Download clock models from Zenodo — download

Downloads the given clock models from the Zenodo record and returns the input augmented with a path column pointing to the local files. Composite clocks (list_clocks) are single files; module clocks (list_module_clocks) come in one archive per module set, which is downloaded once and unpacked into dest_dir.

Usage

download_clocks(
  clocks,
  dest_dir = "clocks",
  record = .TAGE_ZENODO_RECORD,
  overwrite = FALSE,
  quiet = FALSE,
  timeout = 3600
)

Arguments

clocks: A data frame returned by list_clocks or list_module_clocks, or a character vector of composite clock file names.
dest_dir: Directory to save the models into. Created if needed. Default "clocks".
record: Character Zenodo record id. Defaults to the published record.
overwrite: Logical. Re-download files that already exist. Default FALSE.
quiet: Logical. Suppress progress messages. Default FALSE.
timeout: Seconds allowed per file. R's default of 60 s aborts the Bayesian ridge models, which are 0.9-2.4 GB each; elastic net models are about 1 MB and the module clock archives under 1 MB. Default 3600.

Value

The clocks data frame with an added path column.

Details

The whole record is about 64 GB, nearly all of it the 30 Bayesian ridge models. A partial or failed download is removed rather than left on disk, and every file is checked to be a pickle or a zip archive (Zenodo answers some errors with an HTML page, which would otherwise be saved under the model's name).

Examples

if (FALSE) { # \dontrun{
clocks <- list_clocks(type = "EN", outcome = "Mortality",
                      species = "Multispecies", tissue = "Multi-Tissue")
clocks <- download_clocks(clocks, dest_dir = "clocks")
model_paths <- list(
  scaled_diff = clocks$path[clocks$scaling == "Scaled"],
  yugene_diff = clocks$path[clocks$scaling == "YuGene"]
)

modules <- download_clocks(list_module_clocks(outcome = "Mortality", species = "Rodents"),
                           dest_dir = "clocks")
} # }