SED-ML at 
all. And there are indeed XPath libraries that tools can use. Given that 
SED-ML is an XML-based standard, any tool supporting it ought to be able 
to cope with arbitrary XML manipulations. Fixing on a restricted subset 
just means tools don't implement it properly, like with the original 
annotation support in SBML.

Best wishes,
Jonathan

On 24/03/2016 09:08, David Nickerson wrote:
> You are completely right Matthias. However, limitations in current 
> tool support should not force changes or restrictions on the 
> specification. In an ideal world, tools would provide the level of 
> abstraction that you are asking for and should be provided, but we are 
> currently a long way short of that. Until the tools catch up to the 
> user requirements, then we're in a position where we need to have some 
> "best practices" in order to be able to successfully exchange 
> reproducible simulation experiments between tools, but I don't see 
> these belonging in the specification.
>
> These best practices really belong in libsedml - I wouldn't suggest 
> putting them into libsbml as they really are generic manipulations of 
> the XML structure of a document and have nothing to do with SBML, 
> CellML, or whatever other format your model is encoded in.
>
>
> Cheers,
> David.
>
>
>
> On Thu, Mar 24, 2016 at 9:49 PM, Matthias König 
> <kon...@go... <mailto:kon...@go...>> wrote:
>
>     Hi David,
>
>     I agree xpath works. But there should be some abstraction layer
>     around it either by libsbml or libsedml (and for the XMLChanges).
>     See
>     https://www.pivotaltracker.com/n/projects/248655/stories/114704823
>
>     I personally feel very uncomfortable working with the raw XML (I
>     am not a computer scientist, I am a computional biologist).
>     The core idea of libsbml and libsedml is to provide an abstraction
>     layer around the XML which provides a simple way to work with the
>     model/model changes.
>
>     Xpath expressions & the XMLChanges are the only cases where I
>     currently have to really manipulate/work with the XML, i.e. xml
>     parser, xpath parser, changing the dom. All things which are
>     highly error prone and should be abstracted away by the respective
>     libraries. Not providing support for this by libsbml or libsedml
>     will cause a multitude of problems in the long run because none of
>     the xpath and xmlChanges will be standardized. So you suddenly
>     have a standard which allows you to do basically everything
>     (generic XMLChanges, generic xpath expressions) but most of the
>     tools only supporting a very limited subset (only resolving
>     certain xpath expressions, not checking if the XMLChanges are
>     allowed and result in valid models). This will break compatibility
>     on so many levels.
>     This will be the main reason why SEDML will not be reproducible
>     between different tools.
>
>     Matthias
>
>
>
>     On Thu, Mar 24, 2016 at 9:36 AM, David Nickerson
>     <dav...@gm... <mailto:dav...@gm...>> wrote:
>
>         > Would it make sense to restrict xpath expressions so that we don't run into
>         > these sorts of problems?  This is only going to get worse, I
>         think.  We
>         > don't even have recommended guidelines for how to construct
>         XPaths.
>         >
>         > I mean, I'd love to drop xpath use altogether, but I also
>         don't see a great
>         > alternative.  But maybe a subset would be the best of both
>         worlds.
>
>         I think XPath works well when we are dealing with models
>         encoded in
>         XML formats and tools implement support for XPath rather than
>         relying
>         on extracting key information from the XPath expression. I
>         don't think
>         it would be good to enforce a restricted subset of XPath in
>         the SED-ML
>         specification, and I also don't think we need recommended
>         guidelines
>         for constructing XPath expressions. XPath is a fairly well
>         established
>         standard which is well supported across most programming
>         languages so
>         it makes sense to simply use it.
>
>         It might be worth considering supporting additional
>         methods/schemes
>         for identifying entities in source models. Such as URIs or
>         annotation
>         terms.
>
>         Cheers,
>         David.
>
>         ------------------------------------------------------------------------------
>         Transform Data into Opportunity.
>         Accelerate data analysis in your applications with
>         Intel Data Analytics Acceleration Library.
>         Click to learn more.
>         http://pubads.g.doubleclick.net/gampad/clk?id=278785351&iu=/4140
>         _______________________________________________
>         SED-ML-discuss mailing list
>         SED...@li...
>         <mailto:SED...@li...>
>         https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>
>
>
>
>     -- 
>     Matthias König
>     Junior Group Leader LiSym - Systems Medicine of the Liver
>     Humboldt-University Berlin, Institute for Theoretical Biology
>     https://www.livermetabolism.com
>     kon...@go... <mailto:kon...@go...>
>     Tel: +49 30 20938450
>     Tel: +49 176 81168480 <tel:%2B49%20176%2081168480>
>
>     ------------------------------------------------------------------------------
>     Transform Data into Opportunity.
>     Accelerate data analysis in your applications with
>     Intel Data Analytics Acceleration Library.
>     Click to learn more.
>     http://pubads.g.doubleclick.net/gampad/clk?id=278785351&iu=/4140
>     _______________________________________________
>     SED-ML-discuss mailing list
>     SED...@li...
>     <mailto:SED...@li...>
>     https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>
>
>
>
> -- 
> David Nickerson on about.me
> 		
> David Nickerson
> about.me/david.nickerson
>
> <http://about.me/david.nickerson>
>
>
> ------------------------------------------------------------------------------
> Transform Data into Opportunity.
> Accelerate data analysis in your applications with
> Intel Data Analytics Acceleration Library.
> Click to learn more.
> http://pubads.g.doubleclick.net/gampad/clk?id=278785351&iu=/4140
>
>
> _______________________________________________
> SED-ML-discuss mailing list
> SED...@li...
> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss

Re: [SED-ML-discuss] Exists a tool which can handle the L1V2 specification example?

[David N. <dav...@gm...>]

You are completely right Matthias. However, limitations in current tool
support should not force changes or restrictions on the specification. In
an ideal world, tools would provide the level of abstraction that you are
asking for and should be provided, but we are currently a long way short of
that. Until the tools catch up to the user requirements, then we're in a
position where we need to have some "best practices" in order to be able to
successfully exchange reproducible simulation experiments between tools,
but I don't see these belonging in the specification.

These best practices really belong in libsedml - I wouldn't suggest putting
them into libsbml as they really are generic manipulations of the XML
structure of a document and have nothing to do with SBML, CellML, or
whatever other format your model is encoded in.


Cheers,
David.



On Thu, Mar 24, 2016 at 9:49 PM, Matthias König <kon...@go...>
wrote:

> Hi David,
>
> I agree xpath works. But there should be some abstraction layer around it
> either by libsbml or libsedml (and for the XMLChanges).
> See
> https://www.pivotaltracker.com/n/projects/248655/stories/114704823
>
> I personally feel very uncomfortable working with the raw XML (I am not a
> computer scientist, I am a computional biologist).
> The core idea of libsbml and libsedml is to provide an abstraction layer
> around the XML which provides a simple way to work with the model/model
> changes.
>
> Xpath expressions & the XMLChanges are the only cases where I currently
> have to really manipulate/work with the XML, i.e. xml parser, xpath parser,
> changing the dom. All things which are highly error prone and should be
> abstracted away by the respective libraries. Not providing support for this
> by libsbml or libsedml will cause a multitude of problems in the long run
> because none of the xpath and xmlChanges will be standardized. So you
> suddenly have a standard which allows you to do basically everything
> (generic XMLChanges, generic xpath expressions) but most of the tools only
> supporting a very limited subset (only resolving certain xpath expressions,
> not checking if the XMLChanges are allowed and result in valid models).
> This will break compatibility on so many levels.
> This will be the main reason why SEDML will not be reproducible between
> different tools.
>
> Matthias
>
>
> On Thu, Mar 24, 2016 at 9:36 AM, David Nickerson <
> dav...@gm...> wrote:
>
>> > Would it make sense to restrict xpath expressions so that we don't run
>> into
>> > these sorts of problems?  This is only going to get worse, I think.  We
>> > don't even have recommended guidelines for how to construct XPaths.
>> >
>> > I mean, I'd love to drop xpath use altogether, but I also don't see a
>> great
>> > alternative.  But maybe a subset would be the best of both worlds.
>>
>> I think XPath works well when we are dealing with models encoded in
>> XML formats and tools implement support for XPath rather than relying
>> on extracting key information from the XPath expression. I don't think
>> it would be good to enforce a restricted subset of XPath in the SED-ML
>> specification, and I also don't think we need recommended guidelines
>> for constructing XPath expressions. XPath is a fairly well established
>> standard which is well supported across most programming languages so
>> it makes sense to simply use it.
>>
>> It might be worth considering supporting additional methods/schemes
>> for identifying entities in source models. Such as URIs or annotation
>> terms.
>>
>> Cheers,
>> David.
>>
>>
>> ------------------------------------------------------------------------------
>> Transform Data into Opportunity.
>> Accelerate data analysis in your applications with
>> Intel Data Analytics Acceleration Library.
>> Click to learn more.
>> http://pubads.g.doubleclick.net/gampad/clk?id=278785351&iu=/4140
>> _______________________________________________
>> SED-ML-discuss mailing list
>> SED...@li...
>> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>>
>
>
>
> --
> Matthias König
> Junior Group Leader LiSym - Systems Medicine of the Liver
> Humboldt-University Berlin, Institute for Theoretical Biology
>   https://www.livermetabolism.com
> kon...@go...
> Tel: +49 30 20938450
> Tel: +49 176 81168480
>
>
> ------------------------------------------------------------------------------
> Transform Data into Opportunity.
> Accelerate data analysis in your applications with
> Intel Data Analytics Acceleration Library.
> Click to learn more.
> http://pubads.g.doubleclick.net/gampad/clk?id=278785351&iu=/4140
> _______________________________________________
> SED-ML-discuss mailing list
> SED...@li...
> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>
>


-- 


[image: David Nickerson on about.me]

David Nickerson
about.me/david.nickerson
  <http://about.me/david.nickerson>

Re: [SED-ML-discuss] Exists a tool which can handle the L1V2 specification example?

[Matthias K. <kon...@go...>]

Hi David,

I agree xpath works. But there should be some abstraction layer around it
either by libsbml or libsedml (and for the XMLChanges).
See
https://www.pivotaltracker.com/n/projects/248655/stories/114704823

I personally feel very uncomfortable working with the raw XML (I am not a
computer scientist, I am a computional biologist).
The core idea of libsbml and libsedml is to provide an abstraction layer
around the XML which provides a simple way to work with the model/model
changes.

Xpath expressions & the XMLChanges are the only cases where I currently
have to really manipulate/work with the XML, i.e. xml parser, xpath parser,
changing the dom. All things which are highly error prone and should be
abstracted away by the respective libraries. Not providing support for this
by libsbml or libsedml will cause a multitude of problems in the long run
because none of the xpath and xmlChanges will be standardized. So you
suddenly have a standard which allows you to do basically everything
(generic XMLChanges, generic xpath expressions) but most of the tools only
supporting a very limited subset (only resolving certain xpath expressions,
not checking if the XMLChanges are allowed and result in valid models).
This will break compatibility on so many levels.
This will be the main reason why SEDML will not be reproducible between
different tools.

Matthias


On Thu, Mar 24, 2016 at 9:36 AM, David Nickerson <dav...@gm...>
wrote:

> > Would it make sense to restrict xpath expressions so that we don't run
> into
> > these sorts of problems?  This is only going to get worse, I think.  We
> > don't even have recommended guidelines for how to construct XPaths.
> >
> > I mean, I'd love to drop xpath use altogether, but I also don't see a
> great
> > alternative.  But maybe a subset would be the best of both worlds.
>
> I think XPath works well when we are dealing with models encoded in
> XML formats and tools implement support for XPath rather than relying
> on extracting key information from the XPath expression. I don't think
> it would be good to enforce a restricted subset of XPath in the SED-ML
> specification, and I also don't think we need recommended guidelines
> for constructing XPath expressions. XPath is a fairly well established
> standard which is well supported across most programming languages so
> it makes sense to simply use it.
>
> It might be worth considering supporting additional methods/schemes
> for identifying entities in source models. Such as URIs or annotation
> terms.
>
> Cheers,
> David.
>
>
> ------------------------------------------------------------------------------
> Transform Data into Opportunity.
> Accelerate data analysis in your applications with
> Intel Data Analytics Acceleration Library.
> Click to learn more.
> http://pubads.g.doubleclick.net/gampad/clk?id=278785351&iu=/4140
> _______________________________________________
> SED-ML-discuss mailing list
> SED...@li...
> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>



-- 
Matthias König
Junior Group Leader LiSym - Systems Medicine of the Liver
Humboldt-University Berlin, Institute for Theoretical Biology
  https://www.livermetabolism.com
kon...@go...
Tel: +49 30 20938450
Tel: +49 176 81168480

Re: [SED-ML-discuss] Exists a tool which can handle the L1V2 specification example?

[David N. <dav...@gm...>]

> Would it make sense to restrict xpath expressions so that we don't run into
> these sorts of problems?  This is only going to get worse, I think.  We
> don't even have recommended guidelines for how to construct XPaths.
>
> I mean, I'd love to drop xpath use altogether, but I also don't see a great
> alternative.  But maybe a subset would be the best of both worlds.

I think XPath works well when we are dealing with models encoded in
XML formats and tools implement support for XPath rather than relying
on extracting key information from the XPath expression. I don't think
it would be good to enforce a restricted subset of XPath in the SED-ML
specification, and I also don't think we need recommended guidelines
for constructing XPath expressions. XPath is a fairly well established
standard which is well supported across most programming languages so
it makes sense to simply use it.

