tests/testthat/test-GridSearch.R

In joaofgoncalves/SegOptim: Optimization of Image Segmentation Parameters for Object-Based Image Analysis (OBIA)

context("Tests for grid search optimization systematic and random")

library(SegOptim)
library(terra)

# test_that("Test random search optimization", {
#   
#   
#   source("_CONFIG_.R")
#   skip_if(!dir.exists(OTB_BIN_PATH), "Cannot find OTB path")
#   skip_if(!file.exists(S1_SEGM_FEAT_PATH), "Cannot find input segm feature data")
#   skip_if(!file.exists(S1_CLASS_FEAT_PATH), "Cannot find input classification features data")
#   skip_if(!file.exists(S1_TRAIN_AREAS_PATH), "Cannot find input train data")
#   
#   bestParams <- searchOptimSegmentationParams(
#     S1_CLASS_FEAT_PATH,
#     S1_TRAIN_AREAS_PATH,            
#     
#     "OTB_LSMS",         
#     otbBinPath = OTB_BIN_PATH,
#     inputRstPath = S1_SEGM_FEAT_PATH,
#     
#     segmParamList = list(
#       SpectralRange = c(15, 20),
#       SpatialRange  = c(15, 20),
#       MinSize       = c(30,65)),
#     
#     optimMethod = "random",
#     rand.numIter = 1, 
#     rand.nneigh = 1, 
#     rand.initNeighs = (2 * rand.nneigh), 
#     rand.neighSizeProp = 0.025, 
#     rand.iter = 1,
#     trainThresh = 0.5,
#     segmStatsFuns = "mean",
#     classificationMethod = "RF",
#     classificationMethodParams = NULL,
#     balanceTrainData = FALSE,
#     balanceMethod = "ubUnder",
#     evalMethod = "5FCV",
#     trainPerc = 0.8,
#     evalMetric = "Kappa",
#     minTrainCases = 10,
#     minCasesByClassTrain = 5,
#     minCasesByClassTest = 5,
#     minImgSegm = 10,
#     verbose = FALSE,
#     parallel = FALSE,
#     seed = NULL)
#   
#   expect_is(bestParams, "list")
#   expect_equal(names(bestParams), c("bestFitValue", "bestParams"))
#   
# })
# 
# 
# test_that("Test grid search optimization", {
#   
#   
#   source("_CONFIG_.R")
#   skip_if(!dir.exists(OTB_BIN_PATH), "Cannot find OTB path")
#   skip_if(!file.exists(S1_SEGM_FEAT_PATH), "Cannot find input segm feature data")
#   skip_if(!file.exists(S1_CLASS_FEAT_PATH), "Cannot find input classification features data")
#   skip_if(!file.exists(S1_TRAIN_AREAS_PATH), "Cannot find input train data")
#   
#   bestParams <- searchOptimSegmentationParams(
#     S1_CLASS_FEAT_PATH,
#     S1_TRAIN_AREAS_PATH,            
#     
#     "OTB_LSMS",         
#     otbBinPath = OTB_BIN_PATH,
#     inputRstPath = S1_SEGM_FEAT_PATH,
#     
#     segmParamList = list(
#       SpectralRange = c(15, 20),
#       SpatialRange  = c(15, 20),
#       MinSize       = c(30,65)),
#     
#     optimMethod = "grid",
#     grid.searchSize = 1,
#     trainThresh = 0.5,
#     segmStatsFuns = "mean",
#     classificationMethod = "RF",
#     classificationMethodParams = NULL,
#     balanceTrainData = FALSE,
#     balanceMethod = "ubUnder",
#     evalMethod = "5FCV",
#     trainPerc = 0.8,
#     evalMetric = "Kappa",
#     minTrainCases = 10,
#     minCasesByClassTrain = 5,
#     minCasesByClassTest = 5,
#     minImgSegm = 10,
#     verbose = FALSE,
#     parallel = FALSE,
#     seed = NULL)
#   
#   expect_is(bestParams, "data.frame")
#   #expect_equal(names(bestParams), c("bestFitValue", "bestParams"))
#   
# })
# 
# 
# 
#