wip: fighting comments

io/transforms/gsheet2csv/README.md

@@ -15,6 +15,8 @@ This does surprisingly little - you should probably just handle the boilerplate
 - can preserve comments
 - swaps `\r` (Windows) for `\n` (Unix) and ensures trailing newline (a la `encoding/csv`)
 
+Note: The Google Sheet must be shared to **Anyone with the link**.
+
 ## Usage
 
 Same as `encoding/csv` (embedded), but with two extra options:
@@ -37,16 +39,15 @@ func main() {
 		fmt.Fprintf(os.Stderr, "Usage: %s <url>\n", os.Args[0])
 		os.Exit(1)
 	}
-	url := os.Args[1]
+	urlOrPath := os.Args[1]
 
-	gsr := gsheet2csv.NewReaderFromURL(url)
+	gsr := gsheet2csv.NewReaderFrom(urlOrPath)
 	records, err := gsr.ReadAll()
 	if err != nil {
-		fmt.Fprintf(os.Stderr, "Error reading from %s\n", gsr.URL)
+		fmt.Fprintf(os.Stderr, "Error reading from %s: %v\n", gsr.URL, err)
 		os.Exit(1)
 	}
 
-    // distinguishes between comments and quoted fields
 	csvw := gsheet2csv.NewWriter(os.Stdout)
 	csvw.Comment = gsr.Comment
 	if err := csvw.WriteAll(records); err != nil {

io/transforms/gsheet2csv/cmd/gsheet2csv/main.go

@@ -1,6 +1,7 @@
 package main
 
 import (
+	"encoding/csv"
 	"errors"
 	"flag"
 	"fmt"
@@ -19,8 +20,21 @@ const (
 	unitSeparator   = "\x1f"
 )
 
+type CSVReader interface {
+	Read() ([]string, error)
+	ReadAll() ([][]string, error)
+}
+
+type CSVWriter interface {
+	Write([]string) error
+	WriteAll([][]string) error
+	Flush()
+	Error() error
+}
+
 func main() {
 	var commentArg string
+	newline := "\n"
 	format := "CSV"
 	delim := ','
 	if strings.Contains(os.Args[0], "tsv") {
@@ -29,13 +43,17 @@ func main() {
 	}
 
 	// Parse command-line flags
-	flag.StringVar(&commentArg, "comment", "#", "treat lines beginning with this rune as comments")
+	flag.StringVar(&commentArg, "comment", "#", "treat lines beginning with this rune as comments, 0 to disable")
 	outputFile := flag.String("o", "", "Output "+format+" file (default: stdout)")
 	delimString := flag.String("d", string(delim), "field delimiter to use for output file ('\\t' for tab, '^_' for Unit Separator, etc)")
 	useCRLF := flag.Bool("crlf", false, "use CRLF (\\r\\n) as record separator")
 	urlOnly := flag.Bool("print-url", false, "don't download, just print the Google Sheet URL")
 	parseOnly := flag.Bool("print-ids", false, "don't download, just print the Doc ID and Sheet ID (gid)")
 	rawOnly := flag.Bool("raw", false, "don't parse, just download")
+	noReadComments := flag.Bool("no-read-comments", false, "strip comments when reading (gsheet-only)")
+	noWriteComments := flag.Bool("no-write-comments", false, "strip comments when writing")
+	readStyle := flag.String("read-style", "gsheet", "'gsheet' or 'rfc' to read either as a gsheet or rfc CSV")
+	writeStyle := flag.String("write-style", "rfc", "'gsheet' or 'rfc' to write either as a gsheet or rfc CSV")
 	flag.Usage = func() {
 		fmt.Fprintf(os.Stderr, "Usage: %s [flags] <google-sheet-url-or-file-path>\n", os.Args[0])
 		fmt.Fprintf(os.Stderr, "Converts a Google Sheet to %s format.\n\n", format)
@@ -88,6 +106,10 @@ func main() {
 	}
 	delim, _ = utf8.DecodeRuneInString(*delimString)
 
+	if *useCRLF {
+		newline = "\r\n"
+	}
+
 	var rc io.ReadCloser
 	if strings.HasPrefix(url, "https://") || strings.HasPrefix(url, "http://") {
 		docid, gid := gsheet2csv.ParseIDs(url)
@@ -123,7 +145,10 @@ func main() {
 		}
 		rc = f
 	}
-	fmt.Fprintf(os.Stderr, "\n")
+
+	if out == os.Stdout {
+		fmt.Fprintf(os.Stderr, "\n")
+	}
 
 	if *urlOnly || *parseOnly {
 		os.Exit(0)
@@ -138,22 +163,53 @@ func main() {
 		return
 	}
 
