R/revert_parse_data.R

In origin: Explicitly Qualifying Namespaces by Automatically Adding 'pkg::' to Functions

# converse parsed data back to its input
revert_parse_data <- function(parse_data,
                              script,
                              recover_empty_lines = TRUE) {
  
  # Strings that cover multiple lines
  to_split <- parse_data$token == "STR_CONST" &
    parse_data$line1 != parse_data$line2
  
  if (any(to_split)) {
    dfs <- parse_data[to_split, ]
    
    # how many lines does each string cover
    rep <- dfs$line2 - dfs$line1 + 1
    
    # recover one row for each line
    # get string for each line
    string_cut <- unlist(strsplit(dfs$text, split = "\n"))
    # repeat ddf-rows as often
    dfs <- dfs[rep(seq_len(nrow(dfs)), rep), ]
    
    # assign string for eahc line
    dfs$text <- string_cut
    
    # helper for info recovery
    dfs$rep <- unlist(lapply(rep, seq_len)) - 1
    
    # start line is initial start + number of lines covered
    dfs$line1 <- dfs$line1 + dfs$rep
    #start line is end line
    dfs$line2 <- dfs$line1
    
    # recover column values
    dfs[dfs$rep > 0, "col1"] <- 1
    dfs[dfs$rep > 0, "col2"] <- nchar(dfs[dfs$rep > 0, "text"])
    dfs[dfs$rep == 0, "col2"] <-
      nchar(dfs[dfs$rep == 0, "text"]) + dfs[dfs$rep == 0, "col1"]
    dfs$rep <- NULL
    
    # combine initial and recovered data
    parse_data <- rbind(parse_data[!to_split, ],
                        dfs)
    
    # bring in correct order
    parse_data <- parse_data[order(parse_data$file,
                                   parse_data$line1,
                                   parse_data$col1), ]
  }
  
  res <- by(data = parse_data,
            # consider each line separately
            # TODO: exclude loop
            INDICES = list(parse_data$line1, parse_data$file),
            FUN = function(args) {
              # determine number of tokens in the line
              n_cases <- length(args$text)
              
              # determine width of white space needed prior to each token
              # it is its starting position minus 
              # the end position of the precessing token minus 1 
              # since positions are including
              n_ws <- args$col1 - c(0, args$col2[-n_cases]) - 1
              
              # create needed whitespace
              ws <- strrep(" ", times = n_ws)
              
              # add whitespaces to the tokens
              out <- paste(ws, args$text, collapse = "", sep = "")
              return(out)
            })
  res <- as.character(res)
  
  if (is_rmd_file(parse_data$file[1]) && 
      # only do this in case it is called via _dir, _pkg, _file,
      # not via selction/text. 
      any(grepl(x = script,
                pattern = "```",
                fixed = TRUE))) {
    r_lines <- extract_r_chunk_lines(script)
    # replace r-chunks with result
    # this preserves the non-r lines as well as blanks (nzchar) as they were
    script[r_lines & nzchar(trimws(script))] <- res[nzchar(trimws(res))]
    return(script)
    
  } else {
    # recover empty lines if needed
    if (recover_empty_lines) {
      out <- character(max(parse_data$line1))
      out[unique(parse_data$line1)] <- res
      return(out)
    } else {
      return(res)
    }
  }
}