It might be worth considering supporting additional methods/schemes
for identifying entities in source models. Such as URIs or annotation
terms.

Cheers,
David.

Re: [SED-ML-discuss] Exists a tool which can handle the L1V2 specification example?

[Lucian S. <luc...@gm...>]

On Wed, Mar 23, 2016 at 4:18 AM, Ilya Kiselev <ax...@de...
> wrote:

> Hi Mathias.
>
> Sorry for not writing earlier.
>
> BioUML can import and simulate such examples.
> However it currently can not parse complicated xpath expression like "
> */sbml:sbml/sbml:model/descendant::*[@id='PX']*"
> I've changed them to the form "
> */sbml:sbml/sbml:model/sbml:listOfSpecies/sbml:species[@id='PX']*" in
> order to import your sedml to BioUML.
>

Would it make sense to restrict xpath expressions so that we don't run into
these sorts of problems?  This is only going to get worse, I think.  We
don't even have recommended guidelines for how to construct XPaths.

I mean, I'd love to drop xpath use altogether, but I also don't see a great
alternative.  But maybe a subset would be the best of both worlds.

1. curves must have id
>

Why do curves have to have an id?  Literally nothing in SED-ML can
reference them, right?  Can we drop this requirement and make them optional
for the next level/version?

2. *max *functions should be used as special csymbol:
>     <apply>
>         <csymbol encoding="text" definitionURL="http://sed-ml.org/#max">
>             max
>         </csymbol>
>         <ci> task1_____PZ </ci>
>     </apply>
>

Is there software that will parse 'max' in infix and create this output?
Or something that will take libsbml ASTNodes and convert them to this?

-Lucian

Re: [SED-ML-discuss] Exists a tool which can handle the L1V2 specification example?

[Matthias K. <kon...@go...>]

Hi Ilya,
thanks so much. This looks great.

I never used BioUML before so I am a bit lost here. I downloaded BioUML
workbench 0.9.10 from the BioStore and installed it. I looked for something
which resembles SBML/SEDML but could not find anything. Could you provide a
stepwise explanation of what do I have to do/click in the GUI to get the 3
plots from the sedx archive?
Not even sure where to start.
Is there any citation for BioUML, so I can cite you when doing the
roundtrip? Could not find anything on your wiki or pubmed.

I filed the "descendent" issue in phrasedml
https://sourceforge.net/p/phrasedml/tickets/18/
Some more info here
https://github.com/sys-bio/tellurium/issues/114
Technically the expression is correct, but will probably create problems
with most tools.

Thanks for the specification hints. If you find any other issues with the
SEDML, please let me know.
It is very difficult to validate SEDML because there is complete validation
against the specification of files anywhere. Or is there?
If libSEDML is not giving warnings I accept the file as correct.
Filed the two things here
https://sourceforge.net/p/phrasedml/tickets/21/

The best
Matthias


On Wed, Mar 23, 2016 at 12:18 PM, Ilya Kiselev <
ax...@de...> wrote:

> Hi Mathias.
>
> Sorry for not writing earlier.
>
> BioUML can import and simulate such examples.
> However it currently can not parse complicated xpath expression like "
> */sbml:sbml/sbml:model/descendant::*[@id='PX']*"
> I've changed them to the form "
> */sbml:sbml/sbml:model/sbml:listOfSpecies/sbml:species[@id='PX']*" in
> order to import your sedml to BioUML.
>
> Also your sedml document is no quite according to specifiction
>
> 1. curves must have id
> 2. *max *functions should be used as special csymbol:
>     <apply>
>         <csymbol encoding="text" definitionURL="http://sed-ml.org/#max">
>             max
>         </csymbol>
>         <ci> task1_____PZ </ci>
>     </apply>
>
> instead of
>
>    <apply>
>         <ci> max </ci>
>         <ci> task1_____PZ </ci>
>    </apply>
>
> After fixing this and changing xpath to the form compatible with BioUML, I
> was able to import your sedml and simulate in BioUML. Generated plots are
> attached.
> c_0_0, c_0_1,... are new ids of curves.
> Fixed sedml is also attached.
>
>
> [image: Встроенное изображение 1][image: Встроенное изображение 3][image:
> Встроенное изображение 2]
>
> 2016-03-23 14:23 GMT+06:00 Matthias König <kon...@go...>:
>
>> Hi all,
>> I want to roundtrip simple SED-ML examples with other tool to show that a
>> reproducible workflow is possible.
>>
>> Unfortunately, I could not find any tool which is able to handle the core
>> L1V2 specification example in the introduction section consisting of
>> - simple timecourse simulation
>> - change paramters & second simple timecourse simulation
>> - outputplot2d of time~species & time~species/max(species)
>>
>> Combine Archive and plots are attached.
>>
>> It would be very helpful if there would be any tool supporting the core
>> L1V2 features.  Otherwise the conclusion something in the direction:
>> "Currently, it is not possible to provide reproducible workflows which
>> roundtrip between different tools, because currently no other software/tool
>> is supporting the full L1V2 SEDML specification."
>> I would like to have a handful (2-3) example Combine archives which I can
>> roundtrip between different tools which show the core features of L1V2 with
>> the specification examples being one of them.
>>
>> Matthias
>>
>> [image: Inline image 3][image: Inline image 1][image: Inline image 2]
>>
>>
>> ------------------------------------------------------------------------------
>> Transform Data into Opportunity.
>> Accelerate data analysis in your applications with
>> Intel Data Analytics Acceleration Library.
>> Click to learn more.
>> http://pubads.g.doubleclick.net/gampad/clk?id=278785351&iu=/4140
>> _______________________________________________
>> SED-ML-discuss mailing list
>> SED...@li...
>> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>>
>>
>
>
> --
> С уважением,
> Илья.
>
>
> ------------------------------------------------------------------------------
> Transform Data into Opportunity.
> Accelerate data analysis in your applications with
> Intel Data Analytics Acceleration Library.
> Click to learn more.
> http://pubads.g.doubleclick.net/gampad/clk?id=278785351&iu=/4140
> _______________________________________________
> SED-ML-discuss mailing list
> SED...@li...
> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>
>


-- 
Matthias König
Junior Group Leader LiSym - Systems Medicine of the Liver
Humboldt-University Berlin, Institute for Theoretical Biology
  https://www.livermetabolism.com
kon...@go...
Tel: +49 30 20938450
Tel: +49 176 81168480

Re: [SED-ML-discuss] Exists a tool which can handle the L1V2 specification example?

[Ilya K. <ax...@de...>]

Hi Mathias.

Sorry for not writing earlier.

BioUML can import and simulate such examples.
However it currently can not parse complicated xpath expression like "
*/sbml:sbml/sbml:model/descendant::*[@id='PX']*"
I've changed them to the form "
*/sbml:sbml/sbml:model/sbml:listOfSpecies/sbml:species[@id='PX']*" in order
to import your sedml to BioUML.

Also your sedml document is no quite according to specifiction

1. curves must have id
2. *max *functions should be used as special csymbol:
    <apply>
        <csymbol encoding="text" definitionURL="http://sed-ml.org/#max">
            max
        </csymbol>
        <ci> task1_____PZ </ci>
    </apply>

instead of

   <apply>
        <ci> max </ci>
        <ci> task1_____PZ </ci>
   </apply>

After fixing this and changing xpath to the form compatible with BioUML, I
was able to import your sedml and simulate in BioUML. Generated plots are
attached.
c_0_0, c_0_1,... are new ids of curves.
Fixed sedml is also attached.


[image: Встроенное изображение 1][image: Встроенное изображение 3][image:
Встроенное изображение 2]

2016-03-23 14:23 GMT+06:00 Matthias König <kon...@go...>:

> Hi all,
> I want to roundtrip simple SED-ML examples with other tool to show that a
> reproducible workflow is possible.
>
> Unfortunately, I could not find any tool which is able to handle the core
> L1V2 specification example in the introduction section consisting of
> - simple timecourse simulation
> - change paramters & second simple timecourse simulation
> - outputplot2d of time~species & time~species/max(species)
>
> Combine Archive and plots are attached.
>
> It would be very helpful if there would be any tool supporting the core
> L1V2 features.  Otherwise the conclusion something in the direction:
> "Currently, it is not possible to provide reproducible workflows which
> roundtrip between different tools, because currently no other software/tool
> is supporting the full L1V2 SEDML specification."
> I would like to have a handful (2-3) example Combine archives which I can
> roundtrip between different tools which show the core features of L1V2 with
> the specification examples being one of them.
>
> Matthias
>
> [image: Inline image 3][image: Inline image 1][image: Inline image 2]
>
>
> ------------------------------------------------------------------------------
> Transform Data into Opportunity.
> Accelerate data analysis in your applications with
> Intel Data Analytics Acceleration Library.
> Click to learn more.
> http://pubads.g.doubleclick.net/gampad/clk?id=278785351&iu=/4140
> _______________________________________________
> SED-ML-discuss mailing list
> SED...@li...
> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>
>


-- 
С уважением,
Илья.

[SED-ML-discuss] Exists a tool which can handle the L1V2 specification example?

[Matthias K. <kon...@go...>]

Hi all,
I want to roundtrip simple SED-ML examples with other tool to show that a
reproducible workflow is possible.

Unfortunately, I could not find any tool which is able to handle the core
L1V2 specification example in the introduction section consisting of
- simple timecourse simulation
- change paramters & second simple timecourse simulation
- outputplot2d of time~species & time~species/max(species)

Combine Archive and plots are attached.

It would be very helpful if there would be any tool supporting the core
L1V2 features.  Otherwise the conclusion something in the direction:
"Currently, it is not possible to provide reproducible workflows which
roundtrip between different tools, because currently no other software/tool
is supporting the full L1V2 SEDML specification."
I would like to have a handful (2-3) example Combine archives which I can
roundtrip between different tools which show the core features of L1V2 with
the specification examples being one of them.

Matthias

[image: Inline image 3][image: Inline image 1][image: Inline image 2]

[SED-ML-discuss] Soliciting ideas for HARMONY topics

[David N. <dav...@gm...>]

Hi all,

HARMONY 2016 is coming up (and registration is open:
http://co.mbine.org/events/HARMONY_2016) and we are starting to
collect topics for possible SED-ML related breakouts/sessions during
the meeting.

If you have any ideas for topics or burning issues that you would like
to discuss, please let me know or raise them in on this list.


Cheers,
David.

-- 


David Nickerson
about.me/david.nickerson

[SED-ML-discuss] the meaning of this

[awbestesq <awb...@gm...>]

hello,

i have come across the following sed statement which is cryptic to myself.

sed -e '/^ *$/ d;/^$/ d' filename

it is the regex portion that i have never seen. i presume that /^ *$/ means match a line with zero or more blanks with no other chars and the /^$/ means to match a line with no characters; it is the ' d;' and ' d' that i am baffled by.

tia,
awb

reticulate - being or involving evolutionary change dependent on genetic recombination involving diverse interbreeding populations 

Re: [SED-ML-discuss] How to interpret ComputeChange & FunctionalRange & curves of repeatedTask in L1V2

[Chris J. M. <my...@ec...>]

Hi,

iBioSim uses an old version of jlibsedml, and it only really works for simple sed-ml files like those found in the SBML testsuite.  I see though jlibsedml has seen some recent activity.  We will try updating and see if we can add more support.

