mergePipelineResults: mergePipelineResults

In plger/pipeComp: pipeComp pipeline benchmarking framework

Description

Merges the (non-aggregated) results of any number of runs of 'runPipeline' using the same PipelineDefinition (but on different datasets and/or using different parameters). First read the different sets of results using readPipelineResults, and pass them to this function.

Usage

mergePipelineResults(..., rr = NULL, verbose = TRUE)

Arguments

...	Any number of lists of pipeline results, each as produced by readPipelineResults
rr	Alternatively, the pipeline results can be passed as a list (in which case '...' is ignored)
verbose	Whether to print processing information

Value

A list of merged pipeline results.

Examples

# we produce 2 mock pipeline results:
pip <- mockPipeline()
datasets <- list( ds1=1:3, ds2=c(5,10,15) )
tmpdir1 <- paste0(tempdir(),'/')
res <- runPipeline(datasets, pipelineDef=pip, output.prefix=tmpdir1,
                   alternatives=list() )
alternatives <- list(meth1=c('log2','sqrt'), meth2='cumsum')
tmpdir2 <- paste0(tempdir(),'/')
res <- runPipeline(datasets, alternatives, pip, output.prefix=tmpdir2)
# we read the evaluation files:
res1 <- readPipelineResults(tmpdir1)
res2 <- readPipelineResults(tmpdir2)
# we merge them:
res <- mergePipelineResults(res1,res2)
# and we aggregate:
res <- aggregatePipelineResults(res)

plger/pipeComp documentation built on Nov. 2, 2021, 8:40 p.m.