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

Re: [SED-ML-discuss] Functional ranges

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

The reason for including the functionalRange, was that we already had all the complexity in SED-ML, and so we could as well use it. It was meant for all those situations where it is easier to write down a close form expression that describes the range. Certainly you could in many cases achieve the same with other ranges, but perhaps not always. 

One example would be running a repeated task until some condition is met in your model. For that you will not be able to produce a vectorRange or uniformRange beforehand, as you don’t know how many steps it would take.  

Frank

> On Jun 18, 2015, at 8:30 PM, Lucian Smith <luc...@gm...> wrote:
> 
> OK, I guess an example would help.
> 
> Here is a very simple example with a functional range.  A vector range is used by a functional range, which is used by a setValue:
> 
> <repeatedTask id="task3" resetModel="false" range="function">
>  <listOfRanges>
>   <vectorRange id="vector">
>    <value> 1 </value>
>    <value> 4 </value>
>    <value> 10 </value>
>   </vectorRange>
>   <functionalRange id="function" range="vector">
>    <listOfVariables>
>     <variable id="w" name="current parameter value" modelReference="model2"
>        target="/s:sbml/s:model/s:listOfParameters/s:parameter[@id=’w’]" />
>    </listOfVariables>
>    <math xmlns="http://www.w3.org/1998/Math/MathML">
>     <apply>
>      <times/>
>      <ci> w </ci>
>      <ci> vector </ci>
>     </apply>
>    </math>
>   </functionalRange>
>  </listOfRanges>
>  <listOfChanges>
>   <setValue target="/s:sbml/s:model/s:listOfParameters/s:parameter[@id='w']" modelReference="model1" range="function">
>    <math xmlns="http://www.w3.org/1998/Math/MathML">
>     <ci> function </ci>
>    </math>
>   </setValue>
>  </listOfChanges>
>  <listOfSubTasks>
>   <subTask task="task1" />
>  </listOfSubTasks>
> </repeatedTask>
> 
> And in this example, we accomplish exactly the same thing without using a functional range at all.  Instead, the same vectorRange is simply used directly by the setValue:
> 
> <repeatedTask id="task3" resetModel="false" range="vector">
>  <listOfRanges>
>   <vectorRange id="vector">
>    <value> 1 </value>
>    <value> 4 </value>
>    <value> 10 </value>
>   </vectorRange>
>  </listOfRanges>
>  <listOfChanges>
>   <setValue target="/s:sbml/s:model/s:listOfParameters/s:parameter[@id='w']" modelReference="model1" range="vector">
>    <listOfVariables>
>     <variable id="w" name="current parameter value" modelReference="model2"
>        target="/s:sbml/s:model/s:listOfParameters/s:parameter[@id=’w’]" />
>    </listOfVariables>
>    <math xmlns="http://www.w3.org/1998/Math/MathML">
>     <apply>
>      <times/>
>      <ci> w </ci>
>      <ci> vector </ci>
>     </apply>
>    </math>
>   </setValue>
>  </listOfChanges>
>  <listOfSubTasks>
>   <subTask task="task1" />
>  </listOfSubTasks>
> </repeatedTask>
> 
> 
> It seems to me that you could always do this, even for more complicated examples.  All a functionalRange seems to do is wrap up some stuff in a function, which has the scope of that individual repeatedTask.  If you expand the function, you could always translate it to disappear entirely, subsumed into the 'setValue'. This is in contrast to the uniformRange and the vectorRange constructs, which are the only ways to get an actual vector into anything, and therefore are crucial.  Basically, it seems to me that the 'functionalRange' is a really complicated way to get a slightly more global variable into a single repeated task.  Am I wrong?  Does it have other functionality I'm not seeing?  Are the two example models I have above not actually equivalent?
> 
> -Lucian
> 
> On Thu, Jun 18, 2015 at 1:10 AM, Frank Bergmann <fbe...@ca...> wrote:
> >
> > Do they function any differently from the Variable construct?  (Apart from being slightly more global?)
> 
> functional ranges, just like all the other classes that involve computation contain a listOfVariables, and a listOfParameters that can be used in the math expressions. Maybe you can expand at what issue you have, and I can provide more information.
> 
> Frank
> 
> >
> > On Wed, Jun 17, 2015 at 11:44 PM, Frank Bergmann <fbe...@ca...> wrote:
> > >
> > > OK, so why have functional ranges at all, then?  What purpose do they serve?  Is it solely to provide a globally-defined variable for the other children of the repeated task?
> > >
> >
> > functional ranges, are there for all ranges that you want to calculate explicitly based on a compact equation, perhaps involving the state of the model they are running over, so that you can compute the nth element of the range. And yes, the range can then be used in other constructs as an already computed value.
> >
> > best
> > Frank
> >
> > > -Lucian
> > >
> > > On Wed, Jun 17, 2015 at 11:10 PM, Frank Bergmann <fbe...@ca...> wrote:
> > > SED-ML uses the same subset of mathml as covered by SBML L3V1 core, augmented by special functions min, max sum and product. This is described in section 2.2.1 … that means that MathML vectors and matrix operations are currently not supported.
> > >
> > > Cheers
> > > Frank
> > >
> > > > On Jun 17, 2015, at 10:30 PM, Lucian Smith <luc...@gm...> wrote:
> > > >
> > > > I have a question about functional ranges.  From the examples, all of the MathML returns scalar values, unless it in turn references another range.  Is it possible for the function to return a vector on its own?  If so, how?  If not, what's the point of them existing in the first place?  Wouldn't they be exactly equivalent to a Variable defined for another SetValue in the same RepeatedTask?  Or is the point that you then only have to define it once, and use it in multiple RepeatedTasks?
> > > >
> > > > In general, is it legal to use vector constructs in MathML in SED-ML?  If so, would these be equivalent:
> > > >
> > > > <vectorRange id="current">
> > > >  <value> 1 </value>
> > > >  <value> 4 </value>
> > > >  <value> 10 </value>
> > > > </vectorRange>
> > > >
> > > > and:
> > > >
> > > > <functionalRange id="current" >
> > > >   <math xmlns="http://www.w3.org/1998/Math/MathML">
> > > >     <vector>
> > > >       <cn> 1 </cn>
> > > >       <cn> 4 </cn>
> > > >       <cn> 10 </cn>
> > > >     </vector>
> > > >   </math>
> > > > </functionalRange>
> > > > ------------------------------------------------------------------------------
> > > > _______________________________________________
> > > > SED-ML-discuss mailing list
> > > > SED...@li...
> > > > https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
> > >
> > >
> > > ------------------------------------------------------------------------------
> > > _______________________________________________
> > > SED-ML-discuss mailing list
> > > SED...@li...
> > > https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
> > >
> > > ------------------------------------------------------------------------------
> > > _______________________________________________
> > > SED-ML-discuss mailing list
> > > SED...@li...
> > > https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
> >
> >
> > ------------------------------------------------------------------------------
> > _______________________________________________
> > SED-ML-discuss mailing list
> > SED...@li...
> > https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
> >
> > ------------------------------------------------------------------------------
> > _______________________________________________
> > SED-ML-discuss mailing list
> > SED...@li...
> > https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
> 
> 
> ------------------------------------------------------------------------------
> _______________________________________________
> SED-ML-discuss mailing list
> SED...@li...
> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
> 
> ------------------------------------------------------------------------------
> _______________________________________________
> SED-ML-discuss mailing list
> SED...@li...
> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss

Re: [SED-ML-discuss] Functional ranges

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

OK, I guess an example would help.

