Run weathergenr end-to-end (generate and evaluate)

Top-level convenience wrapper that runs: generate_weather() prepare_evaluation_data() evaluate_weather_generator()

All execution and evaluation settings are taken from config. Optional logging writes console output to a timestamped file in out_dir while continuing to display output on the console.

Usage

run_weather_generator(
  obs_data,
  obs_grid,
  obs_dates,
  out_dir,
  config,
  eval_max_grids = 25L,
  log_messages = FALSE
)

Arguments

obs_data

Observed data (e.g. ncdata$data).

obs_grid

Observed grid metadata (e.g. ncdata$grid).

obs_dates

Observed dates (e.g. ncdata$date).

out_dir

Character. Output directory.

config

List. Full simulation/evaluation configuration.

eval_max_grids

Integer. Maximum number of grids to evaluate.

log_messages

Logical. If TRUE, save console output to log_YYYYMMDD_HHMMSS.txt in out_dir.

Value

A list with components:

• gen_output: output of generate_weather()

• evaluation: output of evaluate_weather_generator()

• log_path: path to the log file (or NULL if log_messages=FALSE)