Re: [OpenMOLE-users] csvfile sampling in directorder

2016-12-05 Thread paul chapron 
Ok I see,  so the order is not preserved . Thanks !


2016-12-04 15:17 GMT+01:00 Romain Reuillon :

> I am not sure what you mean by the order of the files in a folder...
>
> In OpenMOLE the order of executions of the parallel stream is not
> determined. An exploration task generate independant execution streams
> which may be executed in any order. However if you use an aggregation
> transition the results are sorted in the same order as the sampling order
> of the matching exploration during the aggregation operation.
>
>
> Le 04/12/2016 à 12:51, paul chapron a écrit :
>
> I'd like to retrieve the results of the execution in the exact order of
> the csv file lines
>
> e.g. first result file or folder in my workdirectory must be the result of
> my model with iput = the first line of the CSV from which I take a
> CSVsampling task , second resulkt file must be the result of the model with
> the second line of the CSV as iputs and so on...
>
> Since it's called "sampling" , I guess there is some randomness at some
> point
>
> 2016-12-01 20:27 GMT+01:00 Romain Reuillon :
>
>> Hi Paul,
>>
>> I don't get what you want. The sampling here assing one seed for each
>> line in the csv file. Therefor the subsequent task will be executed once
>> for each line. What do you want to achive?
>>
>> Romain
>>
>>
>> Le 01/12/2016 à 17:46, paul chapron a écrit :
>>
>> Hi fellow users ,
>>
>>
>> I have replication task to perform (n execution per line in my file, n
>> result file per execution ) , that looks like :
>>
>> _
>> val s = CSVSampling(workDirectory/"population40.csv") set (
>>   columns += ("distReculVoirie", distReculVoirie),
>>   columns += ("distReculFond", distReculFond),
>>   columns += ("distReculLat", distReculLat),
>> [...]
>>  separator := ','
>> )
>>
>> val exploration = ExplorationTask(
>>   s x
>> (seed in UniformDistribution[Long](42) take 1)   //for now , only one
>> execution is needed
>> )
>> [...]
>> __
>>
>>
>> It appears that now I must not sample randomly anymore in the lines of
>> the CSV.
>>
>> Is there a way to tell CSVsampling to go through the file line-by-line
>> and not at random ?
>> (It would savec me a lot of matching on values later)
>>
>>
>> Cheers
>>
>> paul
>>
>>
>>
>>
>>
>> ___
>> OpenMOLE-users mailing 
>> listOpenMOLE-users@iscpif.frhttp://fedex.iscpif.fr/mailman/listinfo/openmole-users
>>
>> ___ OpenMOLE-users mailing
>> list OpenMOLE-users@iscpif.fr http://fedex.iscpif.fr/mailman
>> /listinfo/openmole-users
>
> ___
> OpenMOLE-users mailing 
> listOpenMOLE-users@iscpif.frhttp://fedex.iscpif.fr/mailman/listinfo/openmole-users
>
>
> ___
> OpenMOLE-users mailing list
> OpenMOLE-users@iscpif.fr
> http://fedex.iscpif.fr/mailman/listinfo/openmole-users
>
>
___
OpenMOLE-users mailing list
OpenMOLE-users@iscpif.fr
http://fedex.iscpif.fr/mailman/listinfo/openmole-users

Re: [OpenMOLE-users] csvfile sampling in directorder

2016-12-04 Thread Romain Reuillon 
I am not sure what you mean by the order of the files in a folder...

In OpenMOLE the order of executions of the parallel stream is not
determined. An exploration task generate independant execution streams
which may be executed in any order. However if you use an aggregation
transition the results are sorted in the same order as the sampling
order of the matching exploration during the aggregation operation.

Le 04/12/2016 à 12:51, paul chapron a écrit :
> I'd like to retrieve the results of the execution in the exact order
> of the csv file lines
>  
> e.g. first result file or folder in my workdirectory must be the
> result of my model with iput = the first line of the CSV from which I
> take a CSVsampling task , second resulkt file must be the result of
> the model with the second line of the CSV as iputs and so on...
>
> Since it's called "sampling" , I guess there is some randomness at
> some point 
>
> 2016-12-01 20:27 GMT+01:00 Romain Reuillon  >:
>
> Hi Paul,
>
> I don't get what you want. The sampling here assing one seed for
> each line in the csv file. Therefor the subsequent task will be
> executed once for each line. What do you want to achive?
>
> Romain
>
>
> Le 01/12/2016 à 17:46, paul chapron a écrit :
>> Hi fellow users ,
>>
>>
>> I have replication task to perform (n execution per line in my
>> file, n result file per execution ) , that looks like :
>>
>> _
>> val s = CSVSampling(workDirectory/"population40.csv") set (
>>   columns += ("distReculVoirie", distReculVoirie),
>>   columns += ("distReculFond", distReculFond),
>>   columns += ("distReculLat", distReculLat),
>> [...]
>>  separator := ','
>> )
>>
>> val exploration = ExplorationTask(
>>   s x
>> (seed in UniformDistribution[Long](42) take 1)   //for now , only
>> one execution is needed
>> )
>> [...]
>> __
>>
>>
>> It appears that now I must not sample randomly anymore in the
>> lines of the CSV.
>>
>> Is there a way to tell CSVsampling to go through the file
>> line-by-line and not at random ?
>> (It would savec me a lot of matching on values later)
>>
>>
>> Cheers
>>
>> paul
>>
>>
>>
>>
>>
>> ___
>> OpenMOLE-users mailing list
>> OpenMOLE-users@iscpif.fr 
>> http://fedex.iscpif.fr/mailman/listinfo/openmole-users
>> 
>
> ___ OpenMOLE-users
> mailing list OpenMOLE-users@iscpif.fr
> 
> http://fedex.iscpif.fr/mailman/listinfo/openmole-users
>  
>
> ___
> OpenMOLE-users mailing list
> OpenMOLE-users@iscpif.fr
> http://fedex.iscpif.fr/mailman/listinfo/openmole-users