Here is a very simple example with a functional range.  A vector range is
used by a functional range, which is used by a setValue:

<repeatedTask id="task3" resetModel="false" range="function">
 <listOfRanges>
  <vectorRange id="vector">
   <value> 1 </value>
   <value> 4 </value>
   <value> 10 </value>
  </vectorRange>
  <functionalRange id="function" range="vector">
   <listOfVariables>
    <variable id="w" name="current parameter value" modelReference="model2"
       target="/s:sbml/s:model/s:listOfParameters/s:parameter[@id=’w’]" />
   </listOfVariables>
   <math xmlns="http://www.w3.org/1998/Math/MathML">
    <apply>
     <times/>
     <ci> w </ci>
     <ci> vector </ci>
    </apply>
   </math>
  </functionalRange>
 </listOfRanges>
 <listOfChanges>
  <setValue
target="/s:sbml/s:model/s:listOfParameters/s:parameter[@id='w']"
modelReference="model1" range="function">
   <math xmlns="http://www.w3.org/1998/Math/MathML">
    <ci> function </ci>
   </math>
  </setValue>
 </listOfChanges>
 <listOfSubTasks>
  <subTask task="task1" />
 </listOfSubTasks>
</repeatedTask>

And in this example, we accomplish exactly the same thing without using a
functional range at all.  Instead, the same vectorRange is simply used
directly by the setValue:

<repeatedTask id="task3" resetModel="false" range="vector">
 <listOfRanges>
  <vectorRange id="vector">
   <value> 1 </value>
   <value> 4 </value>
   <value> 10 </value>
  </vectorRange>
 </listOfRanges>
 <listOfChanges>
  <setValue
target="/s:sbml/s:model/s:listOfParameters/s:parameter[@id='w']"
modelReference="model1" range="vector">
   <listOfVariables>
    <variable id="w" name="current parameter value" modelReference="model2"
       target="/s:sbml/s:model/s:listOfParameters/s:parameter[@id=’w’]" />
   </listOfVariables>
   <math xmlns="http://www.w3.org/1998/Math/MathML">
    <apply>
     <times/>
     <ci> w </ci>
     <ci> vector </ci>
    </apply>
   </math>
  </setValue>
 </listOfChanges>
 <listOfSubTasks>
  <subTask task="task1" />
 </listOfSubTasks>
</repeatedTask>


It seems to me that you could always do this, even for more complicated
examples.  All a functionalRange seems to do is wrap up some stuff in a
function, which has the scope of that individual repeatedTask.  If you
expand the function, you could always translate it to disappear entirely,
subsumed into the 'setValue'. This is in contrast to the uniformRange and
the vectorRange constructs, which are the only ways to get an actual vector
into anything, and therefore are crucial.  Basically, it seems to me that
the 'functionalRange' is a really complicated way to get a slightly more
global variable into a single repeated task.  Am I wrong?  Does it have
other functionality I'm not seeing?  Are the two example models I have
above not actually equivalent?

-Lucian

On Thu, Jun 18, 2015 at 1:10 AM, Frank Bergmann <fbe...@ca...>
wrote:

> >
> > Do they function any differently from the Variable construct?  (Apart
> from being slightly more global?)
>
> functional ranges, just like all the other classes that involve
> computation contain a listOfVariables, and a listOfParameters that can be
> used in the math expressions. Maybe you can expand at what issue you have,
> and I can provide more information.
>
> Frank
>
> >
> > On Wed, Jun 17, 2015 at 11:44 PM, Frank Bergmann <fbe...@ca...>
> wrote:
> > >
> > > OK, so why have functional ranges at all, then?  What purpose do they
> serve?  Is it solely to provide a globally-defined variable for the other
> children of the repeated task?
> > >
> >
> > functional ranges, are there for all ranges that you want to calculate
> explicitly based on a compact equation, perhaps involving the state of the
> model they are running over, so that you can compute the nth element of the
> range. And yes, the range can then be used in other constructs as an
> already computed value.
> >
> > best
> > Frank
> >
> > > -Lucian
> > >
> > > On Wed, Jun 17, 2015 at 11:10 PM, Frank Bergmann <fbe...@ca...>
> wrote:
> > > SED-ML uses the same subset of mathml as covered by SBML L3V1 core,
> augmented by special functions min, max sum and product. This is described
> in section 2.2.1 … that means that MathML vectors and matrix operations are
> currently not supported.
> > >
> > > Cheers
> > > Frank
> > >
> > > > On Jun 17, 2015, at 10:30 PM, Lucian Smith <
> luc...@gm...> wrote:
> > > >
> > > > I have a question about functional ranges.  From the examples, all
> of the MathML returns scalar values, unless it in turn references another
> range.  Is it possible for the function to return a vector on its own?  If
> so, how?  If not, what's the point of them existing in the first place?
> Wouldn't they be exactly equivalent to a Variable defined for another
> SetValue in the same RepeatedTask?  Or is the point that you then only have
> to define it once, and use it in multiple RepeatedTasks?
> > > >
> > > > In general, is it legal to use vector constructs in MathML in
> SED-ML?  If so, would these be equivalent:
> > > >
> > > > <vectorRange id="current">
> > > >  <value> 1 </value>
> > > >  <value> 4 </value>
> > > >  <value> 10 </value>
> > > > </vectorRange>
> > > >
> > > > and:
> > > >
> > > > <functionalRange id="current" >
> > > >   <math xmlns="http://www.w3.org/1998/Math/MathML">
> > > >     <vector>
> > > >       <cn> 1 </cn>
> > > >       <cn> 4 </cn>
> > > >       <cn> 10 </cn>
> > > >     </vector>
> > > >   </math>
> > > > </functionalRange>
> > > >
> ------------------------------------------------------------------------------
> > > > _______________________________________________
> > > > SED-ML-discuss mailing list
> > > > SED...@li...
> > > > https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
> > >
> > >
> > >
> ------------------------------------------------------------------------------
> > > _______________________________________________
> > > SED-ML-discuss mailing list
> > > SED...@li...
> > > https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
> > >
> > >
> ------------------------------------------------------------------------------
> > > _______________________________________________
> > > SED-ML-discuss mailing list
> > > SED...@li...
> > > https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
> >
> >
> >
> ------------------------------------------------------------------------------
> > _______________________________________________
> > SED-ML-discuss mailing list
> > SED...@li...
> > https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
> >
> >
> ------------------------------------------------------------------------------
> > _______________________________________________
> > SED-ML-discuss mailing list
> > SED...@li...
> > https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>
>
>
> ------------------------------------------------------------------------------
> _______________________________________________
> SED-ML-discuss mailing list
> SED...@li...
> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>

Re: [SED-ML-discuss] Functional ranges

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

> 
> Do they function any differently from the Variable construct?  (Apart from being slightly more global?)

functional ranges, just like all the other classes that involve computation contain a listOfVariables, and a listOfParameters that can be used in the math expressions. Maybe you can expand at what issue you have, and I can provide more information. 

