Add a VPLanet Output

To add a new output, complete the following steps.

First, determine if the output is applicable to one module, or multiple modules. For example, eccentricity is part of the SpiNBody, EqTide, and DistOrb modules, but the mean anomaly is only used in SpiNBody. If the output should be callable by multiple modules, then you should modify output.[ch], but if it is only appropriate for an individual module, then it belongs in module files (spinbody.[ch] for the previous example).

Next, in the header file, add a OUT macro to give your output a unique ID, e.g.

#define OUT_MYNEWOUTPUT   77

Note that the number of the option must lie in the appropriate range for the module. For new outputs in output.c, the range is 0-1000, for other modules, consult vplanet.h, specifically the text just before MODULEOUTEND is defined. Make sure your new outputs name is easily understood and very unlikely to be duplicated by another variable name.

Then add a block of code in the appropriate InitializeOutput function (InitializeOutputGeneral in output.c or InitializeOutputModuleName, where “ModuleName” = AtmEsc, EqTide, etc., in the module’s c file) like this:

4output[OUT_MASS].bNeg = 1;
5output[OUT_MASS].dNeg = 1./MEARTH;
6output[OUT_MASS].iNum = 1;
7output[OUT_MASS].iModuleBit = 1;
8fnWrite[OUT_MASS] = &WriteMass;

These lines instruct VPLanet on the key features of the output. Line 1 contains the name of the option that will appear in both the log file and/or the argument list to saOutputOrder. Line 2 is the short description of the output; mass is a fundamental property, so doesn’t need much explanation. This description should be now more than 10 words long. Line 3 is the units of the output if the user places a negative sign (“-”) in front of it and should be from the list of unit names used in astropy. (If you must use an alternative unit, please add it to the file in the vplanet folder that is created after pip installation.) This feature is only allowed if the output is positive definite and requires the bNeg flag be set to 1 (Line 4). If set to 0 (or not set), then including a negative sign will cause VPLanet to exit and error. The next line, is a mathematical description of how to convert the units from system units (SI) to the custom unit. The code multiplies the value by this conversion factor (note that MEARTH is a macro defined in vplanet.h). Line 6 tells VPLanet how many elements are part of the output; mass is a scalar, so this value is set to 1.

Next on Line 7 is a variable that explains what options are associated with the output and is a bit more complicated. For ease of use, VPLanet checks if a selected output is actually appropriate for the selected modules. This check is done via a bitwise comparison, with each module assigned a unique bit as defined in vplanet.h. Outputs that are applicable to all modules, like mass, can be set to 1 as shown. Otherwise, set the value to equal the sum of module bits, such as is done for the rotation period:

In other words, each module is assigned a unique integer that is a power of 2 so that all module combinations sum to a unique multiple of 2. When an output is read in, VPLanet checks that at least one of the appropriate bits in the integer is set to 1.

The last line in the example is a function pointer to the subroutine that reads in the output. This function must be defined prior to this line. Finally, not shown is the cLongDescr field that contains a longer description of the output parameter that is used with the -H command line option, e.g. the “precession angle” requires a bit more explanation for a typical user:


Any field that is omitted will be set to the default values that are defined in output.c:InitializeOutput.

Next, let’s turn to the fnWrite subroutine., which assigns the value and unit to the output in preparation for writing to an output file. These functions typically look like this:

void WriteMass(BODY *body,CONTROL *control,OUTPUT *output,SYSTEM *system,UNITS
               *units,UPDATE *update,int iBody,double *dTmp,char cUnit[]) {

  *dTmp = body[iBody].dMass;

  if (output->bDoNeg[iBody]) {
    *dTmp *= output->dNeg;
  } else {
    *dTmp /= fdUnitsMass(units->iMass);

In this function the pointer to dTmp is the value(s) you want to be output. It can be either a scalar or an array, and can be either calculated here or can call another function. In this case, we can just use the current value of the body’s mass. The next blocks of code convert the value of the mass from system units (SI) to the user’s preferred units, which depend on if the user selected the negative option, i.e. the argument to saOutputOrder is either “Mass” or “-Mass”. If the negative sign was not included or is not allowed, then the else block is executed and the value is converted to the units set by the user by options such as sUnitMass, sUnitLength, etc. In this case, mass is simple, but other units like energy or pressure could require numerous operations to convert to the appropriate units. In some cases, conversion subroutines for common unit combinations already exist – see control.c. In both cases, a string that describes the units is also assigned, with complex unit descriptions also available in control.c.

With these steps completed, the output parameter is now available in VPLanet. The code will automatically include it in subsequent compilations and you do not need to ever explicitly call a function like WriteMass.

If you want to contribute your new feature to the code base, you must also write or update a unit test to include your new function.