signature.asc
Description: OpenPGP digital signature
___
OpenMOLE-users mailing list
OpenMOLE-users@iscpif.fr
http://fedex.iscpif.fr/mailman/listinfo/openmole-users

Re: [OpenMOLE-users] csvfile sampling in directorder

2016-12-04 Thread paul chapron 
I'd like to retrieve the results of the execution in the exact order of the
csv file lines

e.g. first result file or folder in my workdirectory must be the result of
my model with iput = the first line of the CSV from which I take a
CSVsampling task , second resulkt file must be the result of the model with
the second line of the CSV as iputs and so on...

Since it's called "sampling" , I guess there is some randomness at some
point

2016-12-01 20:27 GMT+01:00 Romain Reuillon :

> Hi Paul,
>
> I don't get what you want. The sampling here assing one seed for each line
> in the csv file. Therefor the subsequent task will be executed once for
> each line. What do you want to achive?
>
> Romain
>
>
> Le 01/12/2016 à 17:46, paul chapron a écrit :
>
> Hi fellow users ,
>
>
> I have replication task to perform (n execution per line in my file, n
> result file per execution ) , that looks like :
>
> _
> val s = CSVSampling(workDirectory/"population40.csv") set (
>   columns += ("distReculVoirie", distReculVoirie),
>   columns += ("distReculFond", distReculFond),
>   columns += ("distReculLat", distReculLat),
> [...]
>  separator := ','
> )
>
> val exploration = ExplorationTask(
>   s x
> (seed in UniformDistribution[Long](42) take 1)   //for now , only one
> execution is needed
> )
> [...]
> __
>
>
> It appears that now I must not sample randomly anymore in the lines of the
> CSV.
>
> Is there a way to tell CSVsampling to go through the file line-by-line and
> not at random ?
> (It would savec me a lot of matching on values later)
>
>
> Cheers
>
> paul
>
>
>
>
>
> ___
> OpenMOLE-users mailing 
> listOpenMOLE-users@iscpif.frhttp://fedex.iscpif.fr/mailman/listinfo/openmole-users
>
>
>
> ___
> OpenMOLE-users mailing list
> OpenMOLE-users@iscpif.fr
> http://fedex.iscpif.fr/mailman/listinfo/openmole-users
>
>
___
OpenMOLE-users mailing list
OpenMOLE-users@iscpif.fr
http://fedex.iscpif.fr/mailman/listinfo/openmole-users

Re: [OpenMOLE-users] csvfile sampling in directorder

2016-12-01 Thread Romain Reuillon 
Hi Paul,

I don't get what you want. The sampling here assing one seed for each
line in the csv file. Therefor the subsequent task will be executed once
for each line. What do you want to achive?

Romain

Le 01/12/2016 à 17:46, paul chapron a écrit :
> Hi fellow users ,
>
>
> I have replication task to perform (n execution per line in my file, n
> result file per execution ) , that looks like :
>
> _
> val s = CSVSampling(workDirectory/"population40.csv") set (
>   columns += ("distReculVoirie", distReculVoirie),
>   columns += ("distReculFond", distReculFond),
>   columns += ("distReculLat", distReculLat),
> [...]
>  separator := ','
> )
>
> val exploration = ExplorationTask(
>   s x
> (seed in UniformDistribution[Long](42) take 1)   //for now , only one
> execution is needed
> )
> [...]
> __
>
>
> It appears that now I must not sample randomly anymore in the lines of
> the CSV.
>
> Is there a way to tell CSVsampling to go through the file line-by-line
> and not at random ?
> (It would savec me a lot of matching on values later)
>
>
> Cheers
>
> paul
>
>
>
>
>
> ___
> OpenMOLE-users mailing list
> OpenMOLE-users@iscpif.fr
> http://fedex.iscpif.fr/mailman/listinfo/openmole-users




signature.asc
Description: OpenPGP digital signature
___
OpenMOLE-users mailing list
OpenMOLE-users@iscpif.fr
http://fedex.iscpif.fr/mailman/listinfo/openmole-users