Frank

> 
> On Wed, Jun 17, 2015 at 11:44 PM, Frank Bergmann <fbe...@ca...> wrote:
> >
> > OK, so why have functional ranges at all, then?  What purpose do they serve?  Is it solely to provide a globally-defined variable for the other children of the repeated task?
> >
> 
> functional ranges, are there for all ranges that you want to calculate explicitly based on a compact equation, perhaps involving the state of the model they are running over, so that you can compute the nth element of the range. And yes, the range can then be used in other constructs as an already computed value.
> 
> best
> Frank
> 
> > -Lucian
> >
> > On Wed, Jun 17, 2015 at 11:10 PM, Frank Bergmann <fbe...@ca...> wrote:
> > SED-ML uses the same subset of mathml as covered by SBML L3V1 core, augmented by special functions min, max sum and product. This is described in section 2.2.1 … that means that MathML vectors and matrix operations are currently not supported.
> >
> > Cheers
> > Frank
> >
> > > On Jun 17, 2015, at 10:30 PM, Lucian Smith <luc...@gm...> wrote:
> > >
> > > I have a question about functional ranges.  From the examples, all of the MathML returns scalar values, unless it in turn references another range.  Is it possible for the function to return a vector on its own?  If so, how?  If not, what's the point of them existing in the first place?  Wouldn't they be exactly equivalent to a Variable defined for another SetValue in the same RepeatedTask?  Or is the point that you then only have to define it once, and use it in multiple RepeatedTasks?
> > >
> > > In general, is it legal to use vector constructs in MathML in SED-ML?  If so, would these be equivalent:
> > >
> > > <vectorRange id="current">
> > >  <value> 1 </value>
> > >  <value> 4 </value>
> > >  <value> 10 </value>
> > > </vectorRange>
> > >
> > > and:
> > >
> > > <functionalRange id="current" >
> > >   <math xmlns="http://www.w3.org/1998/Math/MathML">
> > >     <vector>
> > >       <cn> 1 </cn>
> > >       <cn> 4 </cn>
> > >       <cn> 10 </cn>
> > >     </vector>
> > >   </math>
> > > </functionalRange>
> > > ------------------------------------------------------------------------------
> > > _______________________________________________
> > > SED-ML-discuss mailing list
> > > SED...@li...
> > > https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
> >
> >
> > ------------------------------------------------------------------------------
> > _______________________________________________
> > SED-ML-discuss mailing list
> > SED...@li...
> > https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
> >
> > ------------------------------------------------------------------------------
> > _______________________________________________
> > SED-ML-discuss mailing list
> > SED...@li...
> > https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
> 
> 
> ------------------------------------------------------------------------------
> _______________________________________________
> SED-ML-discuss mailing list
> SED...@li...
> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
> 
> ------------------------------------------------------------------------------
> _______________________________________________
> SED-ML-discuss mailing list
> SED...@li...
> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss

Re: [SED-ML-discuss] Functional ranges

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

Do they function any differently from the Variable construct?  (Apart from
being slightly more global?)

On Wed, Jun 17, 2015 at 11:44 PM, Frank Bergmann <fbe...@ca...>
wrote:

> >
> > OK, so why have functional ranges at all, then?  What purpose do they
> serve?  Is it solely to provide a globally-defined variable for the other
> children of the repeated task?
> >
>
> functional ranges, are there for all ranges that you want to calculate
> explicitly based on a compact equation, perhaps involving the state of the
> model they are running over, so that you can compute the nth element of the
> range. And yes, the range can then be used in other constructs as an
> already computed value.
>
> best
> Frank
>
> > -Lucian
> >
> > On Wed, Jun 17, 2015 at 11:10 PM, Frank Bergmann <fbe...@ca...>
> wrote:
> > SED-ML uses the same subset of mathml as covered by SBML L3V1 core,
> augmented by special functions min, max sum and product. This is described
> in section 2.2.1 … that means that MathML vectors and matrix operations are
> currently not supported.
> >
> > Cheers
> > Frank
> >
> > > On Jun 17, 2015, at 10:30 PM, Lucian Smith <luc...@gm...>
> wrote:
> > >
> > > I have a question about functional ranges.  From the examples, all of
> the MathML returns scalar values, unless it in turn references another
> range.  Is it possible for the function to return a vector on its own?  If
> so, how?  If not, what's the point of them existing in the first place?
> Wouldn't they be exactly equivalent to a Variable defined for another
> SetValue in the same RepeatedTask?  Or is the point that you then only have
> to define it once, and use it in multiple RepeatedTasks?
> > >
> > > In general, is it legal to use vector constructs in MathML in SED-ML?
> If so, would these be equivalent:
> > >
> > > <vectorRange id="current">
> > >  <value> 1 </value>
> > >  <value> 4 </value>
> > >  <value> 10 </value>
> > > </vectorRange>
> > >
> > > and:
> > >
> > > <functionalRange id="current" >
> > >   <math xmlns="http://www.w3.org/1998/Math/MathML">
> > >     <vector>
> > >       <cn> 1 </cn>
> > >       <cn> 4 </cn>
> > >       <cn> 10 </cn>
> > >     </vector>
> > >   </math>
> > > </functionalRange>
> > >
> ------------------------------------------------------------------------------
> > > _______________________________________________
> > > SED-ML-discuss mailing list
> > > SED...@li...
> > > https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
> >
> >
> >
> ------------------------------------------------------------------------------
> > _______________________________________________
> > SED-ML-discuss mailing list
> > SED...@li...
> > https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
> >
> >
> ------------------------------------------------------------------------------
> > _______________________________________________
> > SED-ML-discuss mailing list
> > SED...@li...
> > https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>
>
>
> ------------------------------------------------------------------------------
> _______________________________________________
> SED-ML-discuss mailing list
> SED...@li...
> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>

Re: [SED-ML-discuss] Functional ranges

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

> 
> OK, so why have functional ranges at all, then?  What purpose do they serve?  Is it solely to provide a globally-defined variable for the other children of the repeated task?
> 

functional ranges, are there for all ranges that you want to calculate explicitly based on a compact equation, perhaps involving the state of the model they are running over, so that you can compute the nth element of the range. And yes, the range can then be used in other constructs as an already computed value. 

best
Frank

> -Lucian
> 
> On Wed, Jun 17, 2015 at 11:10 PM, Frank Bergmann <fbe...@ca...> wrote:
> SED-ML uses the same subset of mathml as covered by SBML L3V1 core, augmented by special functions min, max sum and product. This is described in section 2.2.1 … that means that MathML vectors and matrix operations are currently not supported.
> 
> Cheers
> Frank
> 
> > On Jun 17, 2015, at 10:30 PM, Lucian Smith <luc...@gm...> wrote:
> >
> > I have a question about functional ranges.  From the examples, all of the MathML returns scalar values, unless it in turn references another range.  Is it possible for the function to return a vector on its own?  If so, how?  If not, what's the point of them existing in the first place?  Wouldn't they be exactly equivalent to a Variable defined for another SetValue in the same RepeatedTask?  Or is the point that you then only have to define it once, and use it in multiple RepeatedTasks?
> >
> > In general, is it legal to use vector constructs in MathML in SED-ML?  If so, would these be equivalent:
> >
> > <vectorRange id="current">
> >  <value> 1 </value>
> >  <value> 4 </value>
> >  <value> 10 </value>
> > </vectorRange>
> >
> > and:
> >
> > <functionalRange id="current" >
> >   <math xmlns="http://www.w3.org/1998/Math/MathML">
> >     <vector>
> >       <cn> 1 </cn>
> >       <cn> 4 </cn>
> >       <cn> 10 </cn>
> >     </vector>
> >   </math>
> > </functionalRange>
> > ------------------------------------------------------------------------------
> > _______________________________________________
> > SED-ML-discuss mailing list
> > SED...@li...
> > https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
> 
> 
> ------------------------------------------------------------------------------
> _______________________________________________
> SED-ML-discuss mailing list
> SED...@li...
> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
> 
> ------------------------------------------------------------------------------
> _______________________________________________
> SED-ML-discuss mailing list
> SED...@li...
> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss

Re: [SED-ML-discuss] Functional ranges

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

OK, so why have functional ranges at all, then?  What purpose do they
serve?  Is it solely to provide a globally-defined variable for the other
children of the repeated task?

