Introduce a new ic tag

[pohaoc2]

Summary

Introduce a new ic tag for parameters to enforce constant initial conditions for cellular properties, while ensuring that parameters for daughter cells continue to be drawn from a distribution.

Changes

• Create a new distribution - DegenerateDistribution, which always returns the same value
• Draw a constant value from DegenerateDistribution for parent cell agents' parameters
• Draw values from other distributions defined in the input xml file for daughter agents' parameters

Examples

In the input .xml file, we can set a ic tag for population.parameter.

1. Initialize all cells with constant volume = 50.0

<population.parameter id="CELL_VOLUME" value="NORMAL(MU=2250,SIGMA=200)" ic="50"/>

2. Initialize all cells volume with the expected value of the distribution (only for NORMAL distribution)

<population.parameter id="CELL_VOLUME" value="NORMAL(MU=2250,SIGMA=200)" ic="MU"/>

3. Initialize all cells volume with the MIN value of the distribution (only for UNIFORM distribution)

<population.parameter id="CELL_VOLUME" value="UNIFORM(MIN=2000,MAX=3000)" ic="MIN"/>

4. Initialize all cells volume with the NaN value of the distribution if the ic tag does not match the distribution

<population.parameter id="CELL_VOLUME" value="UNIFORM(MIN=2000,MAX=3000)" ic="MU"/>

Comments

I overrided the parseParameter by introducing a new argument - MiniBox ics instead of modify the original one.

Estimated size: Small to Medium
Resolve #177.

[allison-li-1016]

Wouldn't the daughter cells also inherit the DegenerateDistribution? Maybe Im wrong, but Im of the understanding that the daughter cell inherits the parent cell's parameters in convert() for CellFactories in the proliferation module, so the parent cell's distribution for that parameter will always get passed on.

[allison-li-1016]

Wouldn't the daughter cells also inherit the DegenerateDistribution? Maybe Im wrong, but Im of the understanding that the daughter cell inherits the parent cell's parameters in convert() for CellFactories in the proliferation module, so the parent cell's distribution for that parameter will always get passed on.

Just to make sure I understand the code correctly, does line 45 in the parameters class make sure that the daughter cells get the original distribution?

if (distribution instanceof DegenerateDistribution) { distribution = popParameters.getDistribution(key, random);

[cainja]

Should degenerate distributions be added as an accessible distribution option from the input xml? If so, we should update minibox with the degenerate distribution.

[cainja]

So this is the important test - you need to make sure the daughter cells inherit the original distribution and not the degenerate distribution. This is how this happens in PatchModuleProliferation:

CellContainer newContainer = cell.make(newID, State.UNDEFINED, random);
Parameters newParameters = cell.getParameters();
PatchCell newCell =
        (PatchCell)
                newContainer.convert(
                        sim.getCellFactory(),
                        newLocation,
                        random,
                        newParameters);

This test isn't actually testing how new cells are created, which is related to @allison-li-1016's question.