Port get_schema_via_mysqldump and other useful DB metadata functions to redcapcustodian

[ChemiKyle]

Archiving and shutting down projects has created several (possibly generically) useful functions, we should consider porting these functions to redcapcustodian.

The code from STP is as follows:

get_schema_via_mysqldump <- function(table_name, host, database, username, password) {
  command <- paste(
    "mysqldump",
    "--no-data --skip-lock-tables --triggers --routines --events --skip-add-locks --skip-comments --no-tablespaces",
    paste0("--host=", host),
    paste0("--user=", username),
    "-p",
    database,
    table_name,
    "| sed -e 's/AUTO_INCREMENT=[[:digit:]]* //'",
    "| sed '/50013 DEFINER=/d'"
  )

  schema <- system(
    command = command,
    input = password,
    intern = TRUE
  )
  project_name <- tolower(word(table_name, 1, sep = "_"))
  schema_filename <- paste0(output_dir, "/", project_name, "_schema", ".sql")
  write(schema, schema_filename)
}