-Lucian

On Wed, Jun 17, 2015 at 11:10 PM, Frank Bergmann <fbe...@ca...>
wrote:

> SED-ML uses the same subset of mathml as covered by SBML L3V1 core,
> augmented by special functions min, max sum and product. This is described
> in section 2.2.1 … that means that MathML vectors and matrix operations are
> currently not supported.
>
> Cheers
> Frank
>
> > On Jun 17, 2015, at 10:30 PM, Lucian Smith <luc...@gm...>
> wrote:
> >
> > I have a question about functional ranges.  From the examples, all of
> the MathML returns scalar values, unless it in turn references another
> range.  Is it possible for the function to return a vector on its own?  If
> so, how?  If not, what's the point of them existing in the first place?
> Wouldn't they be exactly equivalent to a Variable defined for another
> SetValue in the same RepeatedTask?  Or is the point that you then only have
> to define it once, and use it in multiple RepeatedTasks?
> >
> > In general, is it legal to use vector constructs in MathML in SED-ML?
> If so, would these be equivalent:
> >
> > <vectorRange id="current">
> >  <value> 1 </value>
> >  <value> 4 </value>
> >  <value> 10 </value>
> > </vectorRange>
> >
> > and:
> >
> > <functionalRange id="current" >
> >   <math xmlns="http://www.w3.org/1998/Math/MathML">
> >     <vector>
> >       <cn> 1 </cn>
> >       <cn> 4 </cn>
> >       <cn> 10 </cn>
> >     </vector>
> >   </math>
> > </functionalRange>
> >
> ------------------------------------------------------------------------------
> > _______________________________________________
> > SED-ML-discuss mailing list
> > SED...@li...
> > https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>
>
>
> ------------------------------------------------------------------------------
> _______________________________________________
> SED-ML-discuss mailing list
> SED...@li...
> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>

Re: [SED-ML-discuss] Functional ranges

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

SED-ML uses the same subset of mathml as covered by SBML L3V1 core, augmented by special functions min, max sum and product. This is described in section 2.2.1 … that means that MathML vectors and matrix operations are currently not supported. 

Cheers
Frank

> On Jun 17, 2015, at 10:30 PM, Lucian Smith <luc...@gm...> wrote:
> 
> I have a question about functional ranges.  From the examples, all of the MathML returns scalar values, unless it in turn references another range.  Is it possible for the function to return a vector on its own?  If so, how?  If not, what's the point of them existing in the first place?  Wouldn't they be exactly equivalent to a Variable defined for another SetValue in the same RepeatedTask?  Or is the point that you then only have to define it once, and use it in multiple RepeatedTasks?
> 
> In general, is it legal to use vector constructs in MathML in SED-ML?  If so, would these be equivalent:
> 
> <vectorRange id="current">
>  <value> 1 </value>
>  <value> 4 </value>
>  <value> 10 </value>
> </vectorRange>
> 
> and:
> 
> <functionalRange id="current" >
>   <math xmlns="http://www.w3.org/1998/Math/MathML">
>     <vector>
>       <cn> 1 </cn>
>       <cn> 4 </cn>
>       <cn> 10 </cn>
>     </vector>
>   </math>
> </functionalRange>
> ------------------------------------------------------------------------------
> _______________________________________________
> SED-ML-discuss mailing list
> SED...@li...
> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss

[SED-ML-discuss] Functional ranges

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

I have a question about functional ranges.  From the examples, all of the
MathML returns scalar values, unless it in turn references another range.
Is it possible for the function to return a vector on its own?  If so,
how?  If not, what's the point of them existing in the first place?
Wouldn't they be exactly equivalent to a Variable defined for another
SetValue in the same RepeatedTask?  Or is the point that you then only have
to define it once, and use it in multiple RepeatedTasks?

In general, is it legal to use vector constructs in MathML in SED-ML?  If
so, would these be equivalent:

<vectorRange id="current">
 <value> 1 </value>
 <value> 4 </value>
 <value> 10 </value>
</vectorRange>

and:

<functionalRange id="current" >
  <math xmlns="http://www.w3.org/1998/Math/MathML">
    <vector>
      <cn> 1 </cn>
      <cn> 4 </cn>
      <cn> 10 </cn>
    </vector>
  </math>
</functionalRange>

Re: [SED-ML-discuss] Repeated tasks with multiple subtasks

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

On Mon, Jun 8, 2015 at 2:59 PM, David Nickerson <dav...@gm...>
wrote:

> Hi Lucian,
>
> Sorry for the delay in replying!
>

No problem at all; I know my own response times to requests like these vary
widely.


> > Is there any requirement that the referenced model be the same model as
> is
> > referenced by the subtask (or one of the subtasks) in question?
>
> I don't think so. I my own code I just look for set value changes that
> match the "current" model as I iterate over the sub-tasks. If there
> was a change that referenced a model that is not used then I would
> simply ignore it. If I have understood your question correctly?
>

Yup, that was the question.  I think it would be nice if there was a
validation error that could catch these sorts of things, but I do
understand that adding validation rules would be a pretty significant
undertaking.


> Although I think my code would throw an error if the model reference
> matched but the xpath target expression failed to match a suitable
> target for setting the value.
>

Yeah, that'd be another validation issue that would be nice to formalize.

> And in fact, are there any validation rules planned at all?  There don't
> > seem to be any in the spec; I don't know if there are any in any of the
> > implementations.
>
> Frank could better answer this, but there is currently not too much in
> the way of validation rules beyond those of the XML serialisation. If
> you look at the validation results in the SED-ML web tools
> (http://sysbioapps.dyndns.org/SED-ML%20Web%20Tools) its mostly around
> the XML structure and unique id's etc. I'm not sure if the libSEDML
> C++ implementation implements any further validation rules (its
> generated mostly using the same tools as libSBML, if that helps?).


It means that the mechanisms would be in place to allow you to write code
that could catch any validation errors, but that code would still have to
be written by hand for anything that wasn't a basic 'this is a required
element'; 'this attribute doesn't exist', etc.

Thanks!

-Lucian

Re: [SED-ML-discuss] Repeated tasks with multiple subtasks

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

Hi Lucian,

Sorry for the delay in replying!

>> The set value changes allowed in a repeated task only apply to the
>> model identified by the required modelReference attribute. So if
>> multiple sub-tasks refer to different models then you'd need a
>> setValue change defined in the repeated task for each of them.
>
>
> Aha!  OK, I missed this, since it wasn't in the UML.  Thanks!

Looks like the UML needs a fair bit of work to bring it up to scratch!
Thanks for catching and reporting all these errors.

> Is there any requirement that the referenced model be the same model as is
> referenced by the subtask (or one of the subtasks) in question?

I don't think so. I my own code I just look for set value changes that
match the "current" model as I iterate over the sub-tasks. If there
was a change that referenced a model that is not used then I would
simply ignore it. If I have understood your question correctly?
Although I think my code would throw an error if the model reference
matched but the xpath target expression failed to match a suitable
target for setting the value.

> And in fact, are there any validation rules planned at all?  There don't
> seem to be any in the spec; I don't know if there are any in any of the
> implementations.

Frank could better answer this, but there is currently not too much in
the way of validation rules beyond those of the XML serialisation. If
you look at the validation results in the SED-ML web tools
(http://sysbioapps.dyndns.org/SED-ML%20Web%20Tools) its mostly around
the XML structure and unique id's etc. I'm not sure if the libSEDML
C++ implementation implements any further validation rules (its
generated mostly using the same tools as libSBML, if that helps?).


Cheers,
David.

Re: [SED-ML-discuss] Repeated tasks with multiple subtasks

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

On Tue, Jun 2, 2015 at 3:14 PM, David Nickerson <dav...@gm...>
wrote:

> Hi Lucian,
>
> > I have a question about repeated tasks with multiple subtasks.  There
> are no
> > examples in the spec for this situation, so in general, I'm confused
> about
> > what sorts of things would be covered, and why.
>
> yes, more examples would be useful. We should try to get some more up
> on the website, or at least link out to various examples around the
> internet.
>
> > On a more mechanical level, I am wondering about the situation where
> > multiple subtasks refer to multiple models.  The spec seems to allow this
> > situation; from p53-54:
> >
> >   "All subTasks have to be carried out sequentially, each continuing from
> > the current model state (i.e. as at the end of the previous subTask,
> > assuming it simulates the same model) [...]"
> >
> > However, if the tasks do indeed reference different models, what do you
> do
> > with the listOfChanges?  Do you apply all of the changes to all
> referenced
> > models?
>
> The set value changes allowed in a repeated task only apply to the
> model identified by the required modelReference attribute. So if
> multiple sub-tasks refer to different models then you'd need a
> setValue change defined in the repeated task for each of them.
>

Aha!  OK, I missed this, since it wasn't in the UML.  Thanks!

