toolkit/cmd/md-fmt/main.go

// md-fmt takes a csv or tsv input file and outputs a formatted markdown table
// with the data.
package main

import (
	"bufio"
	"encoding/csv"
	"flag"
	"fmt"
	"io"
	"os"
	"strings"
)

var (
	sourcePath = flag.String("source", "", "path to the input file")
	separator  = flag.String("sep", ",", "separator character to use when reading the csv file")
	lazyQuotes = flag.Bool("lazy-quotes", false, "controls the lazy-quotes setting on the csv reader")
	toCSV      = flag.Bool("to-csv", false, "parses markdown table and encodes it to csv (opposite operation)")
)

func main() {
	flag.Parse()

	fd, err := os.Open(*sourcePath)
	if err != nil {
		fmt.Fprintf(os.Stderr, "failed to open source file %s: %s\n", *sourcePath, err)
	}

	if *toCSV {
		markdownToCSV(fd, os.Stdout)
	} else {
		csvToMarkdown(fd, os.Stdout)
	}
}

func csvToMarkdown(r io.ReadSeeker, w io.Writer) {
	sep := []rune(*separator)[0]
	read := csv.NewReader(r)
	read.Comma = sep
	read.TrimLeadingSpace = true
	read.LazyQuotes = *lazyQuotes

	rec, err := read.Read()
	if err != nil {
		fmt.Fprintf(os.Stderr, "error reading from csv file: %s\n", err)
		return
	}

	widths := make([]int, len(rec))
	for i, col := range rec {
		widths[i] = max(widths[i], len(col))
	}
	for {
		rec, err := read.Read()
		if err == io.EOF {
			break
		}
		if err != nil {
			fmt.Fprintf(os.Stderr, "error reading from csv file: %s", err)
			return
		}

		for i, col := range rec {
			widths[i] = max(widths[i], len(col))
		}
	}

	c := make([]string, len(widths))
	for i := 0; i < len(c); i++ {
		c[i] = " %-*s "
	}
	pattern := fmt.Sprintf("|%s|\n", strings.Join(c, "|"))

	// Reset file descriptor cursor and take new CSV reader from it
	r.Seek(0, 0)
	read = csv.NewReader(r)
	read.Comma = sep
	read.TrimLeadingSpace = true
	read.LazyQuotes = *lazyQuotes

	// Format header row
	rec, err = read.Read()
	if err != nil {
		fmt.Fprintf(os.Stderr, "failed to read next csv record: %s", err)
		return
	}

	curr := make([]any, 0, 2*len(widths))
	for i := range widths {
		curr = append(curr, widths[i], rec[i])
	}
	fmt.Fprintf(w, pattern, curr...)

	// Format header separator row
	curr = curr[:0] // empty slice but preserve capacity
	for i := range widths {
		curr = append(curr, widths[i], strings.Repeat("-", widths[i]))
	}
	fmt.Fprintf(w, pattern, curr...)

	for {
		rec, err := read.Read()
		if err == io.EOF {
			break
		}
		if err != nil {
			fmt.Fprintf(os.Stderr, "error reading from csv file: %s", err)
			return
		}

		curr = curr[:0]
		for i := range widths {
			curr = append(curr, widths[i], rec[i])
		}
		fmt.Fprintf(w, pattern, curr...)
	}
}

func markdownToCSV(r io.ReadSeeker, w io.Writer) {
	cw := csv.NewWriter(w)
	defer func() {
		cw.Flush()
		if err := cw.Error(); err != nil {
			fmt.Fprintf(os.Stderr, "error occured during scan: %s", err)
		}
	}()

	// for now we're assuming there's only one table in a markdown file, this
	// flag keeps track of whether or not we've seen a table already, and is
	// used to exit the loop early once we're done reading the table, ignoring
	// subsequent lines in the file.
	var seen bool
	scanner := bufio.NewScanner(r)
	for scanner.Scan() {
		if err := scanner.Err(); err != nil {
			fmt.Fprintln(os.Stderr, "reading standard input:", err)
		}

		line := strings.TrimSpace(scanner.Text())
		if !(strings.HasPrefix(line, "|") && strings.HasSuffix(line, "|")) {
			if seen {
				break // exit early if we've seen a table already
			}
			continue // keep going until we see a table
		}

		// the rest of the body only applies to lines that are part of a table
		// i.e. that start and end with a pipe character (`|`)
		seen = true

		line = strings.Trim(line, "|")
		parts := strings.Split(line, "|")
		for i := 0; i < len(parts); i++ {
			parts[i] = strings.TrimSpace(parts[i])
		}

		// ignore separator line
		if parts[0] == strings.Repeat("-", len(parts[0])) {
			continue
		}

		if err := cw.Write(parts); err != nil {
			fmt.Fprintf(os.Stderr, "failed to write to output: %s\n", err)
			return
		}
	}
}

func max(a, b int) int {
	if a > b {
		return a
	}
	return b
}