Cheers,

Chris

> On Mar 10, 2016, at 2:52 AM, Matthias König <kon...@go...> wrote:
> 
> Hi all, 
> 
> like mentioned earlier I was missing a comprehensive list of examples to test the SED-ML implementation against. I created a set of tests files & used additional files from Frank & Lucian to test my implementation of L1V2.
> https://drive.google.com/file/d/0B5gYPV9g1eEKVmNCZU9SMHBSYkk/view?usp=sharing <https://drive.google.com/file/d/0B5gYPV9g1eEKVmNCZU9SMHBSYkk/view?usp=sharing>
> I added the sedx and the resulting output with tellurium (reports, plot2D, plot3D). 
> I am not completely sure if all outputs are correct so it would be nice to see the results of other tools with the archives. 
> 
> I tried SEDML webtools which could handle many of the examples and if outputs were generated I added the webtools plots in the respective testcase folder.
> 
> I tried CellDesigner 4.4 via:
> Importing model and than
> Simulation Control Panel -> Open Condition -> Select SED-ML for model
> -> For the example I get either  "Can not read your file" messages for the SEDML files or the SBML L3V1 files are not supported (which are most of them).
> 
> I tried Copasi COPASI-4.16.104 
> which worked good for simple timecourse simulations, but did not work for anything more complex
> 
> I tried iBioSim-2.8.4 but there seems to be a bug with reading SED-ML files, so could not get any outputs.
> 
> I saw SEDML support also in the Systems Biology Simultion Core Library but did not try it so far.
> Are there any other tools which have comprehensive SED-ML support I could try?
> 
> It would be interesting to see some reference solutions for all the test cases.
> 
> Here some insights from the implementation:
> [1]
> - the tasks can form a quit complex tree (RepeatedTasks of RepeatedTasks of RepeatedTasks with multiple subtasks)
> -> this makes it challenging to know when what to update and also to handle the merging of the subtasks
> -> also the resetting of models becomes challenging in this scenarios
> 
> I don't think any other tool handles this, but most tools stop at simple repeatedTask. Things like the following with R RepeatedTask and T Task should be handled:
> 
> R
>  | - T
>  | - R
>      | - T
>      | - R
>           | - T
>  | - T
> 
> [2]
> I could not find information on how to annotate plots with additional visual information. Especially the color of the lines would be very useful to annotate because every tool has another default color order which makes it visually challenging to compare plots
> - other things which should have a clear way to annotate are axis limits, axis labels, titel, grid, ..., linestyle, markerstyle, ...
> It would be nice to have one standard way to handle all these additional plot/curve information: i.e. a standard way to annotate these things.
> 
> [3]
> Some things are missing in SED-ML which are important for simulation descriptions and would be nice to have in L1V3. I would use all of the following features right away:
> - NonUniformTimecourse (already mentioned & filed a Ticked. What is the status? Will this be implemented?)
> 
> - real 3D plots: i.e.surface plots, contour plots, ... all based on meshgrids (x vector, y vector, z 2D matrix).
> 
> - linear Algebra operations in the math of dataGenerators.  All other math (like on ComputeChange) can stay like it is, but to have 3D plots and to be able to represent real analyses linear algebra operations are required. For instance things like 
> * mean over repeats (needed in repeated gillespie to get mean trajectory)
> * vector slices (for instance only last element of time course)
> i.e. all basic math operations on vectors, i.e. elementwise sum, product, ..
> This is what everybody is doing running simulations. They collect repeats and than perform some linear algebra/vector operations on them.
> 
> - some interpolate math which allows to interpolate time intervals of dataGenerators. Required for NonUniformTimecourse.
> 
> Matthias
> 
> 
> 
> On Wed, Mar 2, 2016 at 6:27 PM, Matthias König <kon...@go... <mailto:kon...@go...>> wrote:
> Yes, a flow chart would help a lot. The implementation so far was a lot of trial and error and took quit some time to figure out why things are like they are in the spec. Only thing unclear right now is how to handle recursive repeated tasks (repeated tasks which have repeated tasks as subtasks).
> 
> Also a comprehensive list of examples with the expected outputs would be very helpful for implementation (plot2D, plot3D and reports). Which tools beside SED-ML webtools provide full implementation of the specification (with SBML models at least)? Preferable once which can handle combine archives as input?
> I want: Combine Archive -> Tool -> outputs
> M
> 
> 
> On Wed, Mar 2, 2016 at 11:04 AM, Martin Peters <acc...@fr... <mailto:acc...@fr...>> wrote:
> Hello everybody,
> 
> Lately I struggled on a similar issue as Matthias. Thankfully Dagmar
> could answer my questions. But may I suggest to include a flow chart (or
> similar), describing how to parse the Sed-ML in the spec?
> To be more precise a graphical explanation of when to load models, apply
> changes, calculate MathML etc. would have helped me a lot with
> understanding the spec in detail.
> 
> But thanks for the great work so far and sorry for being slightly OT.
> 
> Cheers
> Martin Peters
> 
> On 2016-03-01 20:34, Matthias König wrote:
> > Hi Jonathan,
> > thanks a lot.
> > I handled [1] like you mentioned and was even easier than creating the
> > InitialAssignments.
> > [2] thanks. solved. Somehow overread it in the spec.
> > [3] did not think about multiple axis. This solves it. The plot has to
> > generate multiple axis in the case of mixed lin and log
> >
> > Are there any Combine archives (SBML+SEDML) examples for the use of [A]
> > FunctionalRange or [B] DataGenerators which are calculated via complex
> > formulas (something like piecewise, min, max, ...)
> >
> > @Lucian Can phrasedml generated [A] or [B] examples
> >
> > Matthias
> >
> >
> > On Tue, Mar 1, 2016 at 1:14 PM, Jonathan Cooper
> > <jon...@cs... <mailto:jon...@cs...> <mailto:jon...@cs... <mailto:jon...@cs...>>> wrote:
> >
> >     Hi Matthias,
> >
> >     For 1, ComputeChange is not quite identical to an
> >     InitialAssignmentRule as it happens when the model is set up, not
> >     when a simulation starts. So model3 should have S2=4.3 inherited
> >     from model2.
> >
> >     For 2, the intent is that FunctionalRange is calculated at each
> >     iteration, not all upfront.  I quote from p52:
> >
> >     "In the listOfVariables, variable elements define identifiers
> >     referring to model variables or range values, which may then be used
> >     within the math expression.  These references always retrieve the
> >     current value of the model variable or range at the current
> >     iteration of the enclosing repeatedTask.  For a model not being
> >     simulated by any subTask, the initial state of the model is used."
> >
> >     For 3, I'm not sure of the original reasoning, but it's certainly
> >     not uncommon to have multiple y axes displayed on a plot, one of
> >     which might be log scale and one linear.  Having different x axes is
> >     less likely to arise.
> >
> >     As Andre mentioned in his email, 4 is not defined by the current
> >     specification, as there is no explicit support for multidimensional
> >     arrays. This may come with the integration of NuML and parameter
> >     estimation in the next version(s).
> >
> >     Best wishes,
> >     Jonathan
> >
> >
> >     On 01/03/2016 09:59, Matthias König wrote:
> >>     Hi all,
> >>
> >>     I am currently implementing SEDML support for tellurium and a few
> >>     things in the SEDML specification are unclear for me.
> >>
> >>     [1] ComputeChange
> >>
> >>     The specification says: "The ComputeChange class permits to change
> >>     the numerical value of any element or attribute of a model based
> >>     on a calculation."
> >>     For me it is completely unclear when this calculation is
> >>     performed. Depending on the time of the actual evaluation
> >>     different numerical values are set in the model. After discussion
> >>     with Lucian I understood the ComputeChange as an
> >>     InitialAssignmentRule in SBML.
> >>     But what happens if later model changes (of derived models) change
> >>     the variables used in the ComputeChange?
> >>
> >>     Here an example:
> >>     It is not clear what should happen if for instance later model
> >>     versions change variables which are used in the calculation of the
> >>     computeChange, i.e. something like|
> >>     model1 = model "test"
> >>     model2 = model model1 with S1=0.3, S2=S1+4|
> >>     |model3 = model model2 with S1=10
> >>     |
> >>     What is the value of S2 in model 3?
> >>     If S2=S1+4 was impmented via an InitialAssignment than init(S2) in
> >>     model3 is 14 instead of 4.3 (which was the computeChange set in
> >>     model2). In my understanding model3 should have an S2=4.3, because
> >>     the computeChange should be calculated based on the state of model2.
> >>     So what is the correct interpretation (S2=14 or S2=4.3 in model3).
> >>     It has to be clearly stated when the calculation of the
> >>     ComputeChange has to be performed (within every model or after all
> >>     changes to all models)?
> >>
> >>     [2] FunctionalRange
> >>     Similar issue here. When is the computation of the functional
> >>     range performed? For instance in a repeatedTask where with every
> >>     repeat values of variables used in the calculation of the
> >>     functional range change.
> >>     Is the complete functional range calculated
> >>     - at the beginning of the loop with the initial model state before
> >>     any repeat
> >>     - or with the actual variables of the model at the beginning of
> >>     every iteration of the loop (which could change with reset=false)
> >>     Both variants would give very differnent numerical values for the
> >>     range.
> >>
> >>     [3] logX/logY/logZ on curves
> >>     Why are the logX,Y,Z not on the Plot, but on the curves?
> >>     What should be done if multiple curves have contradicting log
> >>     settings in a plot? I.e. two curves which have the same
> >>     xDataReference but different log settings.
> >>     <plot2D id="plot_0" name="Example plot"> <listOfCurves> <curve
> >>     logX="false" logY="false" xDataReference="plot_0_0_0"
> >>     yDataReference="plot_0_0_1"/> <curve logX="true" logY="false"
> >>     xDataReference="plot_0_0_0"
> >>     yDataReference="plot_0_1_1"/></listOfCurves>
> >>     [4] How to plot curves of repeatedTasks?
> >>     I could not really find any information on how to plot curves of
> >>     repeatedTasks. I interpreted it as following:
> >>     - when reset=True plot every repeat as a single curve in the plot
> >>     (i.e. Nrep lines per curve)
> >>     - when reset=False concatenate the repeats and plot one line (the
> >>     repeats are time slices of the complete simulation)
> >>     Is this correct?
> >>     And how would I encode the respective other thing in sedml, i.e.
> >>     - a repeatedTask with reset=False where every repeat should be a
> >>     line in the plot
> >>     - a repeatedTask with reset=True which is concatenated
> >>
> >>
> >>     Matthias
> >>
> >>     --
> >>     Matthias König
> >>     Junior Group Leader LiSym - Systems Medicine of the Liver
> >>     Humboldt-University Berlin, Institute for Theoretical Biology
> >>       <https://www.livermetabolism.com <https://www.livermetabolism.com/>>https://www.livermetabolism.com <https://www.livermetabolism.com/>
> >>     <mailto:kon...@go... <mailto:kon...@go...>>kon...@go... <mailto:kon...@go...>
> >>     <mailto:kon...@go... <mailto:kon...@go...>>
> >>     Tel: +49 30 20938450 <tel:%2B49%2030%2020938450>
> >>     Tel: +49 176 81168480 <tel:%2B49%20176%2081168480> <tel:%2B49%20176%2081168480>
> >>
> >>
> >>     ------------------------------------------------------------------------------
> >>     Site24x7 APM Insight: Get Deep Visibility into Application Performance
> >>     APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
> >>     Monitor end-to-end web transactions and take corrective actions now
> >>     Troubleshoot faster and improve end-user experience. Signup Now!
> >>     http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140 <http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140>
> >>
> >>
> >>     _______________________________________________
> >>     SED-ML-discuss mailing list
> >>     SED...@li... <mailto:SED...@li...>
> >>     <mailto:SED...@li... <mailto:SED...@li...>>
> >>     https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss <https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss>
> >
> >
> >     ------------------------------------------------------------------------------
> >     Site24x7 APM Insight: Get Deep Visibility into Application Performance
> >     APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
> >     Monitor end-to-end web transactions and take corrective actions now
> >     Troubleshoot faster and improve end-user experience. Signup Now!
> >     http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140 <http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140>
> >     _______________________________________________
> >     SED-ML-discuss mailing list
> >     SED...@li... <mailto:SED...@li...>
> >     <mailto:SED...@li... <mailto:SED...@li...>>
> >     https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss <https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss>
> >
> >
> >
> >
> > --
> > Matthias König
> > Junior Group Leader LiSym - Systems Medicine of the Liver
> > Humboldt-University Berlin, Institute for Theoretical Biology
> >   https://www.livermetabolism.com <https://www.livermetabolism.com/>
> > kon...@go... <mailto:kon...@go...> <mailto:kon...@go... <mailto:kon...@go...>>
> > Tel: +49 30 20938450 <tel:%2B49%2030%2020938450>
> > Tel: +49 176 81168480 <tel:%2B49%20176%2081168480>
> >
> >
> > ------------------------------------------------------------------------------
> > Site24x7 APM Insight: Get Deep Visibility into Application Performance
> > APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
> > Monitor end-to-end web transactions and take corrective actions now
> > Troubleshoot faster and improve end-user experience. Signup Now!
> > http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140 <http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140>
> >
> >
> >
> > _______________________________________________
> > SED-ML-discuss mailing list
> > SED...@li... <mailto:SED...@li...>
> > https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss <https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss>
> >
> 
> ------------------------------------------------------------------------------
> Site24x7 APM Insight: Get Deep Visibility into Application Performance
> APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
> Monitor end-to-end web transactions and take corrective actions now
> Troubleshoot faster and improve end-user experience. Signup Now!
> http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140 <http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140>
> _______________________________________________
> SED-ML-discuss mailing list
> SED...@li... <mailto:SED...@li...>
> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss <https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss>
> 
> 
> 
> -- 
> Matthias König
> Junior Group Leader LiSym - Systems Medicine of the Liver
> Humboldt-University Berlin, Institute for Theoretical Biology
>   https://www.livermetabolism.com <https://www.livermetabolism.com/>
> kon...@go... <mailto:kon...@go...>
> Tel: +49 30 20938450
> Tel: +49 176 81168480 <tel:%2B49%20176%2081168480>
> 
> 
> 
> -- 
> Matthias König
> Junior Group Leader LiSym - Systems Medicine of the Liver
> Humboldt-University Berlin, Institute for Theoretical Biology
>   https://www.livermetabolism.com <https://www.livermetabolism.com/>
> kon...@go... <mailto:kon...@go...>
> Tel: +49 30 20938450
> Tel: +49 176 81168480
> ------------------------------------------------------------------------------
> Transform Data into Opportunity.
> Accelerate data analysis in your applications with
> Intel Data Analytics Acceleration Library.
> Click to learn more.
> http://pubads.g.doubleclick.net/gampad/clk?id=278785111&iu=/4140_______________________________________________
> SED-ML-discuss mailing list
> SED...@li...
> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss

Re: [SED-ML-discuss] How to interpret ComputeChange & FunctionalRange & curves of repeatedTask in L1V2

[Lucian S. <luc...@gm...>]

Thanks for all the test cases, Mattias!  Those are great.  I haven't yet
gone over them, but I've uploaded them to

https://github.com/luciansmith/sedml-test-suite/tree/master/contributions

and when I get time I'll try to move them to the more standardized part of
the test suite.  I've also made you a contributor, so if you come up with
more tests, you can upload them directly.