Is there any requirement that the referenced model be the same model as is
referenced by the subtask (or one of the subtasks) in question?

And in fact, are there any validation rules planned at all?  There don't
seem to be any in the spec; I don't know if there are any in any of the
implementations.


> > Does this mean that all of the listed changes have to exist in all
> > referenced models, so it only works if (say) both models have a Parameter
> > with the id 'P1'?  If you write more generic XPath queries that happen to
> > have valid targets in both models, is that legal?  For example:
> >
> > /sbml:sbml/sbml:model/descendant::*[@id='P1']/@value
> >
> > might find a Parameter in one model and a LocalParameter in another.  Is
> > that OK?
>
> As above, I don't think this is the case. Is that something that might
> be useful if it was enabled? Something along the lines of a
> modelReference="*" (or perhaps a missing modelReference attribute) and
> then apply the change if a target for the given xpath expression can
> be found in any model used in any sub-task?


No!  I think one change per model is perfectly adequate. I was just trying
to imagine how you'd make a change without referencing a particular
model--but in fact, you *do* reference a model, so we're fine.

Thanks for the quick response!

-Lucian

Re: [SED-ML-discuss] Repeated tasks with multiple subtasks

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

Hi Lucian,

> I have a question about repeated tasks with multiple subtasks.  There are no
> examples in the spec for this situation, so in general, I'm confused about
> what sorts of things would be covered, and why.

yes, more examples would be useful. We should try to get some more up
on the website, or at least link out to various examples around the
internet.

> On a more mechanical level, I am wondering about the situation where
> multiple subtasks refer to multiple models.  The spec seems to allow this
> situation; from p53-54:
>
>   "All subTasks have to be carried out sequentially, each continuing from
> the current model state (i.e. as at the end of the previous subTask,
> assuming it simulates the same model) [...]"
>
> However, if the tasks do indeed reference different models, what do you do
> with the listOfChanges?  Do you apply all of the changes to all referenced
> models?

The set value changes allowed in a repeated task only apply to the
model identified by the required modelReference attribute. So if
multiple sub-tasks refer to different models then you'd need a
setValue change defined in the repeated task for each of them.

> Does this mean that all of the listed changes have to exist in all
> referenced models, so it only works if (say) both models have a Parameter
> with the id 'P1'?  If you write more generic XPath queries that happen to
> have valid targets in both models, is that legal?  For example:
>
> /sbml:sbml/sbml:model/descendant::*[@id='P1']/@value
>
> might find a Parameter in one model and a LocalParameter in another.  Is
> that OK?

As above, I don't think this is the case. Is that something that might
be useful if it was enabled? Something along the lines of a
modelReference="*" (or perhaps a missing modelReference attribute) and
then apply the change if a target for the given xpath expression can
be found in any model used in any sub-task?


Cheers,
David.

[SED-ML-discuss] Repeated tasks with multiple subtasks

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

I have a question about repeated tasks with multiple subtasks.  There are
no examples in the spec for this situation, so in general, I'm confused
about what sorts of things would be covered, and why.

On a more mechanical level, I am wondering about the situation where
multiple subtasks refer to multiple models.  The spec seems to allow this
situation; from p53-54:

  "All subTasks have to be carried out sequentially, each continuing from
the current model state (i.e. as at the end of the previous subTask,
assuming it simulates the same model) [...]"

However, if the tasks do indeed reference different models, what do you do
with the listOfChanges?  Do you apply all of the changes to all referenced
models?  Does this mean that all of the listed changes have to exist in all
referenced models, so it only works if (say) both models have a Parameter
with the id 'P1'?  If you write more generic XPath queries that happen to
have valid targets in both models, is that legal?  For example:

/sbml:sbml/sbml:model/descendant::*[@id='P1']/@value

might find a Parameter in one model and a LocalParameter in another.  Is
that OK?

-Lucian

Re: [SED-ML-discuss] Question about the 'source' attribute

[Pedro M. <ped...@ma...>]

On 05/07/2015 08:23 PM, Jonathan Cooper wrote:
> My opinion would be for Models with a matching id to take precedence, so
> the source is checked against other models first. If no matching Model
> is found, the source is then treated as a URI and either a local or
> remote file looked for. If none is found, an error is raised. Does that
> make sense?

I think this makes perfect sense and should be the way forward.

Pedro

-- 
Pedro Mendes
Professor of Computational Systems Biology
School of Computer Science
Manchester Centre for Integrative Systems Biology
University of Manchester

Manchester Institute of Biotechnology
131 Princess Street
Manchester, M1 7DN, U.K.

Re: [SED-ML-discuss] Question about the 'source' attribute

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

Hi Dave,

It looks like the spec is slightly loose in terminology here, and 
strictly the source attribute is either an SIdRef or a URI *reference* 
(rather than a URI per se).  From the RFC,

    A URI-reference is either a URI or a relative reference.  If the
    URI-reference's prefix does not match the syntax of a scheme followed
    by its colon separator, then the URI-reference is a relative
    reference.

Relative references are resolved relative to the location of the SED-ML 
file itself, according to the algorithm in section 5 of RDF3986.

Best wishes,
Jonathan

On 07/05/2015 20:33, David Brooks wrote:
> I'm not too familiar with the SED-ML spec, but in terms of RFC3986 I 
> understand that a valid URI *must* start with a scheme, which for a 
> file URI (include a relative one) is "file:". Is "file:mod1" a valid 
> SldRef?? If not, a SED-ML processor can distinguish between relative 
> file URIs and SldRefs.
>
>
> Regards,
> Dave
>
> On 7/05/15 9:25 pm, Lucian Smith wrote:
>> That would work for me--would you have IDs for non-Model objects in 
>> the SED-ML document ignored, in this case?  Or would you want that to 
>> be an actual error?
>>
>> -Lucian
>>
>> On Thu, May 7, 2015 at 12:23 PM, Jonathan Cooper 
>> <jon...@cs... <mailto:jon...@cs...>> wrote:
>>
>>     Hi Lucian,
>>
>>     I suspect this is indeed ambiguous in the spec and should be
>>     clarified. My opinion would be for Models with a matching id to
>>     take precedence, so the source is checked against other models
>>     first. If no matching Model is found, the source is then treated
>>     as a URI and either a local or remote file looked for. If none is
>>     found, an error is raised. Does that make sense?
>>
>>     Best wishes,
>>     Jonathan
>>
>>
>>     On 07/05/2015 20:19, Lucian Smith wrote:
>>>     I have a question about the 'source' attribute on the Model
>>>     class.  In the spec, this attribute is described as being either
>>>     a URI or an SIdRef to another Model in the document.  This can
>>>     be potentially ambiguous:  because URIs include relative
>>>     filenames, all SIdRefs are also valid relative filenames.
>>>
>>>     So, what do you do if you have a model with source "mod1", and
>>>     there's another Model in the same document named 'mod1', but
>>>     also a file in the same directory called 'mod1'?  Is it legal to
>>>     have a Model whose own id is 'mod1', and whose source attribute
>>>     is 'mod1', referring to a file?  What if there's no 'mod1'
>>>     object in the SED-ML document, but there is a 'mod1' file in the
>>>     directory?  What if some non-Model object has the id 'mod1' in
>>>     the document, and there's a 'mod1' file?  What if no file can be
>>>     found, and there's no ID with that, either?
>>>
>>>     -Lucian
>>>
>>>
>>>     ------------------------------------------------------------------------------
>>>     One dashboard for servers and applications across Physical-Virtual-Cloud
>>>     Widest out-of-the-box monitoring support with 50+ applications
>>>     Performance metrics, stats and reports that give you Actionable Insights
>>>     Deep dive visibility with transaction tracing using APM Insight.
>>>     http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
>>>
>>>
>>>     _______________________________________________
>>>     SED-ML-discuss mailing list
>>>     SED...@li...  <mailto:SED...@li...>
>>>     https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>>
>>
>>     ------------------------------------------------------------------------------
>>     One dashboard for servers and applications across
>>     Physical-Virtual-Cloud
>>     Widest out-of-the-box monitoring support with 50+ applications
>>     Performance metrics, stats and reports that give you Actionable
>>     Insights
>>     Deep dive visibility with transaction tracing using APM Insight.
>>     http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
>>     _______________________________________________
>>     SED-ML-discuss mailing list
>>     SED...@li...
>>     <mailto:SED...@li...>
>>     https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>>
>>
>>
>>
>> ------------------------------------------------------------------------------
>> One dashboard for servers and applications across Physical-Virtual-Cloud
>> Widest out-of-the-box monitoring support with 50+ applications
>> Performance metrics, stats and reports that give you Actionable Insights
>> Deep dive visibility with transaction tracing using APM Insight.
>> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
>>
>>
>> _______________________________________________
>> SED-ML-discuss mailing list
>> SED...@li...
>> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>
>
>
> ------------------------------------------------------------------------------
> One dashboard for servers and applications across Physical-Virtual-Cloud
> Widest out-of-the-box monitoring support with 50+ applications
> Performance metrics, stats and reports that give you Actionable Insights
> Deep dive visibility with transaction tracing using APM Insight.
> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
>
>
> _______________________________________________
> SED-ML-discuss mailing list
> SED...@li...
> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss

