R/buildPerformanceData.R

In lucasmpavelski/metaOpt: Meta-Learning for Optimization

build_performance_data <- function(
                                   problem_space,
                                   algorithm_space,
                                   solve_function,
                                   irace_scenario = irace::defaultScenario(),
                                   parallel = 1,
                                   quiet = FALSE,
                                   cache_folder = here("data", "performance")) {
  experiments <- expand.grid(problem = problem_space@problems, algorithm = algorithm_space@algorithms)
  results <- pmap_dfr(experiments, function(problem, algorithm) {
    inst_scenario <- irace_scenario
    inst_scenario$instances <- problem@instances
    inst_scenario$targetRunnerData <- list(
      problem = problem,
      algorithm = algorithm
    )
    if (parallel <= 1) {
      inst_scenario$targetRunner <- wrap_irace_target_runner(solve_function)
    } else {
      inst_scenario$targetRunnerParallel <- wrap_irace_target_runner_parallel(solve_function, parallel)
    }
    parameters <- algorithm@parameters

    folder <- file.path(cache_folder, problem@name, algorithm@name)
    if (!is.na(cache_folder)) {
      dir.create(folder, showWarnings = FALSE, recursive = TRUE)
      inst_scenario$logFile <- file.path(folder, "log.Rdata")
      result_file <- file.path(folder, "result.rds")
    } else {
      inst_scenario$logFile <- "/dev/null"
    }

    tunning_result <- NA
    if (!is.na(cache_folder) && file.exists(result_file)) {
      if (!quiet) {
        cat("Reading cached ", result_file, ".")
      }
      tunning_result <- readRDS(result_file)
    } else {
      if (quiet) {
        tunning_result <- quietly(irace)(inst_scenario, parameters)$result
      } else {
        tunning_result <- irace(inst_scenario, parameters)
      }
    }

    if (!is.na(cache_folder)) {
      saveRDS(tunning_result, result_file)
    }

    tibble(
      algorithm_names = algorithm@name,
      algorithms = list(algorithm),
      problems = list(problem),
      problem_names = problem@name,
      results = list(tunning_result)
    )
  })
}