-Lucian

On Thu, Mar 10, 2016 at 1:52 AM, Matthias König <kon...@go...>
wrote:

> Hi all,
>
> like mentioned earlier I was missing a comprehensive list of examples to
> test the SED-ML implementation against. I created a set of tests files &
> used additional files from Frank & Lucian to test my implementation of L1V2.
>
> https://drive.google.com/file/d/0B5gYPV9g1eEKVmNCZU9SMHBSYkk/view?usp=sharing
> I added the sedx and the resulting output with tellurium (reports, plot2D,
> plot3D).
> I am not completely sure if all outputs are correct so it would be nice to
> see the results of other tools with the archives.
>
> I tried *SEDML webtools *which could handle many of the examples and if
> outputs were generated I added the webtools plots in the respective
> testcase folder.
>
> I tried *CellDesigner 4.4* via:
> Importing model and than
> Simulation Control Panel -> Open Condition -> Select SED-ML for model
> -> For the example I get either  "Can not read your file" messages for the
> SEDML files or the SBML L3V1 files are not supported (which are most of
> them).
>
> I tried Copasi
> *COPASI-4.16.104 *
> which worked good for simple timecourse simulations, but did not work for
> anything more complex
>
> I tried *iBioSim-2.8.4* but there seems to be a bug with reading SED-ML
> files, so could not get any outputs.
>
> I saw SEDML support also in the *Systems Biology Simultion Core Library*
> but did not try it so far.
> Are there any other tools which have comprehensive SED-ML support I could
> try?
>
> It would be interesting to see some reference solutions for all the test
> cases.
>
> Here some insights from the implementation:
> [1]
> - the tasks can form a quit complex tree (RepeatedTasks of RepeatedTasks
> of RepeatedTasks with multiple subtasks)
> -> this makes it challenging to know when what to update and also to
> handle the merging of the subtasks
> -> also the resetting of models becomes challenging in this scenarios
>
> I don't think any other tool handles this, but most tools stop at simple
> repeatedTask. Things like the following with R RepeatedTask and T Task
> should be handled:
>
> R
>  | - T
>  | - R
>      | - T
>      | - R
>           | - T
>  | - T
>
> [2]
> I could not find information on how to annotate plots with additional
> visual information. Especially the color of the lines would be very useful
> to annotate because every tool has another default color order which makes
> it visually challenging to compare plots
> - other things which should have a clear way to annotate are axis limits,
> axis labels, titel, grid, ..., linestyle, markerstyle, ...
> It would be nice to have one standard way to handle all these additional
> plot/curve information: i.e. a standard way to annotate these things.
>
> [3]
> Some things are missing in SED-ML which are important for simulation
> descriptions and would be nice to have in L1V3. I would use all of the
> following features right away:
> - NonUniformTimecourse (already mentioned & filed a Ticked. What is the
> status? Will this be implemented?)
>
> - real 3D plots: i.e.surface plots, contour plots, ... all based on
> meshgrids (x vector, y vector, z 2D matrix).
>
> - linear Algebra operations in the math of dataGenerators.  All other math
> (like on ComputeChange) can stay like it is, but to have 3D plots and to be
> able to represent real analyses linear algebra operations are required. For
> instance things like
> * mean over repeats (needed in repeated gillespie to get mean trajectory)
> * vector slices (for instance only last element of time course)
> i.e. all basic math operations on vectors, i.e. elementwise sum, product,
> ..
> This is what everybody is doing running simulations. They collect repeats
> and than perform some linear algebra/vector operations on them.
>
> - some interpolate math which allows to interpolate time intervals of
> dataGenerators. Required for NonUniformTimecourse.
>
> Matthias
>
>
>
> On Wed, Mar 2, 2016 at 6:27 PM, Matthias König <kon...@go...>
> wrote:
>
>> Yes, a flow chart would help a lot. The implementation so far was a lot
>> of trial and error and took quit some time to figure out why things are
>> like they are in the spec. Only thing unclear right now is how to handle
>> recursive repeated tasks (repeated tasks which have repeated tasks as
>> subtasks).
>>
>> Also a comprehensive list of examples with the expected outputs would be
>> very helpful for implementation (plot2D, plot3D and reports). Which tools
>> beside SED-ML webtools provide full implementation of the specification
>> (with SBML models at least)? Preferable once which can handle combine
>> archives as input?
>> I want: Combine Archive -> Tool -> outputs
>> M
>>
>>
>> On Wed, Mar 2, 2016 at 11:04 AM, Martin Peters <
>> acc...@fr...> wrote:
>>
>>> Hello everybody,
>>>
>>> Lately I struggled on a similar issue as Matthias. Thankfully Dagmar
>>> could answer my questions. But may I suggest to include a flow chart (or
>>> similar), describing how to parse the Sed-ML in the spec?
>>> To be more precise a graphical explanation of when to load models, apply
>>> changes, calculate MathML etc. would have helped me a lot with
>>> understanding the spec in detail.
>>>
>>> But thanks for the great work so far and sorry for being slightly OT.
>>>
>>> Cheers
>>> Martin Peters
>>>
>>> On 2016-03-01 20:34, Matthias König wrote:
>>> > Hi Jonathan,
>>> > thanks a lot.
>>> > I handled [1] like you mentioned and was even easier than creating the
>>> > InitialAssignments.
>>> > [2] thanks. solved. Somehow overread it in the spec.
>>> > [3] did not think about multiple axis. This solves it. The plot has to
>>> > generate multiple axis in the case of mixed lin and log
>>> >
>>> > Are there any Combine archives (SBML+SEDML) examples for the use of [A]
>>> > FunctionalRange or [B] DataGenerators which are calculated via complex
>>> > formulas (something like piecewise, min, max, ...)
>>> >
>>> > @Lucian Can phrasedml generated [A] or [B] examples
>>> >
>>> > Matthias
>>> >
>>> >
>>> > On Tue, Mar 1, 2016 at 1:14 PM, Jonathan Cooper
>>> > <jon...@cs... <mailto:jon...@cs...>>
>>> wrote:
>>> >
>>> >     Hi Matthias,
>>> >
>>> >     For 1, ComputeChange is not quite identical to an
>>> >     InitialAssignmentRule as it happens when the model is set up, not
>>> >     when a simulation starts. So model3 should have S2=4.3 inherited
>>> >     from model2.
>>> >
>>> >     For 2, the intent is that FunctionalRange is calculated at each
>>> >     iteration, not all upfront.  I quote from p52:
>>> >
>>> >     "In the listOfVariables, variable elements de fine identifiers
>>> >     referring to model variables or range values, which may then be
>>> used
>>> >     within the math expression.  These references always retrieve the
>>> >     current value of the model variable or range at the current
>>> >     iteration of the enclosing repeatedTask.  For a model not being
>>> >     simulated by any subTask, the initial state of the model is used."
>>> >
>>> >     For 3, I'm not sure of the original reasoning, but it's certainly
>>> >     not uncommon to have multiple y axes displayed on a plot, one of
>>> >     which might be log scale and one linear.  Having different x axes
>>> is
>>> >     less likely to arise.
>>> >
>>> >     As Andre mentioned in his email, 4 is not defined by the current
>>> >     specification, as there is no explicit support for multidimensional
>>> >     arrays. This may come with the integration of NuML and parameter
>>> >     estimation in the next version(s).
>>> >
>>> >     Best wishes,
>>> >     Jonathan
>>> >
>>> >
>>> >     On 01/03/2016 09:59, Matthias König wrote:
>>> >>     Hi all,
>>> >>
>>> >>     I am currently implementing SEDML support for tellurium and a few
>>> >>     things in the SEDML specification are unclear for me.
>>> >>
>>> >>     [1] ComputeChange
>>> >>
>>> >>     The specification says: "The ComputeChange class permits to change
>>> >>     the numerical value of any element or attribute of a model based
>>> >>     on a calculation."
>>> >>     For me it is completely unclear when this calculation is
>>> >>     performed. Depending on the time of the actual evaluation
>>> >>     different numerical values are set in the model. After discussion
>>> >>     with Lucian I understood the ComputeChange as an
>>> >>     InitialAssignmentRule in SBML.
>>> >>     But what happens if later model changes (of derived models) change
>>> >>     the variables used in the ComputeChange?
>>> >>
>>> >>     Here an example:
>>> >>     It is not clear what should happen if for instance later model
>>> >>     versions change variables which are used in the calculation of the
>>> >>     computeChange, i.e. something like|
>>> >>     model1 = model "test"
>>> >>     model2 = model model1 with S1=0.3, S2=S1+4|
>>> >>     |model3 = model model2 with S1=10
>>> >>     |
>>> >>     What is the value of S2 in model 3?
>>> >>     If S2=S1+4 was impmented via an InitialAssignment than init(S2) in
>>> >>     model3 is 14 instead of 4.3 (which was the computeChange set in
>>> >>     model2). In my understanding model3 should have an S2=4.3, because
>>> >>     the computeChange should be calculated based on the state of
>>> model2.
>>> >>     So what is the correct interpretation (S2=14 or S2=4.3 in model3).
>>> >>     It has to be clearly stated when the calculation of the
>>> >>     ComputeChange has to be performed (within every model or after all
>>> >>     changes to all models)?
>>> >>
>>> >>     [2] FunctionalRange
>>> >>     Similar issue here. When is the computation of the functional
>>> >>     range performed? For instance in a repeatedTask where with every
>>> >>     repeat values of variables used in the calculation of the
>>> >>     functional range change.
>>> >>     Is the complete functional range calculated
>>> >>     - at the beginning of the loop with the initial model state before
>>> >>     any repeat
>>> >>     - or with the actual variables of the model at the beginning of
>>> >>     every iteration of the loop (which could change with reset=false)
>>> >>     Both variants would give very differnent numerical values for the
>>> >>     range.
>>> >>
>>> >>     [3] logX/logY/logZ on curves
>>> >>     Why are the logX,Y,Z not on the Plot, but on the curves?
>>> >>     What should be done if multiple curves have contradicting log
>>> >>     settings in a plot? I.e. two curves which have the same
>>> >>     xDataReference but different log settings.
>>> >>     <plot2D id="plot_0" name="Example plot"> <listOfCurves> <curve
>>> >>     logX="false" logY="false" xDataReference="plot_0_0_0"
>>> >>     yDataReference="plot_0_0_1"/> <curve logX="true" logY="false"
>>> >>     xDataReference="plot_0_0_0"
>>> >>     yDataReference="plot_0_1_1"/></listOfCurves>
>>> >>     [4] How to plot curves of repeatedTasks?
>>> >>     I could not really find any information on how to plot curves of
>>> >>     repeatedTasks. I interpreted it as following:
>>> >>     - when reset=True plot every repeat as a single curve in the plot
>>> >>     (i.e. Nrep lines per curve)
>>> >>     - when reset=False concatenate the repeats and plot one line (the
>>> >>     repeats are time slices of the complete simulation)
>>> >>     Is this correct?
>>> >>     And how would I encode the respective other thing in sedml, i.e.
>>> >>     - a repeatedTask with reset=False where every repeat should be a
>>> >>     line in the plot
>>> >>     - a repeatedTask with reset=True which is concatenated
>>> >>
>>> >>
>>> >>     Matthias
>>> >>
>>> >>     --
>>> >>     Matthias König
>>> >>     Junior Group Leader LiSym - Systems Medicine of the Liver
>>> >>     Humboldt-University Berlin, Institute for Theoretical Biology
>>> >>       <https://www.livermetabolism.com>
>>> https://www.livermetabolism.com
>>> >>     <mailto:kon...@go...>kon...@go...
>>> >>     <mailto:kon...@go...>
>>> >>     Tel: +49 30 20938450
>>> >>     Tel: +49 176 81168480 <tel:%2B49%20176%2081168480>
>>> >>
>>> >>
>>> >>
>>>  ------------------------------------------------------------------------------
>>> >>     Site24x7 APM Insight: Get Deep Visibility into Application
>>> Performance
>>> >>     APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
>>> >>     Monitor end-to-end web transactions and take corrective actions
>>> now
>>> >>     Troubleshoot faster and improve end-user experience. Signup Now!
>>> >>     http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140
>>> >>
>>> >>
>>> >>     _______________________________________________
>>> >>     SED-ML-discuss mailing list
>>> >>     SED...@li...
>>> >>     <mailto:SED...@li...>
>>> >>     https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>>> >
>>> >
>>> >
>>>  ------------------------------------------------------------------------------
>>> >     Site24x7 APM Insight: Get Deep Visibility into Application
>>> Performance
>>> >     APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
>>> >     Monitor end-to-end web transactions and take corrective actions now
>>> >     Troubleshoot faster and improve end-user experience. Signup Now!
>>> >     http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140
>>> >     _______________________________________________
>>> >     SED-ML-discuss mailing list
>>> >     SED...@li...
>>> >     <mailto:SED...@li...>
>>> >     https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>>> >
>>> >
>>> >
>>> >
>>> > --
>>> > Matthias König
>>> > Junior Group Leader LiSym - Systems Medicine of the Liver
>>> > Humboldt-University Berlin, Institute for Theoretical Biology
>>> >   https://www.livermetabolism.com
>>> > kon...@go... <mailto:kon...@go...>
>>> > Tel: +49 30 20938450
>>> > Tel: +49 176 81168480
>>> >
>>> >
>>> >
>>> ------------------------------------------------------------------------------
>>> > Site24x7 APM Insight: Get Deep Visibility into Application Performance
>>> > APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
>>> > Monitor end-to-end web transactions and take corrective actions now
>>> > Troubleshoot faster and improve end-user experience. Signup Now!
>>> > http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140
>>> >
>>> >
>>> >
>>> > _______________________________________________
>>> > SED-ML-discuss mailing list
>>> > SED...@li...
>>> > https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>>> >
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> Site24x7 APM Insight: Get Deep Visibility into Application Performance
>>> APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
>>> Monitor end-to-end web transactions and take corrective actions now
>>> Troubleshoot faster and improve end-user experience. Signup Now!
>>> http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140
>>> _______________________________________________
>>> SED-ML-discuss mailing list
>>> SED...@li...
>>> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>>>
>>
>>
>>
>> --
>> Matthias König
>> Junior Group Leader LiSym - Systems Medicine of the Liver
>> Humboldt-University Berlin, Institute for Theoretical Biology
>>   https://www.livermetabolism.com
>> kon...@go...
>> Tel: +49 30 20938450
>> Tel: +49 176 81168480
>>
>
>
>
> --
> Matthias König
> Junior Group Leader LiSym - Systems Medicine of the Liver
> Humboldt-University Berlin, Institute for Theoretical Biology
>   https://www.livermetabolism.com
> kon...@go...
> Tel: +49 30 20938450
> Tel: +49 176 81168480
>
>
> ------------------------------------------------------------------------------
> Transform Data into Opportunity.
> Accelerate data analysis in your applications with
> Intel Data Analytics Acceleration Library.
> Click to learn more.
> http://pubads.g.doubleclick.net/gampad/clk?id=278785111&iu=/4140
> _______________________________________________
> SED-ML-discuss mailing list
> SED...@li...
> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>
>

Re: [SED-ML-discuss] How to interpret ComputeChange & FunctionalRange & curves of repeatedTask in L1V2

[Matthias K. <kon...@go...>]

Hi all,

like mentioned earlier I was missing a comprehensive list of examples to
test the SED-ML implementation against. I created a set of tests files &
used additional files from Frank & Lucian to test my implementation of L1V2.
https://drive.google.com/file/d/0B5gYPV9g1eEKVmNCZU9SMHBSYkk/view?usp=sharing
I added the sedx and the resulting output with tellurium (reports, plot2D,
plot3D).
I am not completely sure if all outputs are correct so it would be nice to
see the results of other tools with the archives.