Re: [SED-ML-discuss] Question about the 'source' attribute

[David B. <d.b...@au...>]

I'm not too familiar with the SED-ML spec, but in terms of RFC3986 I 
understand that a valid URI *must* start with a scheme, which for a file 
URI (include a relative one) is "file:". Is "file:mod1" a valid SldRef?? 
If not, a SED-ML processor can distinguish between relative file URIs 
and SldRefs.


Regards,
Dave

On 7/05/15 9:25 pm, Lucian Smith wrote:
> That would work for me--would you have IDs for non-Model objects in 
> the SED-ML document ignored, in this case? Or would you want that to 
> be an actual error?
>
> -Lucian
>
> On Thu, May 7, 2015 at 12:23 PM, Jonathan Cooper 
> <jon...@cs... <mailto:jon...@cs...>> wrote:
>
>     Hi Lucian,
>
>     I suspect this is indeed ambiguous in the spec and should be
>     clarified. My opinion would be for Models with a matching id to
>     take precedence, so the source is checked against other models
>     first. If no matching Model is found, the source is then treated
>     as a URI and either a local or remote file looked for. If none is
>     found, an error is raised. Does that make sense?
>
>     Best wishes,
>     Jonathan
>
>
>     On 07/05/2015 20:19, Lucian Smith wrote:
>>     I have a question about the 'source' attribute on the Model
>>     class.  In the spec, this attribute is described as being either
>>     a URI or an SIdRef to another Model in the document.  This can be
>>     potentially ambiguous:  because URIs include relative filenames,
>>     all SIdRefs are also valid relative filenames.
>>
>>     So, what do you do if you have a model with source "mod1", and
>>     there's another Model in the same document named 'mod1', but also
>>     a file in the same directory called 'mod1'?  Is it legal to have
>>     a Model whose own id is 'mod1', and whose source attribute is
>>     'mod1', referring to a file?  What if there's no 'mod1' object in
>>     the SED-ML document, but there is a 'mod1' file in the
>>     directory?  What if some non-Model object has the id 'mod1' in
>>     the document, and there's a 'mod1' file?  What if no file can be
>>     found, and there's no ID with that, either?
>>
>>     -Lucian
>>
>>
>>     ------------------------------------------------------------------------------
>>     One dashboard for servers and applications across Physical-Virtual-Cloud
>>     Widest out-of-the-box monitoring support with 50+ applications
>>     Performance metrics, stats and reports that give you Actionable Insights
>>     Deep dive visibility with transaction tracing using APM Insight.
>>     http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
>>
>>
>>     _______________________________________________
>>     SED-ML-discuss mailing list
>>     SED...@li...  <mailto:SED...@li...>
>>     https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>
>
>     ------------------------------------------------------------------------------
>     One dashboard for servers and applications across
>     Physical-Virtual-Cloud
>     Widest out-of-the-box monitoring support with 50+ applications
>     Performance metrics, stats and reports that give you Actionable
>     Insights
>     Deep dive visibility with transaction tracing using APM Insight.
>     http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
>     _______________________________________________
>     SED-ML-discuss mailing list
>     SED...@li...
>     <mailto:SED...@li...>
>     https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>
>
>
>
> ------------------------------------------------------------------------------
> One dashboard for servers and applications across Physical-Virtual-Cloud
> Widest out-of-the-box monitoring support with 50+ applications
> Performance metrics, stats and reports that give you Actionable Insights
> Deep dive visibility with transaction tracing using APM Insight.
> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
>
>
> _______________________________________________
> SED-ML-discuss mailing list
> SED...@li...
> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss

Re: [SED-ML-discuss] Question about the 'source' attribute

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

That would work for me--would you have IDs for non-Model objects in the
SED-ML document ignored, in this case?  Or would you want that to be an
actual error?

-Lucian

On Thu, May 7, 2015 at 12:23 PM, Jonathan Cooper <
jon...@cs...> wrote:

>  Hi Lucian,
>
> I suspect this is indeed ambiguous in the spec and should be clarified. My
> opinion would be for Models with a matching id to take precedence, so the
> source is checked against other models first. If no matching Model is
> found, the source is then treated as a URI and either a local or remote
> file looked for. If none is found, an error is raised. Does that make sense?
>
> Best wishes,
> Jonathan
>
>
> On 07/05/2015 20:19, Lucian Smith wrote:
>
> I have a question about the 'source' attribute on the Model class.  In the
> spec, this attribute is described as being either a URI or an SIdRef to
> another Model in the document.  This can be potentially ambiguous:  because
> URIs include relative filenames, all SIdRefs are also valid relative
> filenames.
>
>  So, what do you do if you have a model with source "mod1", and there's
> another Model in the same document named 'mod1', but also a file in the
> same directory called 'mod1'?  Is it legal to have a Model whose own id is
> 'mod1', and whose source attribute is 'mod1', referring to a file?  What if
> there's no 'mod1' object in the SED-ML document, but there is a 'mod1' file
> in the directory?  What if some non-Model object has the id 'mod1' in the
> document, and there's a 'mod1' file?  What if no file can be found, and
> there's no ID with that, either?
>
>  -Lucian
>
>
> ------------------------------------------------------------------------------
> One dashboard for servers and applications across Physical-Virtual-Cloud
> Widest out-of-the-box monitoring support with 50+ applications
> Performance metrics, stats and reports that give you Actionable Insights
> Deep dive visibility with transaction tracing using APM Insight.http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
>
>
>
> _______________________________________________
> SED-ML-discuss mailing lis...@li...https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>
>
>
>
> ------------------------------------------------------------------------------
> One dashboard for servers and applications across Physical-Virtual-Cloud
> Widest out-of-the-box monitoring support with 50+ applications
> Performance metrics, stats and reports that give you Actionable Insights
> Deep dive visibility with transaction tracing using APM Insight.
> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
> _______________________________________________
> SED-ML-discuss mailing list
> SED...@li...
> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>
>

Re: [SED-ML-discuss] Question about the 'source' attribute

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

Hi Lucian,

I suspect this is indeed ambiguous in the spec and should be clarified. 
My opinion would be for Models with a matching id to take precedence, so 
the source is checked against other models first. If no matching Model 
is found, the source is then treated as a URI and either a local or 
remote file looked for. If none is found, an error is raised. Does that 
make sense?

