Loads a tiny network and runs the NicheNet pipeline with low number of iterations in the optimization process. This way the pipeline runs in a reasonable time in order to test the code. Due to the random subsampling disconnected networks might be produced sometimes. If you see an error like "Error in if (sd(prediction_vector) == 0) ... missing value where TRUE/FALSE needed", the random subsampled input is not appropriate. In this case just interrupt and call again. This test ensures the computational integrity of the pipeline. If it fails during the optimization process, try to start it over several times, even restarting R. The unpredictability is related to codemlrMBO and nichenetr not being prepared to handle certain conditions, and it's also difficult to find out which conditions lead to which errors. At least 3 different errors appear time to time, depending on the input. It also seems like restarting R sometimes helps, suggesting that the entire system might be somehow stateful. You can ignore the  Parallelization was not stopped warnings on repeated runs.

## Usage

nichenet_test(...)

## Arguments

...

Passed to nichenet_main.

## Value

A named list with the intermediate and final outputs of the pipeline: networks, expression, optimized_parameters, weighted_networks and ligand_target_matrix.

## Examples

if (FALSE) {
nnt <- nichenet_test()
}