initial commit

Added repo for mock-up forecasts in R, but pushed to the wrong branch because I wasn't paying enough attention.

R/forecast_functions.R

+
+#' Make a mock-up daily forecast by sampling from input catalogue
+#' Returns random samples from a catalogue. Assumes required duration is 1 day but can be generalised.
+#'
+#' @param catalog catalogue to sample events from
+#' @param date date for which to generate catalogue
+#' @param n_cats number of catalogues to simulate
+#'
+#' @return r dataframe object containing forecast columns
+#' @export
+
+mock_daily_forecast <- function(catalog, date, n_cats ){
+  ## Set up list of forecasts
+  forecasts <-  list(length=length(n_cats))
+  ## Set up catalog we're using
+  cat_to_use <- catalog[as.POSIXct(catalog$time_string, tz="UTC") < date,]
+
+  ## set up rate
+  rate <- nrow(cat_to_use)/abs(as.numeric(difftime(as.POSIXct(cat_to_use$time_string[1]), as.POSIXct(cat_to_use$time_string[nrow(cat_to_use)], tz="UTC"), units="days")))
+  for(i in 1:n_cats){
+    ## choose number of events for catalogue from poisson with rate = catalogue rate
+    num_events_cat <- rpois(1, rate)
+    newcat <-  cat_to_use[round(runif(num_events_cat, 1, length(cat_to_use))),]
+    ## change the catalog_id to reflect this mock-up:
+    ## these events get catalog_id from whatever catalog number we are currently making
+    newcat$catalog_id <- i
+    forecasts[[i]] <- newcat
+  }
+  forecast_df = do.call(rbind, forecasts)
+  return(forecast_df)
+}
+
+#' Write forecast to file
+#' Write a forecast to a txt file that will be read by pycsep
+#'
+#' @param date string with date of forecast (for file naming)
+#' @param forecast_df dataframe of forecast data (assumes structure of mock_daily_forecast)
+#' @param output_path path of location to save forecast
+#' @return Writes forecast to file at specified location
+#' @export
+
+write_forecast <- function(date, forecast_df, output_path){
+  forecast_name <- paste0(output_path,"mockup_forecast_", date, ".txt")
+  write.table(format(as.data.frame(forecast_df, digits = 6)), forecast_name, quote = FALSE, row.names = FALSE, col.names = FALSE, sep=",")
+}
+