Best wishes,
Jonathan

On 07/05/2015 20:19, Lucian Smith wrote:
> I have a question about the 'source' attribute on the Model class.  In 
> the spec, this attribute is described as being either a URI or an 
> SIdRef to another Model in the document.  This can be potentially 
> ambiguous:  because URIs include relative filenames, all SIdRefs are 
> also valid relative filenames.
>
> So, what do you do if you have a model with source "mod1", and there's 
> another Model in the same document named 'mod1', but also a file in 
> the same directory called 'mod1'?  Is it legal to have a Model whose 
> own id is 'mod1', and whose source attribute is 'mod1', referring to a 
> file?  What if there's no 'mod1' object in the SED-ML document, but 
> there is a 'mod1' file in the directory?  What if some non-Model 
> object has the id 'mod1' in the document, and there's a 'mod1' file?  
> What if no file can be found, and there's no ID with that, either?
>
> -Lucian
>
>
> ------------------------------------------------------------------------------
> One dashboard for servers and applications across Physical-Virtual-Cloud
> Widest out-of-the-box monitoring support with 50+ applications
> Performance metrics, stats and reports that give you Actionable Insights
> Deep dive visibility with transaction tracing using APM Insight.
> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
>
>
> _______________________________________________
> SED-ML-discuss mailing list
> SED...@li...
> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss

[SED-ML-discuss] Question about the 'source' attribute

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

I have a question about the 'source' attribute on the Model class.  In the
spec, this attribute is described as being either a URI or an SIdRef to
another Model in the document.  This can be potentially ambiguous:  because
URIs include relative filenames, all SIdRefs are also valid relative
filenames.

So, what do you do if you have a model with source "mod1", and there's
another Model in the same document named 'mod1', but also a file in the
same directory called 'mod1'?  Is it legal to have a Model whose own id is
'mod1', and whose source attribute is 'mod1', referring to a file?  What if
there's no 'mod1' object in the SED-ML document, but there is a 'mod1' file
in the directory?  What if some non-Model object has the id 'mod1' in the
document, and there's a 'mod1' file?  What if no file can be found, and
there's no ID with that, either?

-Lucian

Re: [SED-ML-discuss] Questions about repeated task

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

Hello Ilya, 

> On Mar 27, 2015, at 1:22 PM, Ilya Kiselev <ax...@de...> wrote:
> 
> Dear SedML community.
> 
> We are currently working on support of sed ml l1v2 in BioUML. So far we have few questions about RepeatedTask construct.
> 
> In repeated task we can have multiple subTasks. It seems to me that it is never stated that all subTasks should refer to the same model. However in data generator variables may refer only to task and not to the model inside the task. Then my questions are:
> 

I see the problem you are talking about. It could be resolved however with some xpath trickery. (not actually implemented in my tool though). let us assume that “task1” is a repeated task involving two models, each being SBML models, one with id=“model1” the other with id=“model2”. Then you could construct a datagenerator totalADP like so: 

<dataGenerator id=“totalAADP”>
  <listOfVariables>
    <variable id=“adp_1” taskReference=“task1” target=“/sbml:sbml/sbml:model[@id=‘model1’]/sbml:listOfSpecies/sbml:species[@id=‘ADP’]”/>
    <variable id=“adp_2” taskReference=“task1” target=“/sbml:sbml/sbml:model[@id=‘model2’]/sbml:listOfSpecies/sbml:species[@id=‘ADP’]”/>
  </listOfVariables>
  <math xmlns=“…”>
   <apply> 
   <plus>
   <ci>adp_1</ci>
   <ci>adp_2</ci>
   </apply>
  </math>
</dataGenerator>




> 1. Is it true that all subTasks in repeatedTask must refer to the same model?

not this is not in the current spec .. and i don’t think it is desirable … 

> 2. Can several repeatedTasks use the same task as subTasks?

yes

> 3. If task is used as a subtask then variable in data generator can not refer to it directly and should refer to its "parent" RepeatedTask. Is that correct?
> 

if you want to refer to the results as available from running the repeatedTask, then i would point to the repeated task. In my implementation I only run those tasks that are referred to by datagenerators, as when no output is needed to be computed, there is no reason to run the task. 

> Another question is regarding "resetModel" attribute. In Sed-ML specification it is stated that "resetModel = false" means that model in the task will not be reset to its normal state at the start of each iteration. 
> 
> Say we have subTask which refers to UniformTimeCourse simulation with initialTime = 0 and outputEndTime = 100.
> 
> This means that next iteration will start from time = 0 but variables of the model will be taken from the last time point of previous simulation which is time = 100. This situation can lead to strange and nonsensical results especially when time is explicitly involved in model equations.
> 
> For now it seems to me that "resetModel = false" makes sense only for onestepper and steady state and "resetModel = true" for onestepper will be also quite strange. Can anybody share showcases for such situations (I mean "resetModel=false" for time course simulation and "resetModel = true" for one step simulation)?
> 
> One showcase with time course simulation and resetModel = "false" I can think of is not compatible with current Sed-ML specifications (but it is doable in BioUML). It relies on changing of simulation parameters in RepeatedTask iterations. 
> For example: perform timecourse simulation from 0 to 100 with step 0.1. At the second iteration change initial time to 100 and completion time to 200, do some other changes to model parameters and continue simulation.
> 

this is precisely what I implemented, if i don’t reset the model, I don’t reset the models time variable, thus the second run for me would go from 100 to 200. 

> Also, are there any tests for sedml? It will be very helpful for tool developers. Currently we try to reproduce examples from specifications but they lacks different constructions (e.g. multiple subtasks inside repeatedTask or mentioned above situations). 
> 

that sounds really useful … i have a couple of tests, but we could do with a whole suite of them … 

Frank

> -- 
> Best regards, Ilya
> ------------------------------------------------------------------------------
> Dive into the World of Parallel Programming The Go Parallel Website, sponsored
> by Intel and developed in partnership with Slashdot Media, is your hub for all
> things parallel software development, from weekly thought leadership blogs to
> news, videos, case studies, tutorials and more. Take a look and join the 
> conversation now. http://goparallel.sourceforge.net/_______________________________________________
> SED-ML-discuss mailing list
> SED...@li...
> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss

[SED-ML-discuss] Questions about repeated task

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

Dear SedML community.

We are currently working on support of sed ml l1v2 in BioUML. So far we
have few questions about RepeatedTask construct.

In repeated task we can have multiple subTasks. It seems to me that it is
never stated that all subTasks should refer to the same model. However in
data generator variables may refer only to task and not to the model inside
the task. Then my questions are:

1. Is it true that all subTasks in repeatedTask must refer to the same
model?
2. Can several repeatedTasks use the same task as subTasks?
3. If task is used as a subtask then variable in data generator can not
refer to it directly and should refer to its "parent" RepeatedTask. Is that
correct?

Another question is regarding "resetModel" attribute. In Sed-ML
specification it is stated that "resetModel = false" means that model in
the task will not be reset to its normal state at the start of each
iteration.

Say we have subTask which refers to UniformTimeCourse simulation with
initialTime = 0 and outputEndTime = 100.

This means that next iteration will start from time = 0 but variables of
the model will be taken from the last time point of previous simulation
which is time = 100. This situation can lead to strange and nonsensical
results especially when time is explicitly involved in model equations.

For now it seems to me that "resetModel = false" makes sense only for
onestepper and steady state and "resetModel = true" for onestepper will be
also quite strange. Can anybody share showcases for such situations (I mean
"resetModel=false" for time course simulation and "resetModel = true" for
one step simulation)?

One showcase with time course simulation and resetModel = "false" I can
think of is not compatible with current Sed-ML specifications (but it is
doable in BioUML). It relies on changing of simulation parameters in
RepeatedTask iterations.
For example: perform timecourse simulation from 0 to 100 with step 0.1. At
the second iteration change initial time to 100 and completion time to 200,
do some other changes to model parameters and continue simulation.