I tried *SEDML webtools *which could handle many of the examples and if
outputs were generated I added the webtools plots in the respective
testcase folder.

I tried *CellDesigner 4.4* via:
Importing model and than
Simulation Control Panel -> Open Condition -> Select SED-ML for model
-> For the example I get either  "Can not read your file" messages for the
SEDML files or the SBML L3V1 files are not supported (which are most of
them).

I tried Copasi
*COPASI-4.16.104 *
which worked good for simple timecourse simulations, but did not work for
anything more complex

I tried *iBioSim-2.8.4* but there seems to be a bug with reading SED-ML
files, so could not get any outputs.

I saw SEDML support also in the *Systems Biology Simultion Core Library*
but did not try it so far.
Are there any other tools which have comprehensive SED-ML support I could
try?

It would be interesting to see some reference solutions for all the test
cases.

Here some insights from the implementation:
[1]
- the tasks can form a quit complex tree (RepeatedTasks of RepeatedTasks of
RepeatedTasks with multiple subtasks)
-> this makes it challenging to know when what to update and also to handle
the merging of the subtasks
-> also the resetting of models becomes challenging in this scenarios

I don't think any other tool handles this, but most tools stop at simple
repeatedTask. Things like the following with R RepeatedTask and T Task
should be handled:

R
 | - T
 | - R
     | - T
     | - R
          | - T
 | - T

[2]
I could not find information on how to annotate plots with additional
visual information. Especially the color of the lines would be very useful
to annotate because every tool has another default color order which makes
it visually challenging to compare plots
- other things which should have a clear way to annotate are axis limits,
axis labels, titel, grid, ..., linestyle, markerstyle, ...
It would be nice to have one standard way to handle all these additional
plot/curve information: i.e. a standard way to annotate these things.

[3]
Some things are missing in SED-ML which are important for simulation
descriptions and would be nice to have in L1V3. I would use all of the
following features right away:
- NonUniformTimecourse (already mentioned & filed a Ticked. What is the
status? Will this be implemented?)

- real 3D plots: i.e.surface plots, contour plots, ... all based on
meshgrids (x vector, y vector, z 2D matrix).

- linear Algebra operations in the math of dataGenerators.  All other math
(like on ComputeChange) can stay like it is, but to have 3D plots and to be
able to represent real analyses linear algebra operations are required. For
instance things like
* mean over repeats (needed in repeated gillespie to get mean trajectory)
* vector slices (for instance only last element of time course)
i.e. all basic math operations on vectors, i.e. elementwise sum, product, ..
This is what everybody is doing running simulations. They collect repeats
and than perform some linear algebra/vector operations on them.

- some interpolate math which allows to interpolate time intervals of
dataGenerators. Required for NonUniformTimecourse.

Matthias



On Wed, Mar 2, 2016 at 6:27 PM, Matthias König <kon...@go...>
wrote:

> Yes, a flow chart would help a lot. The implementation so far was a lot of
> trial and error and took quit some time to figure out why things are like
> they are in the spec. Only thing unclear right now is how to handle
> recursive repeated tasks (repeated tasks which have repeated tasks as
> subtasks).
>
> Also a comprehensive list of examples with the expected outputs would be
> very helpful for implementation (plot2D, plot3D and reports). Which tools
> beside SED-ML webtools provide full implementation of the specification
> (with SBML models at least)? Preferable once which can handle combine
> archives as input?
> I want: Combine Archive -> Tool -> outputs
> M
>
>
> On Wed, Mar 2, 2016 at 11:04 AM, Martin Peters <
> acc...@fr...> wrote:
>
>> Hello everybody,
>>
>> Lately I struggled on a similar issue as Matthias. Thankfully Dagmar
>> could answer my questions. But may I suggest to include a flow chart (or
>> similar), describing how to parse the Sed-ML in the spec?
>> To be more precise a graphical explanation of when to load models, apply
>> changes, calculate MathML etc. would have helped me a lot with
>> understanding the spec in detail.
>>
>> But thanks for the great work so far and sorry for being slightly OT.
>>
>> Cheers
>> Martin Peters
>>
>> On 2016-03-01 20:34, Matthias König wrote:
>> > Hi Jonathan,
>> > thanks a lot.
>> > I handled [1] like you mentioned and was even easier than creating the
>> > InitialAssignments.
>> > [2] thanks. solved. Somehow overread it in the spec.
>> > [3] did not think about multiple axis. This solves it. The plot has to
>> > generate multiple axis in the case of mixed lin and log
>> >
>> > Are there any Combine archives (SBML+SEDML) examples for the use of [A]
>> > FunctionalRange or [B] DataGenerators which are calculated via complex
>> > formulas (something like piecewise, min, max, ...)
>> >
>> > @Lucian Can phrasedml generated [A] or [B] examples
>> >
>> > Matthias
>> >
>> >
>> > On Tue, Mar 1, 2016 at 1:14 PM, Jonathan Cooper
>> > <jon...@cs... <mailto:jon...@cs...>>
>> wrote:
>> >
>> >     Hi Matthias,
>> >
>> >     For 1, ComputeChange is not quite identical to an
>> >     InitialAssignmentRule as it happens when the model is set up, not
>> >     when a simulation starts. So model3 should have S2=4.3 inherited
>> >     from model2.
>> >
>> >     For 2, the intent is that FunctionalRange is calculated at each
>> >     iteration, not all upfront.  I quote from p52:
>> >
>> >     "In the listOfVariables, variable elements de fine identifiers
>> >     referring to model variables or range values, which may then be used
>> >     within the math expression.  These references always retrieve the
>> >     current value of the model variable or range at the current
>> >     iteration of the enclosing repeatedTask.  For a model not being
>> >     simulated by any subTask, the initial state of the model is used."
>> >
>> >     For 3, I'm not sure of the original reasoning, but it's certainly
>> >     not uncommon to have multiple y axes displayed on a plot, one of
>> >     which might be log scale and one linear.  Having different x axes is
>> >     less likely to arise.
>> >
>> >     As Andre mentioned in his email, 4 is not defined by the current
>> >     specification, as there is no explicit support for multidimensional
>> >     arrays. This may come with the integration of NuML and parameter
>> >     estimation in the next version(s).
>> >
>> >     Best wishes,
>> >     Jonathan
>> >
>> >
>> >     On 01/03/2016 09:59, Matthias König wrote:
>> >>     Hi all,
>> >>
>> >>     I am currently implementing SEDML support for tellurium and a few
>> >>     things in the SEDML specification are unclear for me.
>> >>
>> >>     [1] ComputeChange
>> >>
>> >>     The specification says: "The ComputeChange class permits to change
>> >>     the numerical value of any element or attribute of a model based
>> >>     on a calculation."
>> >>     For me it is completely unclear when this calculation is
>> >>     performed. Depending on the time of the actual evaluation
>> >>     different numerical values are set in the model. After discussion
>> >>     with Lucian I understood the ComputeChange as an
>> >>     InitialAssignmentRule in SBML.
>> >>     But what happens if later model changes (of derived models) change
>> >>     the variables used in the ComputeChange?
>> >>
>> >>     Here an example:
>> >>     It is not clear what should happen if for instance later model
>> >>     versions change variables which are used in the calculation of the
>> >>     computeChange, i.e. something like|
>> >>     model1 = model "test"
>> >>     model2 = model model1 with S1=0.3, S2=S1+4|
>> >>     |model3 = model model2 with S1=10
>> >>     |
>> >>     What is the value of S2 in model 3?
>> >>     If S2=S1+4 was impmented via an InitialAssignment than init(S2) in
>> >>     model3 is 14 instead of 4.3 (which was the computeChange set in
>> >>     model2). In my understanding model3 should have an S2=4.3, because
>> >>     the computeChange should be calculated based on the state of
>> model2.
>> >>     So what is the correct interpretation (S2=14 or S2=4.3 in model3).
>> >>     It has to be clearly stated when the calculation of the
>> >>     ComputeChange has to be performed (within every model or after all
>> >>     changes to all models)?
>> >>
>> >>     [2] FunctionalRange
>> >>     Similar issue here. When is the computation of the functional
>> >>     range performed? For instance in a repeatedTask where with every
>> >>     repeat values of variables used in the calculation of the
>> >>     functional range change.
>> >>     Is the complete functional range calculated
>> >>     - at the beginning of the loop with the initial model state before
>> >>     any repeat
>> >>     - or with the actual variables of the model at the beginning of
>> >>     every iteration of the loop (which could change with reset=false)
>> >>     Both variants would give very differnent numerical values for the
>> >>     range.
>> >>
>> >>     [3] logX/logY/logZ on curves
>> >>     Why are the logX,Y,Z not on the Plot, but on the curves?
>> >>     What should be done if multiple curves have contradicting log
>> >>     settings in a plot? I.e. two curves which have the same
>> >>     xDataReference but different log settings.
>> >>     <plot2D id="plot_0" name="Example plot"> <listOfCurves> <curve
>> >>     logX="false" logY="false" xDataReference="plot_0_0_0"
>> >>     yDataReference="plot_0_0_1"/> <curve logX="true" logY="false"
>> >>     xDataReference="plot_0_0_0"
>> >>     yDataReference="plot_0_1_1"/></listOfCurves>
>> >>     [4] How to plot curves of repeatedTasks?
>> >>     I could not really find any information on how to plot curves of
>> >>     repeatedTasks. I interpreted it as following:
>> >>     - when reset=True plot every repeat as a single curve in the plot
>> >>     (i.e. Nrep lines per curve)
>> >>     - when reset=False concatenate the repeats and plot one line (the
>> >>     repeats are time slices of the complete simulation)
>> >>     Is this correct?
>> >>     And how would I encode the respective other thing in sedml, i.e.
>> >>     - a repeatedTask with reset=False where every repeat should be a
>> >>     line in the plot
>> >>     - a repeatedTask with reset=True which is concatenated
>> >>
>> >>
>> >>     Matthias
>> >>
>> >>     --
>> >>     Matthias König
>> >>     Junior Group Leader LiSym - Systems Medicine of the Liver
>> >>     Humboldt-University Berlin, Institute for Theoretical Biology
>> >>       <https://www.livermetabolism.com>https://www.livermetabolism.com
>> >>     <mailto:kon...@go...>kon...@go...
>> >>     <mailto:kon...@go...>
>> >>     Tel: +49 30 20938450
>> >>     Tel: +49 176 81168480 <tel:%2B49%20176%2081168480>
>> >>
>> >>
>> >>
>>  ------------------------------------------------------------------------------
>> >>     Site24x7 APM Insight: Get Deep Visibility into Application
>> Performance
>> >>     APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
>> >>     Monitor end-to-end web transactions and take corrective actions now
>> >>     Troubleshoot faster and improve end-user experience. Signup Now!
>> >>     http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140
>> >>
>> >>
>> >>     _______________________________________________
>> >>     SED-ML-discuss mailing list
>> >>     SED...@li...
>> >>     <mailto:SED...@li...>
>> >>     https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>> >
>> >
>> >
>>  ------------------------------------------------------------------------------
>> >     Site24x7 APM Insight: Get Deep Visibility into Application
>> Performance
>> >     APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
>> >     Monitor end-to-end web transactions and take corrective actions now
>> >     Troubleshoot faster and improve end-user experience. Signup Now!
>> >     http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140
>> >     _______________________________________________
>> >     SED-ML-discuss mailing list
>> >     SED...@li...
>> >     <mailto:SED...@li...>
>> >     https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>> >
>> >
>> >
>> >
>> > --
>> > Matthias König
>> > Junior Group Leader LiSym - Systems Medicine of the Liver
>> > Humboldt-University Berlin, Institute for Theoretical Biology
>> >   https://www.livermetabolism.com
>> > kon...@go... <mailto:kon...@go...>
>> > Tel: +49 30 20938450
>> > Tel: +49 176 81168480
>> >
>> >
>> >
>> ------------------------------------------------------------------------------
>> > Site24x7 APM Insight: Get Deep Visibility into Application Performance
>> > APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
>> > Monitor end-to-end web transactions and take corrective actions now
>> > Troubleshoot faster and improve end-user experience. Signup Now!
>> > http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140
>> >
>> >
>> >
>> > _______________________________________________
>> > SED-ML-discuss mailing list
>> > SED...@li...
>> > https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>> >
>>
>>
>> ------------------------------------------------------------------------------
>> Site24x7 APM Insight: Get Deep Visibility into Application Performance
>> APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
>> Monitor end-to-end web transactions and take corrective actions now
>> Troubleshoot faster and improve end-user experience. Signup Now!
>> http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140
>> _______________________________________________
>> SED-ML-discuss mailing list
>> SED...@li...
>> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>>
>
>
>
> --
> Matthias König
> Junior Group Leader LiSym - Systems Medicine of the Liver
> Humboldt-University Berlin, Institute for Theoretical Biology
>   https://www.livermetabolism.com
> kon...@go...
> Tel: +49 30 20938450
> Tel: +49 176 81168480
>



-- 
Matthias König
Junior Group Leader LiSym - Systems Medicine of the Liver
Humboldt-University Berlin, Institute for Theoretical Biology
  https://www.livermetabolism.com
kon...@go...
Tel: +49 30 20938450
Tel: +49 176 81168480

[SED-ML-discuss] HARMONY 2016 - Registration now open

[David N. <dav...@gm...>]

Hi all,

We are very pleased to announce that registration is now open for HARMONY 2016.

