csv

#!/usr/bin/env Rscript

# csv - a linux CLI for nicely displaying and summarizing CSV files

suppressPackageStartupMessages(library(optparse))
suppressPackageStartupMessages(library(data.table))

file_ext <- function(x) {
	pos <- regexpr("\\.([[:alnum:]]+)$", x)
	tolower(ifelse(pos > -1L, substring(x, pos + 1L), ""))
}

opt.list <- list(
	make_option(c('-a','--all'), action = 'store_true', default = FALSE,
				help = 'Print all rows.'),
	make_option(c('-t','--tail'), action = 'store_true', default = FALSE,
				help = 'Print only tail. By default only the head is printed.\n\t\tIgnored if the [-a|--all] flag is set.'),
	make_option(c('-n','--num-rows'), type = 'integer', default = 6, metavar = 'INTEGER',
				help = 'Specify number of rows to print (default is %default). Ignored\n\t\tif the [-a|--all] flag is set.'),
	make_option(c('-s','--summary'), action = 'store_true', default = FALSE,
				help = 'Summarize all rows by column.'),
	make_option(c('-c','--col-names'), action = 'store_true', default = FALSE,
				help = 'Only print column names.'),
	make_option(c('-k','--keep-cols'), type = 'character', default = NULL, metavar = 'STRING',
				help = 'Column selection. Provde names or numbers of columns to keep\n\t\tor drop, separated by commas (e.g., "var1,var5", or "1:4,8").\n\t\tUse "!" to drop columns (e.g., "!1,!5" or "!var1,!var5").'),
	make_option(c('-d','--drop-cols'), type = 'character', default = NULL, metavar = 'STRING',
				help = 'Column selection. Provde names or numbers of columns to keep\n\t\tor drop, separated by commas (e.g., "var1,var5", or "1:4,8").\n\t\tUse "!" to drop columns (e.g., "!1,!5" or "!var1,!var5").'),
	make_option(c('-D','--dims'), action = 'store_true', default = FALSE,
				help = 'Only print CSV dimensions (i.e., number of rows and columns).'),
	make_option(c('-N','--no-header'), action = 'store_true', default = FALSE,
				help = 'Flag that CSV has no header row.'),
	make_option(c('-i','--remove-index'), action = 'store_true', default = FALSE,
				help = 'Do not print row index.'),
	make_option(c('-v','--verbose'), action = 'store_true', default = FALSE,
				help = 'Prints progress and information messages.')
)

opt.parser <- OptionParser(option_list = opt.list,
						   usage = 'usage: %prog <input.csv> [options]\n',
						   description = paste0('A simple command-line tool to print a CSV file to the terminal.\n\n',
						   					 'Arguments:\n',
						   					 '\t<input.csv>\n',
						   					 '\t\tPath to CSV file.'))

arguments <- parse_args(opt.parser, positional_arguments = T)
pos.args <- arguments$args
opt.args <- arguments$options

csv.file <- pos.args[1]
print.all <- opt.args$all
print.tail <- opt.args$tail
print.nrows <- opt.args$`num-rows`
print.sum <- opt.args$summary
print.colnames <- opt.args$`col-names`
keep.cols <- opt.args$`keep-cols`
drop.cols <- opt.args$`drop-cols`
print.dims <- opt.args$dims
hdr.exists <- !opt.args$`no-header`
rem.ind <- opt.args$`remove-index`
verbose <- opt.args$verbose

if (!is.null(keep.cols) & !is.null(drop.cols)) stop('use either "-k|--keep-cols" or "-d|--drop-cols", but not both. ')

if (is.na(csv.file)) {
	print_help(opt.parser)
} else if (!file.exists(csv.file)) {
	stop('input file does not exist.')
} else if (file_ext(csv.file) != 'csv') {
	stop('input file must be a CSV.')
} else {
	if (verbose) cat(paste('Input table exists ...'), fill = T)

	tmp.df <- fread(csv.file, header = F, nrows = 1, showProgress = F, data.table = F)

	if (hdr.exists) {
		hdr <- as.character(as.vector(tmp.df[1,]))
	} else {
		hdr <- colnames(tmp.df)
	}

	if (print.colnames) {
		if (hdr.exists) {
			cat(paste(hdr, collapse = ', '), fill = T)
		} else {
			stop('no header.')
		}
	} else {