Also, are there any tests for sedml? It will be very helpful for tool
developers. Currently we try to reproduce examples from specifications but
they lacks different constructions (e.g. multiple subtasks inside
repeatedTask or mentioned above situations).

-- 
Best regards, Ilya

[SED-ML-discuss] HARMONY 2015

[Falk S. <fal...@mo...>]

I apologise if you receive this email via different lists

Dear COMBINE community!

This is a kind reminder that the deadline for registration for HARMONY 2015 is 
Monday, 30 March, 2015.

Please register if you plan to join us for the Hackathon.

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. 

HARMONY 2015 will take place 

Mon. 20- Fri. 24 April 2015 in at the Leucorea in Wittenberg, Germany

More information can be found here: http://www.co.mbine.org/events/HARMONY_2015

We hope to see you in Wittenberg!

Dagmar Waltemath, Tobias Czauderna & Falk Schreiber

[SED-ML-discuss] Help to define a workflow

[Nicolas R. <nik...@gm...>]

Hi all,

We are trying to find out if we can define one or more sed-ml files to 
describe a specific workflow.

Here is the quick description of what we want to achieve:

- we have three different models (M1, M2, M3).
- we run a repeated task on M1 (RT_1).
- we run a repeated task on model2 (RT_2).
- then we would like to modify model3 with some values coming from 
model1 and model2.
- last we would like to run a repeated task on M3 (RT_3).

This is a simplified version but there could be further tasks that make 
use of the output coming from model3.
I don't think we can do that using the current specs.

It could be achieve with two sed-ml files, if we can save the output of 
model1 and model2 in one sed-ml file and
then read these files to modify model3 in a second sed-ml file. I 
believe that the second sed-ml file would be possible
to define using L1V3 although I did not check in details the specs. But 
I don't think there is anything to allow us
to set an output file name that could be re-use later on other sed-ml files?

But it could almost be done in one sed-ml file although there are some 
limitations at the moment:

1) The order of the tasks. One solution would be to put the optional 
'order' attribute onto the main task element so
that we can be sure that some tasks are run before others. Otherwise, we 
can abuse the repeated task element
and create a repeated task RT_4 that is repeated only once and that as 
RT_1, RT_2 and RT_3 as sub tasks, in the right order.
Additionally, we might have situations were we define a lot of different 
tasks that are re-used as sub task of other bigger tasks
and we don't really want them to be run on their own, so may be it would 
be useful to add an attribute on the task element to
say that it cannot be run on it's own, that it is just a definition 
re-used later.

2) Changing some values of M3 only at the start of RT_3. It should be 
possible to do that with the current spec, using a range and
a piecewise operator, for example, if we want to set the value of the 
parameter P1 in M3. At the first iteration, we set the value to
the value of the parameter P in M1 otherwise do nothing, meaning keep 
the value of P1 in M3.
Although if this kind of manipulation would be considered as something 
useful, I wonder if it would be worth changing the specs a bit
to add a listOfInitialChanges to the repeated task or an attribute to 
the SetValue class to tell to do it only once at the start ?
An example in the specs of something like that would be nice as well.

Thanks,
Nico

Re: [SED-ML-discuss] Status of the SED-ML tools

[Herbert S. <hs...@gm...>]

We also use Frank's Python bindings to libSEDML which works quite well.

Herbert

On Mon, Jan 26, 2015 at 11:30 AM, Frank Bergmann <fbe...@ca...>
wrote:

> Hello Nico,
>
> My .net project http://libsedml.sourceforge.net/ supports 1.2 as well as
> the first data proposal that i was sent around at the last hackathon. All
> projects that use it already support it:
>
> - SBW: (windows installer under):
> http://128.208.17.26/fbergman/files/latest/SetupSBW.exe
> - SED-ML Script Editor: http://sourceforge.net/projects/libsedml/files/
>
> In case you happen not to use windows, you could try the online version,
> the SED-ML Web Tools as available from: http://sysbioapps.dyndns.org.
>
> If you are looking for a portable library, or you simply want to use a
> library in one of the many SWIG bindings, you could have a look at:
>
> https://github.com/fbergmann/libSEDML
>
> It also supports reading / writing of SED-ML L1V2 models. I plan to add
> L1V3 support before the hackathon.
>
> All the best
> Frank
>
> p.s: good idea about noting what level of sed-ml is supported by a tool
> for the showcase … I will update the survey to reflect that.
>
> > On Jan 26, 2015, at 6:05 PM, Nicolas Rodriguez <nik...@gm...>
> wrote:
> >
> > Hello,
> >
> > I would be nice to get an update of which tools have been updated to
> > support SED-ML L1V2.
> > And this info would be useful in the showcase part of the sedml.org web
> > site.
> >
> > Even more, is there already some tools or libraries that have been
> > updated to support the L1V3 draft specification or
> > intend to do it in the near future? In particular, it might be useful
> > for the whole-cell modeling summer school.
> >
> > Thanks,
> > Nico
> >
> >
> >
> ------------------------------------------------------------------------------
> > Dive into the World of Parallel Programming. The Go Parallel Website,
> > sponsored by Intel and developed in partnership with Slashdot Media, is
> your
> > hub for all things parallel software development, from weekly thought
> > leadership blogs to news, videos, case studies, tutorials and more. Take
> a
> > look and join the conversation now. http://goparallel.sourceforge.net/
> > _______________________________________________
> > SED-ML-discuss mailing list
> > SED...@li...
> > https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>
>
>
> ------------------------------------------------------------------------------
> Dive into the World of Parallel Programming. The Go Parallel Website,
> sponsored by Intel and developed in partnership with Slashdot Media, is
> your
> hub for all things parallel software development, from weekly thought
> leadership blogs to news, videos, case studies, tutorials and more. Take a
> look and join the conversation now. http://goparallel.sourceforge.net/
> _______________________________________________
> SED-ML-discuss mailing list
> SED...@li...
> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss
>

Re: [SED-ML-discuss] Status of the SED-ML tools

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

Hello Nico, 

My .net project http://libsedml.sourceforge.net/ supports 1.2 as well as the first data proposal that i was sent around at the last hackathon. All projects that use it already support it: 

- SBW: (windows installer under): http://128.208.17.26/fbergman/files/latest/SetupSBW.exe
- SED-ML Script Editor: http://sourceforge.net/projects/libsedml/files/

In case you happen not to use windows, you could try the online version, the SED-ML Web Tools as available from: http://sysbioapps.dyndns.org. 

If you are looking for a portable library, or you simply want to use a library in one of the many SWIG bindings, you could have a look at: 

https://github.com/fbergmann/libSEDML

It also supports reading / writing of SED-ML L1V2 models. I plan to add L1V3 support before the hackathon. 

All the best
Frank

p.s: good idea about noting what level of sed-ml is supported by a tool for the showcase … I will update the survey to reflect that. 

> On Jan 26, 2015, at 6:05 PM, Nicolas Rodriguez <nik...@gm...> wrote:
> 
> Hello,
> 
> I would be nice to get an update of which tools have been updated to 
> support SED-ML L1V2.
> And this info would be useful in the showcase part of the sedml.org web 
> site.
> 
> Even more, is there already some tools or libraries that have been 
> updated to support the L1V3 draft specification or
> intend to do it in the near future? In particular, it might be useful 
> for the whole-cell modeling summer school.
> 
> Thanks,
> Nico
> 
> 
> ------------------------------------------------------------------------------
> Dive into the World of Parallel Programming. The Go Parallel Website,
> sponsored by Intel and developed in partnership with Slashdot Media, is your
> hub for all things parallel software development, from weekly thought
> leadership blogs to news, videos, case studies, tutorials and more. Take a
> look and join the conversation now. http://goparallel.sourceforge.net/
> _______________________________________________
> SED-ML-discuss mailing list
> SED...@li...
> https://lists.sourceforge.net/lists/listinfo/sed-ml-discuss