Please see http://co.mbine.org/events/HARMONY_2016 for all the details.

HARMONY 2016 will be held June 7-11, in Auckland, New Zealand and will
be co-located with the 10th International CellML Workshop.

HARMONY is a hackathon-type meeting, with a focus on development of
the standards, interoperability and infrastructure. There are
generally not many general discussions or oral presentations during
HARMONY; instead, the time is devoted to allowing hands-on hacking and
interaction between people focused on practical development of
software and standards.

The annual International CellML Workshop is an opportunity for
scientists to present and discuss ongoing work related to
physiological modelling and simulation with particular focus on model
and data exchange & reuse, archiving, and versioning and related
software support.

Please note that there are a limited number of rooms available at the
HARMONY venue on a first come first served basis.

We hope to see you in Auckland!

David, on behalf of the HARMONY 2016 organisers.

-- 


David Nickerson
about.me/david.nickerson

Re: [SED-ML-discuss] L1V1 -> L1V2 conversion & applyChanges

[Matthias K. <kon...@go...>]

Hi Frank,
thanks so much. Will try this in detail.
M

On Tue, Mar 1, 2016 at 8:48 PM, Frank Bergmann <fbe...@ca...> wrote:

> Hello Matthias,
>
> it took a bit longer, but i finished that command line tool for getting
> the processed models or to convert between SED-ML Script and SEDML. Details
> are in the attached PDF.
>
> I have created a Ubuntu 15.10 x64 binary (that contains only files needed
> for the functions mentioned before):
>
> https://sourceforge.net/projects/libsedml/files/Ubuntu15.10-x64-sedml-console-only.tar.gz/download
>
> and a new Windows installer, that also contains the full CellML API and
> will allow you to simulate the models just like the SED-ML Web Tools do,
> along with the script editor:
>
> https://sourceforge.net/projects/libsedml/files/SetupSedML-win32-1.22.exe/download
>
> I hope that helps,
> best
> Frank
>
>
>
>
>
> > On Feb 24, 2016, at 9:35 AM, Matthias König <kon...@go...>
> wrote:
> >
> > Thanks. Very helpful resource. The source is at
> > https://sourceforge.net/p/libsedml/code/HEAD/tree/
> > There are a multitude of examples of sem-dl and sedx in the repo, for
> anybody looking for examples.
> > M
> >
> >
> >
> > On Tue, Feb 23, 2016 at 11:02 PM, David Nickerson <
> dav...@gm...> wrote:
> > I believe this is the C# code Frank was referring to:
> > http://libsedml.sourceforge.net/libSedML/Welcome.html
> >
> > Cheers,
> > David.
> >
> > On Wed, Feb 24, 2016 at 10:57 AM, Matthias König
> > <kon...@go...> wrote:
> > > Hi Frank,
> > > This helps a lot. Could you post a link to the c# source repository?
> > > Thanks. M
> > >
> > > On Feb 23, 2016 7:25 PM, "Frank Bergmann" <fbe...@ca...>
> wrote:
> > >>
> > >>
> > >> > On Feb 23, 2016, at 5:31 PM, Matthias König <
> kon...@go...>
> > >> > wrote:
> > >> >
> > >> > Hi all,
> > >> > I have some questions in regard to libsedml (python bindings)
> > >> >
> > >> > [1] Is there a conversion function to convert SedDocuments from
> L1V1 to
> > >> > L1V2 similar to libsbml? This would be very nice because than I
> have to
> > >> > read/implement only one specification.
> > >> > Should L1V1 still be supported or did the majority of the community
> > >> > already move on to L1V2 so that I can just ignore L1V1?
> > >> >
> > >>
> > >> L1V2 is a strict superset of L1V1, so if you just treat every model
> you
> > >> read as if it were an L1V2 model, then your code will work no matter
> whether
> > >> it is an l1v1 or l1v2 model. A converter as such is not necessary or
> even
> > >> possible there is just no way to express the l1v2 features in l1v1.
> > >>
> > >> > [2] I looked for a helper function which gives me for a model id
> the XML
> > >> > with all the applied changes. I.e. a function which reads the XML
> from the
> > >> > model source, and applies all the changes defined to the XML (in
> the process
> > >> > handling the model dependencies and the stepwise application of
> changes).
> > >> > This would be a very helpful function which everybody working with
> sedml
> > >> > needs and probably reimplements.
> > >> > Something like
> > >> > sedml_doc.getProcessedModelXML(model_id)
> > >> >
> > >>
> > >> I take contributions for the C++ project. I currently do not have the
> time
> > >> to add that. My C# library has all these features. so I could expose
> a SOAP
> > >> service / REST endpoint, or a command line tool, that would make it
> possible
> > >> to get the processed model.
> > >>
> > >> Cheers
> > >> Frank
> > >>
> > >> > The best
> > >> > Matthias
> > >> >
> > >> >
> > >> > --
> > >> > Matthias König
> > >> > Junior Group Leader LiSym - Systems Medicine of the Liver
> > >> > Humboldt-University Berlin, Institute for Theoretical Biology
> > >> >   https://www.livermetabolism.com
> > >> > kon...@go...
> > >> > Tel: +49 30 20938450
> > >> > Tel: +49 176 81168480
> > >> >
> > >> >
> ------------------------------------------------------------------------------
> > >> > Site24x7 APM Insight: Get Deep Visibility into Application
> Performance
> > >> > APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
> > >> > Monitor end-to-end web transactions and take corrective actions now
> > >> > Troubleshoot faster and improve end-user experience. Signup Now!
> > >> >
> > >> >
> http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140_______________________________________________
> > >> > SED-ML-discuss mailing list
> > >> > SED...@li...
> > >> > https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
> > >>
> > >>
> > >>
> > >>
> ------------------------------------------------------------------------------
> > >> Site24x7 APM Insight: Get Deep Visibility into Application Performance
> > >> APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
> > >> Monitor end-to-end web transactions and take corrective actions now
> > >> Troubleshoot faster and improve end-user experience. Signup Now!
> > >> http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140
> > >> _______________________________________________
> > >> SED-ML-discuss mailing list
> > >> SED...@li...
> > >> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
> > >
> > >
> > >
> ------------------------------------------------------------------------------
> > > Site24x7 APM Insight: Get Deep Visibility into Application Performance
> > > APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
> > > Monitor end-to-end web transactions and take corrective actions now
> > > Troubleshoot faster and improve end-user experience. Signup Now!
> > > http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140
> > > _______________________________________________
> > > SED-ML-discuss mailing list
> > > SED...@li...
> > > https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
> > >
> >
> >
> >
> > --
> >
> >
> > David Nickerson
> > about.me/david.nickerson
> >
> >
> ------------------------------------------------------------------------------
> > Site24x7 APM Insight: Get Deep Visibility into Application Performance
> > APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
> > Monitor end-to-end web transactions and take corrective actions now
> > Troubleshoot faster and improve end-user experience. Signup Now!
> > http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140
> > _______________________________________________
> > SED-ML-discuss mailing list
> > SED...@li...
> > https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
> >
> >
> >
> > --
> > Matthias König
> > Junior Group Leader LiSym - Systems Medicine of the Liver
> > Humboldt-University Berlin, Institute for Theoretical Biology
> >   https://www.livermetabolism.com
> > kon...@go...
> > Tel: +49 30 20938450
> > Tel: +49 176 81168480
> >
> ------------------------------------------------------------------------------
> > Site24x7 APM Insight: Get Deep Visibility into Application Performance
> > APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
> > Monitor end-to-end web transactions and take corrective actions now
> > Troubleshoot faster and improve end-user experience. Signup Now!
> >
> http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140_______________________________________________
> > SED-ML-discuss mailing list
> > SED...@li...
> > https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>
>
>
> ------------------------------------------------------------------------------
> Site24x7 APM Insight: Get Deep Visibility into Application Performance
> APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
> Monitor end-to-end web transactions and take corrective actions now
> Troubleshoot faster and improve end-user experience. Signup Now!
> http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140
> _______________________________________________
> SED-ML-discuss mailing list
> SED...@li...
> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>
>


-- 
Matthias König
Junior Group Leader LiSym - Systems Medicine of the Liver
Humboldt-University Berlin, Institute for Theoretical Biology
  https://www.livermetabolism.com
kon...@go...
Tel: +49 30 20938450
Tel: +49 176 81168480

Re: [SED-ML-discuss] How to interpret ComputeChange & FunctionalRange & curves of repeatedTask in L1V2

[Matthias K. <kon...@go...>]

Yes, a flow chart would help a lot. The implementation so far was a lot of
trial and error and took quit some time to figure out why things are like
they are in the spec. Only thing unclear right now is how to handle
recursive repeated tasks (repeated tasks which have repeated tasks as
subtasks).

Also a comprehensive list of examples with the expected outputs would be
very helpful for implementation (plot2D, plot3D and reports). Which tools
beside SED-ML webtools provide full implementation of the specification
(with SBML models at least)? Preferable once which can handle combine
archives as input?
I want: Combine Archive -> Tool -> outputs
M


On Wed, Mar 2, 2016 at 11:04 AM, Martin Peters <
acc...@fr...> wrote:

