Common Errors#
Start with the model application, object, and variable named in the error. An application is a model configured to run on selected objects. These names tell you where to look: for example, a light calculation on a particular canopy may be missing LAI. PlantSimEngine catches many such problems before running any equation.
| Symptom | What it means | First action |
|---|---|---|
| A required input is missing | You have not supplied a starting value or a model that calculates it | Supply the value or connect a model that provides it |
| A selector finds too few or too many objects | For example, One expects one match but finds two | Check the object labels and where the selector searches |
| More than one source matches | Several models could supply the input | Name the intended application and, if needed, var |
| Several models set the same variable | PlantSimEngine cannot decide which value to keep | Choose one model, or use Updates if one model is meant to change another's result |
| Variable contracts differ | The connected variables declare different units or physical meanings | Check what each equation expects; add a conversion model where appropriate |
| A cadence is rejected | The chosen interval does not fit the base step or is not supported by the model | Choose compatible time steps |
| There is a dependency cycle | Two or more models each wait for a result from the others | Decide whether one input should come from the previous step or the equations must be solved together |
Example: the light model needs LAI#
Beer needs the canopy's LAI to calculate absorbed light. If no other model calculates LAI, supply its value yourself:
julia
using PlantSimEngine
using PlantSimEngine.Examples
model = CompositeModel(Beer(0.6);
status=(LAI=2.0,), environment=(Ri_PAR_f=100.0,))
simulation = run!(model)
final_state(simulation).aPPFD319.35424515612357
Alternatively, add a model that calculates LAI and supply that model's inputs. An arbitrary zero may remove a missing-input error while changing the scientific question. Use a value with a clear meaning.
Check the objects a selector actually sees#
Inspect Diagnostics.explain_objects(model) to see labels and parent links. For a plant application, Many(scale=:Leaf, within=Subtree()) searches that plant and its descendants. For a leaf application, the same selector starts at that leaf. Use SelfPlant() when the search should cover its whole plant.
Choose Many when the receiving model accepts a collection. Choosing it only to silence an error can pass a vector to an equation that expects one value.
Errors while writing a model#
Declare each status input with Required(T) if it must be supplied, or Default(value) if the model provides a fallback value. For example, use inputs_(::MyModel) = (LAI=Required(Real),) with the function qualified as PlantSimEngine.inputs_. Writing a number directly in inputs_ does not declare a required input. Define the equation as PlantSimEngine.run!(...) so that Julia adds your model's method to the package function.
For the complete sequence, see Write and test a first model. For further investigation, see Inspect a simulation and Dependency cycles.
On this page