Input files

Input files are parsed a little differently from the command line. Perhaps the most important feature are the '#'-comments that allow to add text into the file without being actually parsed:

    # -*- getpot -*-  activate emacs 'getpot-mode'
    # FILE:    "example.pot"
    # PURPOSE: some examples how to use GetPot for input file parsing.
    #
    # (C) 2001 Frank R. Schaefer
    #
    # License Terms: GNU Lesser GPL, ABSOLUTELY NO WARRANTY
    #####################################################################
    v0 = 34.2    # [m/s^2] initial speed of point mass

As can be seen in the same example, the variable assignments now allow whitespace between a left hand value v0 the assignment operator ' =' and a right hand value 34.2.

One thing is crucially different in input files: sections. In input files, the basic advantage of sections is to reduce the variable length. Imagine, you have a parameter weight occurring multiple times: for the total vehicle, for each tire, for a load in the trunk etc. Now in order give each one a unique identifier one would have to call them total_vehicle_weight, vehicle_tires_front_left_weight and so on. Here is where sections become handy:

...
    [vehicle]
    length  = 2.65             # [m]
    initial-xyz = '100.  0.1  5.0' # [m]

    # coefficients of magic formula [Pajeka, et al.]
    [./tires/front/right]  
    B = 3.7976    C = 1.2    E = -0.5    D = 64322.404 
    [./tires/front/left]  
    B = 3.7976    C = 1.2    E = -0.5    D = 64322.404 
    [./tires/rear/right]  
    B = 3.546    C = 1.135   E = -0.4    D = 59322.32
    [./tires/rear/left]  
    B = 3.546    C = 1.11    E = -0.32   D = 59322.32


    [../chassis]  # i.e. vehicle/chassis
    Roh = 1.21;   # [kg/m^3] density of air 
    S   = 5.14;   # [m^2]    reference surface
    Cd  = 0.45;   # [1]      air drag coefficient
...

In section 4 it was explained how to read out variables. The same functions work, of course, for a database build up on a parsed file. The only difference is that sections produce suffixes in front of variables and options. A database fed with a file as the above one, can be queried as follows:

...
    front_tire.B = ifile("vehicle/tires/front/right/B", 0.);
    front_tire.C = ifile("vehicle/tires/front/right/C", 0.);
    front_tire.D = ifile("vehicle/tires/front/right/D", 0.);
    front_tire.E = ifile("vehicle/tires/front/right/E", 0.);
...

Note that there are two special indicators in a section label:

./

Take the actual section and append the following name to it.

../

Go back to the parent section and create a name appended to the name of the parent.

Finally, an empty label like [], resets the section suffix to nothing. Speaking in name space terminology, we are back to the global name space.