> Hello everybody,
>
> Lately I struggled on a similar issue as Matthias. Thankfully Dagmar
> could answer my questions. But may I suggest to include a flow chart (or
> similar), describing how to parse the Sed-ML in the spec?
> To be more precise a graphical explanation of when to load models, apply
> changes, calculate MathML etc. would have helped me a lot with
> understanding the spec in detail.
>
> But thanks for the great work so far and sorry for being slightly OT.
>
> Cheers
> Martin Peters
>
> On 2016-03-01 20:34, Matthias König wrote:
> > Hi Jonathan,
> > thanks a lot.
> > I handled [1] like you mentioned and was even easier than creating the
> > InitialAssignments.
> > [2] thanks. solved. Somehow overread it in the spec.
> > [3] did not think about multiple axis. This solves it. The plot has to
> > generate multiple axis in the case of mixed lin and log
> >
> > Are there any Combine archives (SBML+SEDML) examples for the use of [A]
> > FunctionalRange or [B] DataGenerators which are calculated via complex
> > formulas (something like piecewise, min, max, ...)
> >
> > @Lucian Can phrasedml generated [A] or [B] examples
> >
> > Matthias
> >
> >
> > On Tue, Mar 1, 2016 at 1:14 PM, Jonathan Cooper
> > <jon...@cs... <mailto:jon...@cs...>>
> wrote:
> >
> >     Hi Matthias,
> >
> >     For 1, ComputeChange is not quite identical to an
> >     InitialAssignmentRule as it happens when the model is set up, not
> >     when a simulation starts. So model3 should have S2=4.3 inherited
> >     from model2.
> >
> >     For 2, the intent is that FunctionalRange is calculated at each
> >     iteration, not all upfront.  I quote from p52:
> >
> >     "In the listOfVariables, variable elements de fine identifiers
> >     referring to model variables or range values, which may then be used
> >     within the math expression.  These references always retrieve the
> >     current value of the model variable or range at the current
> >     iteration of the enclosing repeatedTask.  For a model not being
> >     simulated by any subTask, the initial state of the model is used."
> >
> >     For 3, I'm not sure of the original reasoning, but it's certainly
> >     not uncommon to have multiple y axes displayed on a plot, one of
> >     which might be log scale and one linear.  Having different x axes is
> >     less likely to arise.
> >
> >     As Andre mentioned in his email, 4 is not defined by the current
> >     specification, as there is no explicit support for multidimensional
> >     arrays. This may come with the integration of NuML and parameter
> >     estimation in the next version(s).
> >
> >     Best wishes,
> >     Jonathan
> >
> >
> >     On 01/03/2016 09:59, Matthias König wrote:
> >>     Hi all,
> >>
> >>     I am currently implementing SEDML support for tellurium and a few
> >>     things in the SEDML specification are unclear for me.
> >>
> >>     [1] ComputeChange
> >>
> >>     The specification says: "The ComputeChange class permits to change
> >>     the numerical value of any element or attribute of a model based
> >>     on a calculation."
> >>     For me it is completely unclear when this calculation is
> >>     performed. Depending on the time of the actual evaluation
> >>     different numerical values are set in the model. After discussion
> >>     with Lucian I understood the ComputeChange as an
> >>     InitialAssignmentRule in SBML.
> >>     But what happens if later model changes (of derived models) change
> >>     the variables used in the ComputeChange?
> >>
> >>     Here an example:
> >>     It is not clear what should happen if for instance later model
> >>     versions change variables which are used in the calculation of the
> >>     computeChange, i.e. something like|
> >>     model1 = model "test"
> >>     model2 = model model1 with S1=0.3, S2=S1+4|
> >>     |model3 = model model2 with S1=10
> >>     |
> >>     What is the value of S2 in model 3?
> >>     If S2=S1+4 was impmented via an InitialAssignment than init(S2) in
> >>     model3 is 14 instead of 4.3 (which was the computeChange set in
> >>     model2). In my understanding model3 should have an S2=4.3, because
> >>     the computeChange should be calculated based on the state of model2.
> >>     So what is the correct interpretation (S2=14 or S2=4.3 in model3).
> >>     It has to be clearly stated when the calculation of the
> >>     ComputeChange has to be performed (within every model or after all
> >>     changes to all models)?
> >>
> >>     [2] FunctionalRange
> >>     Similar issue here. When is the computation of the functional
> >>     range performed? For instance in a repeatedTask where with every
> >>     repeat values of variables used in the calculation of the
> >>     functional range change.
> >>     Is the complete functional range calculated
> >>     - at the beginning of the loop with the initial model state before
> >>     any repeat
> >>     - or with the actual variables of the model at the beginning of
> >>     every iteration of the loop (which could change with reset=false)
> >>     Both variants would give very differnent numerical values for the
> >>     range.
> >>
> >>     [3] logX/logY/logZ on curves
> >>     Why are the logX,Y,Z not on the Plot, but on the curves?
> >>     What should be done if multiple curves have contradicting log
> >>     settings in a plot? I.e. two curves which have the same
> >>     xDataReference but different log settings.
> >>     <plot2D id="plot_0" name="Example plot"> <listOfCurves> <curve
> >>     logX="false" logY="false" xDataReference="plot_0_0_0"
> >>     yDataReference="plot_0_0_1"/> <curve logX="true" logY="false"
> >>     xDataReference="plot_0_0_0"
> >>     yDataReference="plot_0_1_1"/></listOfCurves>
> >>     [4] How to plot curves of repeatedTasks?
> >>     I could not really find any information on how to plot curves of
> >>     repeatedTasks. I interpreted it as following:
> >>     - when reset=True plot every repeat as a single curve in the plot
> >>     (i.e. Nrep lines per curve)
> >>     - when reset=False concatenate the repeats and plot one line (the
> >>     repeats are time slices of the complete simulation)
> >>     Is this correct?
> >>     And how would I encode the respective other thing in sedml, i.e.
> >>     - a repeatedTask with reset=False where every repeat should be a
> >>     line in the plot
> >>     - a repeatedTask with reset=True which is concatenated
> >>
> >>
> >>     Matthias
> >>
> >>     --
> >>     Matthias König
> >>     Junior Group Leader LiSym - Systems Medicine of the Liver
> >>     Humboldt-University Berlin, Institute for Theoretical Biology
> >>       <https://www.livermetabolism.com>https://www.livermetabolism.com
> >>     <mailto:kon...@go...>kon...@go...
> >>     <mailto:kon...@go...>
> >>     Tel: +49 30 20938450
> >>     Tel: +49 176 81168480 <tel:%2B49%20176%2081168480>
> >>
> >>
> >>
>  ------------------------------------------------------------------------------
> >>     Site24x7 APM Insight: Get Deep Visibility into Application
> Performance
> >>     APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
> >>     Monitor end-to-end web transactions and take corrective actions now
> >>     Troubleshoot faster and improve end-user experience. Signup Now!
> >>     http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140
> >>
> >>
> >>     _______________________________________________
> >>     SED-ML-discuss mailing list
> >>     SED...@li...
> >>     <mailto:SED...@li...>
> >>     https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
> >
> >
> >
>  ------------------------------------------------------------------------------
> >     Site24x7 APM Insight: Get Deep Visibility into Application
> Performance
> >     APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
> >     Monitor end-to-end web transactions and take corrective actions now
> >     Troubleshoot faster and improve end-user experience. Signup Now!
> >     http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140
> >     _______________________________________________
> >     SED-ML-discuss mailing list
> >     SED...@li...
> >     <mailto:SED...@li...>
> >     https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
> >
> >
> >
> >
> > --
> > Matthias König
> > Junior Group Leader LiSym - Systems Medicine of the Liver
> > Humboldt-University Berlin, Institute for Theoretical Biology
> >   https://www.livermetabolism.com
> > kon...@go... <mailto:kon...@go...>
> > Tel: +49 30 20938450
> > Tel: +49 176 81168480
> >
> >
> >
> ------------------------------------------------------------------------------
> > Site24x7 APM Insight: Get Deep Visibility into Application Performance
> > APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
> > Monitor end-to-end web transactions and take corrective actions now
> > Troubleshoot faster and improve end-user experience. Signup Now!
> > http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140
> >
> >
> >
> > _______________________________________________
> > SED-ML-discuss mailing list
> > SED...@li...
> > https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
> >
>
>
> ------------------------------------------------------------------------------
> Site24x7 APM Insight: Get Deep Visibility into Application Performance
> APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
> Monitor end-to-end web transactions and take corrective actions now
> Troubleshoot faster and improve end-user experience. Signup Now!
> http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140
> _______________________________________________
> SED-ML-discuss mailing list
> SED...@li...
> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>



-- 
Matthias König
Junior Group Leader LiSym - Systems Medicine of the Liver
Humboldt-University Berlin, Institute for Theoretical Biology
  https://www.livermetabolism.com
kon...@go...
Tel: +49 30 20938450
Tel: +49 176 81168480

Re: [SED-ML-discuss] How to interpret ComputeChange & FunctionalRange & curves of repeatedTask in L1V2

[Martin P. <acc...@fr...>]

Hello everybody,

Lately I struggled on a similar issue as Matthias. Thankfully Dagmar
could answer my questions. But may I suggest to include a flow chart (or
similar), describing how to parse the Sed-ML in the spec?
To be more precise a graphical explanation of when to load models, apply
changes, calculate MathML etc. would have helped me a lot with
understanding the spec in detail.

But thanks for the great work so far and sorry for being slightly OT.

Cheers
Martin Peters

On 2016-03-01 20:34, Matthias König wrote:
> Hi Jonathan,
> thanks a lot.
> I handled [1] like you mentioned and was even easier than creating the
> InitialAssignments.
> [2] thanks. solved. Somehow overread it in the spec.
> [3] did not think about multiple axis. This solves it. The plot has to
> generate multiple axis in the case of mixed lin and log
> 
> Are there any Combine archives (SBML+SEDML) examples for the use of [A]
> FunctionalRange or [B] DataGenerators which are calculated via complex
> formulas (something like piecewise, min, max, ...)
> 
> @Lucian Can phrasedml generated [A] or [B] examples
> 
> Matthias
> 
> 
> On Tue, Mar 1, 2016 at 1:14 PM, Jonathan Cooper
> <jon...@cs... <mailto:jon...@cs...>> wrote:
> 
>     Hi Matthias,
> 
>     For 1, ComputeChange is not quite identical to an
>     InitialAssignmentRule as it happens when the model is set up, not
>     when a simulation starts. So model3 should have S2=4.3 inherited
>     from model2.
> 
>     For 2, the intent is that FunctionalRange is calculated at each
>     iteration, not all upfront.  I quote from p52:
> 
>     "In the listOfVariables, variable elements define identifiers
>     referring to model variables or range values, which may then be used
>     within the math expression.  These references always retrieve the
>     current value of the model variable or range at the current
>     iteration of the enclosing repeatedTask.  For a model not being
>     simulated by any subTask, the initial state of the model is used."
> 
>     For 3, I'm not sure of the original reasoning, but it's certainly
>     not uncommon to have multiple y axes displayed on a plot, one of
>     which might be log scale and one linear.  Having different x axes is
>     less likely to arise.
> 
>     As Andre mentioned in his email, 4 is not defined by the current
>     specification, as there is no explicit support for multidimensional
>     arrays. This may come with the integration of NuML and parameter
>     estimation in the next version(s).
> 
>     Best wishes,
>     Jonathan
> 
> 
>     On 01/03/2016 09:59, Matthias König wrote:
>>     Hi all,
>>
>>     I am currently implementing SEDML support for tellurium and a few
>>     things in the SEDML specification are unclear for me.
>>
>>     [1] ComputeChange
>>
>>     The specification says: "The ComputeChange class permits to change
>>     the numerical value of any element or attribute of a model based
>>     on a calculation."
>>     For me it is completely unclear when this calculation is
>>     performed. Depending on the time of the actual evaluation
>>     different numerical values are set in the model. After discussion
>>     with Lucian I understood the ComputeChange as an
>>     InitialAssignmentRule in SBML.
>>     But what happens if later model changes (of derived models) change
>>     the variables used in the ComputeChange?
>>
>>     Here an example:
>>     It is not clear what should happen if for instance later model
>>     versions change variables which are used in the calculation of the
>>     computeChange, i.e. something like|
>>     model1 = model "test"
>>     model2 = model model1 with S1=0.3, S2=S1+4|
>>     |model3 = model model2 with S1=10
>>     |
>>     What is the value of S2 in model 3?
>>     If S2=S1+4 was impmented via an InitialAssignment than init(S2) in
>>     model3 is 14 instead of 4.3 (which was the computeChange set in
>>     model2). In my understanding model3 should have an S2=4.3, because
>>     the computeChange should be calculated based on the state of model2.
>>     So what is the correct interpretation (S2=14 or S2=4.3 in model3).
>>     It has to be clearly stated when the calculation of the
>>     ComputeChange has to be performed (within every model or after all
>>     changes to all models)?
>>
>>     [2] FunctionalRange
>>     Similar issue here. When is the computation of the functional
>>     range performed? For instance in a repeatedTask where with every
>>     repeat values of variables used in the calculation of the
>>     functional range change.
>>     Is the complete functional range calculated
>>     - at the beginning of the loop with the initial model state before
>>     any repeat
>>     - or with the actual variables of the model at the beginning of
>>     every iteration of the loop (which could change with reset=false)
>>     Both variants would give very differnent numerical values for the
>>     range.
>>
>>     [3] logX/logY/logZ on curves
>>     Why are the logX,Y,Z not on the Plot, but on the curves?
>>     What should be done if multiple curves have contradicting log
>>     settings in a plot? I.e. two curves which have the same
>>     xDataReference but different log settings.
>>     <plot2D id="plot_0" name="Example plot"> <listOfCurves> <curve
>>     logX="false" logY="false" xDataReference="plot_0_0_0"
>>     yDataReference="plot_0_0_1"/> <curve logX="true" logY="false"
>>     xDataReference="plot_0_0_0"
>>     yDataReference="plot_0_1_1"/></listOfCurves>
>>     [4] How to plot curves of repeatedTasks?
>>     I could not really find any information on how to plot curves of
>>     repeatedTasks. I interpreted it as following:
>>     - when reset=True plot every repeat as a single curve in the plot
>>     (i.e. Nrep lines per curve)
>>     - when reset=False concatenate the repeats and plot one line (the
>>     repeats are time slices of the complete simulation)
>>     Is this correct?
>>     And how would I encode the respective other thing in sedml, i.e.
>>     - a repeatedTask with reset=False where every repeat should be a
>>     line in the plot
>>     - a repeatedTask with reset=True which is concatenated
>>
>>
>>     Matthias
>>
>>     -- 
>>     Matthias König
>>     Junior Group Leader LiSym - Systems Medicine of the Liver
>>     Humboldt-University Berlin, Institute for Theoretical Biology
>>       <https://www.livermetabolism.com>https://www.livermetabolism.com
>>     <mailto:kon...@go...>kon...@go...
>>     <mailto:kon...@go...>
>>     Tel: +49 30 20938450
>>     Tel: +49 176 81168480 <tel:%2B49%20176%2081168480>
>>
>>
>>     ------------------------------------------------------------------------------
>>     Site24x7 APM Insight: Get Deep Visibility into Application Performance
>>     APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
>>     Monitor end-to-end web transactions and take corrective actions now
>>     Troubleshoot faster and improve end-user experience. Signup Now!
>>     http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140
>>
>>
>>     _______________________________________________
>>     SED-ML-discuss mailing list
>>     SED...@li...
>>     <mailto:SED...@li...>
>>     https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
> 
> 
>     ------------------------------------------------------------------------------
>     Site24x7 APM Insight: Get Deep Visibility into Application Performance
>     APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
>     Monitor end-to-end web transactions and take corrective actions now
>     Troubleshoot faster and improve end-user experience. Signup Now!
>     http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140
>     _______________________________________________
>     SED-ML-discuss mailing list
>     SED...@li...
>     <mailto:SED...@li...>
>     https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
> 
> 
> 
> 
> -- 
> Matthias König
> Junior Group Leader LiSym - Systems Medicine of the Liver
> Humboldt-University Berlin, Institute for Theoretical Biology
>   https://www.livermetabolism.com
> kon...@go... <mailto:kon...@go...>
> Tel: +49 30 20938450
> Tel: +49 176 81168480
> 
> 
> ------------------------------------------------------------------------------
> Site24x7 APM Insight: Get Deep Visibility into Application Performance
> APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
> Monitor end-to-end web transactions and take corrective actions now
> Troubleshoot faster and improve end-user experience. Signup Now!
> http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140
> 
> 
> 
> _______________________________________________
> SED-ML-discuss mailing list
> SED...@li...
> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
> 

Re: [SED-ML-discuss] L1V1 -> L1V2 conversion & applyChanges

[Frank B. <fbe...@ca...>]

Hello Matthias, 

it took a bit longer, but i finished that command line tool for getting the processed models or to convert between SED-ML Script and SEDML. Details are in the attached PDF. 

