fix(FIMSFrame): Assume yyyy-mm-dd format for dates

When writing a data frame to a csv and reading it back in, the date
formatting can be lost, e.g., 0001-01-01 turns into 1-1-1. FIMS requires
a yyyy-mm-dd format. Now the use of the as.Date() function will create
a date object from a character object but only if it is in the correct
format. If it is in the wrong format, e.g., yyyy/mm/dd, then the function
will error out.

Right now the start_year and end_year are formatted as integers because for
plotting, I thought it would be better to have year 1 versus year 0001 but
we can change this. @ian-taylor-NOAA what do you think?

R/fimsframe.R

@@ -226,11 +226,24 @@ setValidity(
     }
 
     errors <- c(errors, validate_data_colnames(object@data))
+
+    # Add checks for other slots
+    # Check the format for acceptable variants of the ideal yyyy-mm-dd
+    grepl_datestart <- grepl(
+      "[0-9]{1,4}-[0-9]{1,2}-[0-9]{1-2}",
+      data_mile1[["datestart"]]
+    )
+    grepl_dateend <- grepl(
+      "[0-9]{1,4}-[0-9]{1,2}-[0-9]{1-2}",
+      data_mile1[["dateend"]]
+    )
+    if (!all(grepl_datestart)) {
+      errors <- c(errors, "datestart must be in 'yyyy-mm-dd' format")
     }
+    if (!all(grepl_dateend)) {
+      errors <- c(errors, "dateend must be in 'yyyy-mm-dd' format")
     }
 
-    # TODO: Add checks for other slots
-
     # Return
     if (length(errors) == 0) {
       return(TRUE)
@@ -299,14 +312,23 @@ FIMSFrame <- function(data) {
       paste(errors, sep = "\n", collapse = "\n")
     )
   }
-  # Get the earliest and latest year of data and use to calculate n years for
-  # population simulation
-  start_year <- as.integer(
-    strsplit(min(data[["datestart"]], na.rm = TRUE), "-")[[1]][1]
-  )
-  end_year <- as.integer(
-    strsplit(max(data[["dateend"]], na.rm = TRUE), "-")[[1]][1]
-  )
+  # datestart and dateend need to be date classes so leading zeros are present
+  # but writing and reading from csv file removes the classes so they must be
+  # enforced here
+  # e.g., 0004-01-01 for January 01 0004
+  date_formats <- c("%Y-%m-%d")
+  data[["datestart"]] <- as.Date(data[["datestart"]], tryFormats = date_formats)
+  data[["dateend"]] <- as.Date(data[["dateend"]], tryFormats = date_formats)
+
+  # Get the earliest and latest year formatted as a string of 4 integers
+  start_year <- as.integer(format(
+    as.Date(min(data[["datestart"]], na.rm = TRUE), tryFormats = date_formats),
+    "%Y"
+  ))
+  end_year <- as.integer(format(
+    as.Date(max(data[["dateend"]], na.rm = TRUE), tryFormats = date_formats),
+    "%Y"
+  ))
   n_years <- as.integer(end_year - start_year + 1)
   years <- start_year:end_year