		# count total number of rows in CSV
		tot.num.rows <- as.numeric(system(paste('sed -n "$="', csv.file), intern = T)) # includes header
		if (hdr.exists) tot.num.rows <- tot.num.rows - 1

		if (print.dims) {
			cat(paste(tot.num.rows, 'rows,', length(hdr), 'columns'), fill = T)
		} else {

			if (print.nrows > tot.num.rows) {
				print.all <- T
			}

			if (hdr.exists) {
				skip.nrows <- 1 # skip header
			} else {
				skip.nrows <- 0
			}

			if (print.all | print.sum) {
				print.nrows <- Inf
				substr <- 'full'
			} else if (print.tail) {
				skip.nrows <- tot.num.rows - print.nrows + skip.nrows
				substr <- paste('last', print.nrows, 'rows of')
			} else {
				tot.num.rows <- print.nrows
				substr <- paste('first', print.nrows, 'rows of')
			}

			# get total number of columns
			tot.num.cols <- length(hdr)
			cols.num <- 1:tot.num.cols

			if (!is.null(keep.cols) | !is.null(drop.cols)) {

				if (!is.null(keep.cols)) user.cols <- keep.cols
				if (!is.null(drop.cols)) user.cols <- drop.cols

				# split input into vector of strings
				cols.splt <- unlist(strsplit(user.cols, ','))

				# if a number sequence exists in vector, expand it
				cols.seq <- eval(parse(text = cols.splt[grepl(':', cols.splt)]))

				# combine sequence back with other elements of vector
				cols.str <- c(cols.seq, cols.splt[!grepl(':', cols.splt)])

				# convert vector from character to numeric
				cols.num <- suppressWarnings(as.numeric(cols.str))

				# count number of NAs created when converting from character to numeric
				cols.na.cnt <- sum(is.na(cols.num))

				# if no NAs, then user provided column numbers, otherwise user provided column names
				if (cols.na.cnt == 0) {
					if (!all(abs(cols.num) %in% 1:tot.num.cols)) stop('all columns must exist in CSV.')
				} else if (cols.na.cnt == length(cols.str)) {
					if (!all(gsub('^-', '', cols.str) %in% hdr)) stop('all columns must exist in CSV.')
					cols.num <- charmatch(cols.str, hdr) # return column numbers to keep, in ascending order
				} else {
					stop('must provide either column names or numbers, not both.')
				}

				# remove duplicate column selections
				cols.num <- unique(cols.num)

				# invert column selection if user chose to drop columns
				if (!is.null(drop.cols)) cols.num <- (1:tot.num.cols)[-cols.num]

				# final column name selection
				hdr <- hdr[cols.num]

			}

			if (verbose) cat(paste('Reading', substr, 'table ...'), fill = T)
			df <- fread(csv.file, select = cols.num, nrows = print.nrows, skip = skip.nrows, header = F, col.names = hdr, showProgress = verbose, data.table = F, stringsAsFactors = T)

			width <- system('tput cols', intern = T)
			options(width = width)

			if (print.sum) {
				if (verbose) cat('Summarizing table ...', fill = T)
				print(summary(df))
			} else {
				if (verbose) cat('Printing table ...', fill = T)
				if (!hdr.exists) skip.nrows <- skip.nrows + 1
				if (rem.ind) {
					row.names <- F
				} else {
					row.names <- as.character(skip.nrows:tot.num.rows)
				}
				print.data.frame(df, row.names = row.names)
			}
		}
	}
}