I have created a Ubuntu 15.10 x64 binary (that contains only files needed for the functions mentioned before): 
https://sourceforge.net/projects/libsedml/files/Ubuntu15.10-x64-sedml-console-only.tar.gz/download

and a new Windows installer, that also contains the full CellML API and will allow you to simulate the models just like the SED-ML Web Tools do, along with the script editor: 
https://sourceforge.net/projects/libsedml/files/SetupSedML-win32-1.22.exe/download

I hope that helps, 
best
Frank

Re: [SED-ML-discuss] How to interpret ComputeChange & FunctionalRange & curves of repeatedTask in L1V2

[Matthias K. <kon...@go...>]

Hi Jonathan,
thanks a lot.
I handled [1] like you mentioned and was even easier than creating the
InitialAssignments.
[2] thanks. solved. Somehow overread it in the spec.
[3] did not think about multiple axis. This solves it. The plot has to
generate multiple axis in the case of mixed lin and log

Are there any Combine archives (SBML+SEDML) examples for the use of [A]
FunctionalRange or [B] DataGenerators which are calculated via complex
formulas (something like piecewise, min, max, ...)

@Lucian Can phrasedml generated [A] or [B] examples

Matthias


On Tue, Mar 1, 2016 at 1:14 PM, Jonathan Cooper <jon...@cs...
> wrote:

> Hi Matthias,
>
> For 1, ComputeChange is not quite identical to an InitialAssignmentRule as
> it happens when the model is set up, not when a simulation starts. So
> model3 should have S2=4.3 inherited from model2.
>
> For 2, the intent is that FunctionalRange is calculated at each iteration,
> not all upfront.  I quote from p52:
>
> "In the listOfVariables, variable elements de fine identifiers referring
> to model variables or range values, which may then be used within the math
> expression.  These references always retrieve the current value of the
> model variable or range at the current iteration of the enclosing
> repeatedTask.  For a model not being simulated by any subTask, the initial
> state of the model is used."
>
> For 3, I'm not sure of the original reasoning, but it's certainly not
> uncommon to have multiple y axes displayed on a plot, one of which might be
> log scale and one linear.  Having different x axes is less likely to arise.
>
> As Andre mentioned in his email, 4 is not defined by the current
> specification, as there is no explicit support for multidimensional arrays.
> This may come with the integration of NuML and parameter estimation in the
> next version(s).
>
> Best wishes,
> Jonathan
>
>
> On 01/03/2016 09:59, Matthias König wrote:
>
> Hi all,
>
> I am currently implementing SEDML support for tellurium and a few things
> in the SEDML specification are unclear for me.
>
> [1] ComputeChange
>
> The specification says: "The ComputeChange class permits to change the
> numerical value of any element or attribute of a model based on a
> calculation."
> For me it is completely unclear when this calculation is performed.
> Depending on the time of the actual evaluation different numerical values
> are set in the model. After discussion with Lucian I understood the
> ComputeChange as an InitialAssignmentRule in SBML.
> But what happens if later model changes (of derived models) change the
> variables used in the ComputeChange?
>
> Here an example:
> It is not clear what should happen if for instance later model versions
> change variables which are used in the calculation of the computeChange,
> i.e. something like
> model1 = model "test"
> model2 = model model1 with S1=0.3, S2=S1+4
> model3 = model model2 with S1=10
> What is the value of S2 in model 3?
> If S2=S1+4 was impmented via an InitialAssignment than init(S2) in model3
> is 14 instead of 4.3 (which was the computeChange set in model2). In my
> understanding model3 should have an S2=4.3, because the computeChange
> should be calculated based on the state of model2.
> So what is the correct interpretation (S2=14 or S2=4.3 in model3).
> It has to be clearly stated when the calculation of the ComputeChange has
> to be performed (within every model or after all changes to all models)?
>
> [2] FunctionalRange
> Similar issue here. When is the computation of the functional range
> performed? For instance in a repeatedTask where with every repeat values of
> variables used in the calculation of the functional range change.
> Is the complete functional range calculated
> - at the beginning of the loop with the initial model state before any
> repeat
> - or with the actual variables of the model at the beginning of every
> iteration of the loop (which could change with reset=false)
> Both variants would give very differnent numerical values for the range.
>
> [3] logX/logY/logZ on curves
> Why are the logX,Y,Z not on the Plot, but on the curves?
> What should be done if multiple curves have contradicting log settings in
> a plot? I.e. two curves which have the same xDataReference but different
> log settings.
>
> <plot2D id="plot_0" name="Example plot">  <listOfCurves>    <curve logX="false" logY="false" xDataReference="plot_0_0_0" yDataReference="plot_0_0_1"/>    <curve logX="true" logY="false" xDataReference="plot_0_0_0" yDataReference="plot_0_1_1"/>  </listOfCurves>
>
> [4] How to plot curves of repeatedTasks?
> I could not really find any information on how to plot curves of
> repeatedTasks. I interpreted it as following:
> - when reset=True plot every repeat as a single curve in the plot (i.e.
> Nrep lines per curve)
> - when reset=False concatenate the repeats and plot one line (the repeats
> are time slices of the complete simulation)
> Is this correct?
> And how would I encode the respective other thing in sedml, i.e.
> - a repeatedTask with reset=False where every repeat should be a line in
> the plot
> - a repeatedTask with reset=True which is concatenated
>
>
> Matthias
>
> --
> Matthias König
> Junior Group Leader LiSym - Systems Medicine of the Liver
> Humboldt-University Berlin, Institute for Theoretical Biology
>   <https://www.livermetabolism.com>https://www.livermetabolism.com
> <kon...@go...>kon...@go...
> Tel: +49 30 20938450
> Tel: +49 176 81168480
>
>
> ------------------------------------------------------------------------------
> Site24x7 APM Insight: Get Deep Visibility into Application Performance
> APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
> Monitor end-to-end web transactions and take corrective actions now
> Troubleshoot faster and improve end-user experience. Signup Now!http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140
>
>
>
> _______________________________________________
> SED-ML-discuss mailing lis...@li...https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>
>
>
>
> ------------------------------------------------------------------------------
> Site24x7 APM Insight: Get Deep Visibility into Application Performance
> APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
> Monitor end-to-end web transactions and take corrective actions now
> Troubleshoot faster and improve end-user experience. Signup Now!
> http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140
> _______________________________________________
> SED-ML-discuss mailing list
> SED...@li...
> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>
>


-- 
Matthias König
Junior Group Leader LiSym - Systems Medicine of the Liver
Humboldt-University Berlin, Institute for Theoretical Biology
  https://www.livermetabolism.com
kon...@go...
Tel: +49 30 20938450
Tel: +49 176 81168480

Re: [SED-ML-discuss] How to interpret ComputeChange & FunctionalRange & curves of repeatedTask in L1V2

[Jonathan C. <jon...@cs...>]

Hi Matthias,

For 1, ComputeChange is not quite identical to an InitialAssignmentRule 
as it happens when the model is set up, not when a simulation starts. So 
model3 should have S2=4.3 inherited from model2.

For 2, the intent is that FunctionalRange is calculated at each 
iteration, not all upfront.  I quote from p52:

"In the listOfVariables, variable elements define identifiers referring 
to model variables or range values, which may then be used within the 
math expression.  These references always retrieve the current value of 
the model variable or range at the current iteration of the enclosing 
repeatedTask.  For a model not being simulated by any subTask, the 
initial state of the model is used."

For 3, I'm not sure of the original reasoning, but it's certainly not 
uncommon to have multiple y axes displayed on a plot, one of which might 
be log scale and one linear.  Having different x axes is less likely to 
arise.

As Andre mentioned in his email, 4 is not defined by the current 
specification, as there is no explicit support for multidimensional 
arrays. This may come with the integration of NuML and parameter 
estimation in the next version(s).

Best wishes,
Jonathan

On 01/03/2016 09:59, Matthias König wrote:
> Hi all,
>
> I am currently implementing SEDML support for tellurium and a few 
> things in the SEDML specification are unclear for me.
>
> [1] ComputeChange
>
> The specification says: "The ComputeChange class permits to change the 
> numerical value of any element or attribute of a model based on a 
> calculation."
> For me it is completely unclear when this calculation is performed. 
> Depending on the time of the actual evaluation different numerical 
> values are set in the model. After discussion with Lucian I understood 
> the ComputeChange as an InitialAssignmentRule in SBML.
> But what happens if later model changes (of derived models) change the 
> variables used in the ComputeChange?
>
> Here an example:
> It is not clear what should happen if for instance later model 
> versions change variables which are used in the calculation of the 
> computeChange, i.e. something like|
> model1 = model "test"
> model2 = model model1 with S1=0.3, S2=S1+4|
> |model3 = model model2 with S1=10
> |
> What is the value of S2 in model 3?
> If S2=S1+4 was impmented via an InitialAssignment than init(S2) in 
> model3 is 14 instead of 4.3 (which was the computeChange set in 
> model2). In my understanding model3 should have an S2=4.3, because the 
> computeChange should be calculated based on the state of model2.
> So what is the correct interpretation (S2=14 or S2=4.3 in model3).
> It has to be clearly stated when the calculation of the ComputeChange 
> has to be performed (within every model or after all changes to all 
> models)?
>
> [2] FunctionalRange
> Similar issue here. When is the computation of the functional range 
> performed? For instance in a repeatedTask where with every repeat 
> values of variables used in the calculation of the functional range 
> change.
> Is the complete functional range calculated
> - at the beginning of the loop with the initial model state before any 
> repeat
> - or with the actual variables of the model at the beginning of every 
> iteration of the loop (which could change with reset=false)
> Both variants would give very differnent numerical values for the range.
>
> [3] logX/logY/logZ on curves
> Why are the logX,Y,Z not on the Plot, but on the curves?
> What should be done if multiple curves have contradicting log settings 
> in a plot? I.e. two curves which have the same xDataReference but 
> different log settings.
> <plot2D id="plot_0" name="Example plot"> <listOfCurves> <curve 
> logX="false" logY="false" xDataReference="plot_0_0_0" 
> yDataReference="plot_0_0_1"/> <curve logX="true" logY="false" 
> xDataReference="plot_0_0_0" yDataReference="plot_0_1_1"/></listOfCurves>
> [4] How to plot curves of repeatedTasks?
> I could not really find any information on how to plot curves of 
> repeatedTasks. I interpreted it as following:
> - when reset=True plot every repeat as a single curve in the plot 
> (i.e. Nrep lines per curve)
> - when reset=False concatenate the repeats and plot one line (the 
> repeats are time slices of the complete simulation)
> Is this correct?
> And how would I encode the respective other thing in sedml, i.e.
> - a repeatedTask with reset=False where every repeat should be a line 
> in the plot
> - a repeatedTask with reset=True which is concatenated
>
>
> Matthias
>
> -- 
> Matthias König
> Junior Group Leader LiSym - Systems Medicine of the Liver
> Humboldt-University Berlin, Institute for Theoretical Biology
> https://www.livermetabolism.com
> kon...@go... <mailto:kon...@go...>
> Tel: +49 30 20938450
> Tel: +49 176 81168480
>
>
> ------------------------------------------------------------------------------
> Site24x7 APM Insight: Get Deep Visibility into Application Performance
> APM + Mobile APM + RUM: Monitor 3 App instances at just $35/Month
> Monitor end-to-end web transactions and take corrective actions now
> Troubleshoot faster and improve end-user experience. Signup Now!
> http://pubads.g.doubleclick.net/gampad/clk?id=272487151&iu=/4140
>
>
> _______________________________________________
> SED-ML-discuss mailing list
> SED...@li...
> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss

Re: [SED-ML-discuss] How to interpret ComputeChange & FunctionalRange & curves of repeatedTask in L1V2

[David N. <dav...@gm...>]

Hi Matthias,

Apologies for leaving the "hard" questions for now, but here is a quick
response to your final query:

[4] How to plot curves of repeatedTasks?
> I could not really find any information on how to plot curves of
> repeatedTasks. I interpreted it as following:
> - when reset=True plot every repeat as a single curve in the plot (i.e.
> Nrep lines per curve)
> - when reset=False concatenate the repeats and plot one line (the repeats
> are time slices of the complete simulation)
> Is this correct?
> And how would I encode the respective other thing in sedml, i.e.
> - a repeatedTask with reset=False where every repeat should be a line in
> the plot
> - a repeatedTask with reset=True which is concatenated
>

There is nothing in the spec regarding plots and repeated tasks. The
interpretation of this is completely application dependent. My suggestion
would be that "correct" is to follow the example set by Frank with his C#
libsedml code and the SED-ML web tools. That seemed intuitive to me and
gave the behaviour I expected. I can't recall off the top of my head if he
uses the reset flag in any way when drawing the plots.

Your hard questions seem similar to some Lucian was asking on the list
prior to COMBINE last year, so hopefully he will speak up with what he
discovered back then.

Cheers,
David.