From d4d9394a1f699d29029eac8b1e8a74ec8c6e231d Mon Sep 17 00:00:00 2001
From: William Perron <hey@wperron.io>
Date: Thu, 19 Oct 2023 16:02:56 -0400
Subject: [PATCH] add markdown-to-csv option

---
 cmd/md-fmt/main.go | 87 +++++++++++++++++++++++++++++++++++++++-------
 1 file changed, 75 insertions(+), 12 deletions(-)

diff --git a/cmd/md-fmt/main.go b/cmd/md-fmt/main.go
index 53383ff..8df8e82 100644
--- a/cmd/md-fmt/main.go
+++ b/cmd/md-fmt/main.go
@@ -3,11 +3,11 @@
 package main
 
 import (
+	"bufio"
 	"encoding/csv"
 	"flag"
 	"fmt"
 	"io"
-	"log"
 	"os"
 	"strings"
 )
@@ -16,6 +16,7 @@ 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() {
@@ -23,18 +24,27 @@ func main() {
 
 	fd, err := os.Open(*sourcePath)
 	if err != nil {
-		log.Fatalf("failed to open source file %s: %s\n", *sourcePath, err)
+		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(fd)
+	read := csv.NewReader(r)
 	read.Comma = sep
 	read.TrimLeadingSpace = true
 	read.LazyQuotes = *lazyQuotes
 
 	rec, err := read.Read()
 	if err != nil {
-		log.Fatalf("error reading from csv file: %s", err)
+		fmt.Fprintf(os.Stderr, "error reading from csv file: %s\n", err)
+		return
 	}
 
 	widths := make([]int, len(rec))
@@ -47,7 +57,8 @@ func main() {
 			break
 		}
 		if err != nil {
-			log.Fatalf("error reading from csv file: %s", err)
+			fmt.Fprintf(os.Stderr, "error reading from csv file: %s", err)
+			return
 		}
 
 		for i, col := range rec {
@@ -62,8 +73,8 @@ func main() {
 	pattern := fmt.Sprintf("|%s|\n", strings.Join(c, "|"))
 
 	// Reset file descriptor cursor and take new CSV reader from it
-	fd.Seek(0, 0)
-	read = csv.NewReader(fd)
+	r.Seek(0, 0)
+	read = csv.NewReader(r)
 	read.Comma = sep
 	read.TrimLeadingSpace = true
 	read.LazyQuotes = *lazyQuotes
@@ -71,21 +82,22 @@ func main() {
 	// Format header row
 	rec, err = read.Read()
 	if err != nil {
-		log.Fatalf("failed to read next csv record: %s", err)
+		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.Printf(pattern, curr...)
+	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.Printf(pattern, curr...)
+	fmt.Fprintf(w, pattern, curr...)
 
 	for {
 		rec, err := read.Read()
@@ -93,14 +105,65 @@ func main() {
 			break
 		}
 		if err != nil {
-			log.Fatalf("error reading from csv file: %s", err)
+			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.Printf(pattern, curr...)
+		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
+		}
 	}
 }