12.6. Requirements to describe a post-processing after an ADAO calculation¶
Results processing is usually needed after an ADAO calculation, to insert it in a complete study. After execution of a calculation case, the main information is the variable “Analysis” which contains a result of optimal estimation. In addition, all calculation variables that have been requested in intermediate storage are also available by means of the special algorithm variable “StoreSupplementaryCalculations”.
The simplest processing is often represented by a few lines of Python, which can easily be repeated or carried over between studies. But more complex processing of results, in the full SALOME study environment, is often done by explicit parts of additional post-processing, either in YACS nodes or by Python commands in TUI, or other methods. It is therefore often interesting to identify at least part of the calculations following the ADAO estimation, and to associate them in the calculation case.
ADAO thus gives the ability to define a general post-processing for each calculation case. This definition is done by indicating the commands to be performed at the output of the ADAO calculation.
12.6.1. Save a dedicated post-processing¶
In the graphical interface EFICAS for ADAO, there are 3 convenient methods to integrate a dedicated post-processing for an ADAO case. The method is chosen using the “FROM” keyword of the main entry “UserPostAnalysis”, as shown in the following figure:
Choose your input type for the recorded post-processing
Post-processing can be provided as an explicit script (input of type “String”), as a script contained in an external file (input of type “Script”), or by using a template (input of type “Template”). Templates are provided by default when using the graphical editor EFICAS for ADAO or the TUI interface, and are detailed in the Inventory of simple templates of post-processing (“Template”) section that follows. These are simple scripts that can be adapted by the user, either in the integrated case editing step with EFICAS for ADAO, or in the schema editing step before execution, to improve the performance of the ADAO calculation in the SALOME execution supervisor.
In the ADAO TUI textual interface (see the [DocR] Textual User Interface for ADAO (TUI/API) section), the same information can be given using the “setUserPostAnalysis” command. The arguments of this command allow to define the treatment either as a template (argument “Template”) referring to one of the scripts detailed in the part Inventory of simple templates of post-processing (“Template”), or as an explicit script (argument “String”), or as a script contained in an external file (argument “Script”)
12.6.2. General form of a script to define a dedicated post-processing¶
A set of post-processing commands is a special Python script, which is automatically activated at the end of the estimation calculations in ADAO. Any Python commands, which a user can add after a GUI calculation in ADAO, can be part of this post-processing. Several command set templates are available by default,essentially to give the simplest possible example of recording these series.
To be usable in an automatic way, it is required that any call of the ADAO calculation case, to recover a variable, is done only with the reserved name “ADD”. As an example, here is a very simple script (very similar to the “ValuePrinter” template), usable to display the value of the optimal estimate:
import numpy
xa = numpy.ravel(ADD.get('Analysis')[-1])
print(' === Analysis =',xa)
If the command “ADD.get(…)”, used to obtain a result variable, does not use the reserved name “ADD” for the calculation case, then the call will lead to a runtime error and will warn about the missing case name.
To illustrate, the declaration of a model, in TUI interface, is done by using the command:
ADD.setUserPostAnalysis(Template = "AnalysisPrinter")
Warning
If not using the default available templates, it is up to the user to make carefully established function scripts or external programs that do not crash before being registered as an “observer” function. The debugging can otherwise be really difficult!
Hereinafter we give the identifier and the contents of all the available simple templates.
12.6.3. Inventory of simple templates of post-processing (“Template”)¶
12.6.3.1. Modèle AnalysisPrinter¶
Print on standard output the optimal value.
print('# Post-analysis')
import numpy
xa=ADD.get('Analysis')[-1]
print('Analysis',xa)
12.6.3.2. Modèle AnalysisSaver¶
Save the optimal value in a file of the ‘/tmp’ directory named ‘analysis.txt’.
print('# Post-analysis')
import numpy
xa=ADD.get('Analysis')[-1]
f='/tmp/analysis.txt'
print('Analysis saved in "%s"'%f)
numpy.savetxt(f,xa)
12.6.3.3. Modèle AnalysisPrinterAndSaver¶
Print on standard output and, in the same time save in a file of the ‘/tmp’ directory, the optimal value.
print('# Post-analysis')
import numpy
xa=ADD.get('Analysis')[-1]
print('Analysis',xa)
f='/tmp/analysis.txt'
print('Analysis saved in "%s"'%f)
numpy.savetxt(f,xa)
12.6.3.4. Modèle AnalysisSeriePrinter¶
Print on standard output the optimal value series.
print('# Post-analysis')
import numpy
xa=ADD.get('Analysis')
print('Analysis',xa)
12.6.3.5. Modèle AnalysisSerieSaver¶
Save the optimal value series in a file of the ‘/tmp’ directory named ‘analysis.txt’.
print('# Post-analysis')
import numpy
xa=ADD.get('Analysis')
f='/tmp/analysis.txt'
print('Analysis saved in "%s"'%f)
numpy.savetxt(f,xa)
12.6.3.6. Modèle AnalysisSeriePrinterAndSaver¶
Print on standard output and, in the same time save in a file of the ‘/tmp’ directory, the optimal value series.
print('# Post-analysis')
import numpy
xa=ADD.get('Analysis')
print('Analysis',xa)
f='/tmp/analysis.txt'
print('Analysis saved in "%s"'%f)
numpy.savetxt(f,xa)