-	comment, _ := utf8.DecodeRuneInString(commentArg)
+	var comment rune
+	if commentArg == "0" {
+		comment = 0
+	} else {
+		comment, _ = utf8.DecodeRuneInString(commentArg)
+	}
 
 	// Create a reader for the Google Sheet
-	gsr := gsheet2csv.NewReader(rc)
-	gsr.QuotedComments = false
-	gsr.Comment = 0
-	gsr.ReuseRecord = true
+	var csvr CSVReader
+	if *readStyle == "rfc" {
+		rfcr := csv.NewReader(rc)
+		rfcr.Comment = comment
+		csvr = rfcr
+	} else {
+		gsr := gsheet2csv.NewReader(rc)
+		gsr.QuotedComments = false
+		if *noReadComments {
+			gsr.Comment = '#'
+		} else {
+			gsr.Comment = 0
+		}
+		gsr.ReuseRecord = true
+		csvr = gsr
+	}
 
 	// Create CSV writer
-	csvw := gsheet2csv.NewWriter(out)
-	csvw.Comma = delim // Set delimiter to tab for TSV
-	csvw.Comment = comment
-	csvw.UseCRLF = *useCRLF
+	var csvw CSVWriter
+	if *writeStyle == "gsheet" {
+		gsw := gsheet2csv.NewWriter(out)
+		gsw.Comment = comment
+		gsw.Comma = delim // Set delimiter to tab for TSV
+		gsw.UseCRLF = *useCRLF
+		csvw = gsw
+	} else {
+		rfcw := csv.NewWriter(out)
+		rfcw.Comma = delim
+		rfcw.UseCRLF = *useCRLF
+		csvw = rfcw
+	}
+
+	const quote = `"`
+	const dquote = quote + quote
+	var commaStr = string(delim)
+	var commentStr = string(comment)
 	for {
 		// Convert each record
-		record, err := gsr.Read()
+		record, err := csvr.Read()
 		if err != nil {
 			if errors.Is(err, io.EOF) {
 				break
@@ -162,7 +218,39 @@ func main() {
 			os.Exit(1)
 		}
 
-		if err := csvw.Write(record); err != nil {
+		lastNonEmpty := len(record) - 1
+		for lastNonEmpty > -1 && len(record[lastNonEmpty]) == 0 {
+			lastNonEmpty -= 1
+		}
+
+		if *writeStyle != "rfc" || comment == 0 || len(record) == 0 || !strings.HasPrefix(record[0], commentStr) {
+			if *noWriteComments && len(record) > 0 && strings.HasPrefix(record[0], commentStr) && lastNonEmpty == 0 {
+				continue
+			}
+
+			if err := csvw.Write(record); err != nil {
+				fmt.Fprintf(os.Stderr, "Error writing "+format+": %v\n", err)
+				os.Exit(1)
+			}
+			continue
+		}
+
+		if lastNonEmpty == 0 {
+			record = record[:1]
+			if *noWriteComments {
+				continue
+			}
+		} else {
+			for i, f := range record {
+				if i == 0 || strings.Contains(f, quote) || strings.Contains(f, string(commaStr)) {
+					f = strings.ReplaceAll(f, quote, dquote)
+					record[i] = quote + f + quote
+				}
+			}
+		}
+
+		csvw.Flush()
+		if _, err := out.Write([]byte(strings.Join(record, commaStr) + newline)); err != nil {
 			fmt.Fprintf(os.Stderr, "Error writing "+format+": %v\n", err)
 			os.Exit(1)
 		}
@@ -173,4 +261,8 @@ func main() {
 		fmt.Fprintf(os.Stderr, "Error flushing "+format+" writer: %v\n", err)
 		os.Exit(1)
 	}
+
+	if out != os.Stdout {
+		fmt.Fprintf(os.Stderr, "wrote %s\n", *outputFile)
+	}
 }

io/transforms/gsheet2csv/cmd/gsheet2tsv/main.go

@@ -123,7 +123,10 @@ func main() {
 		}
 		rc = f
 	}
-	fmt.Fprintf(os.Stderr, "\n")
+
+	if out == os.Stdout {
+		fmt.Fprintf(os.Stderr, "\n")
+	}
 
 	if *urlOnly || *parseOnly {
 		os.Exit(0)
@@ -173,4 +176,8 @@ func main() {
 		fmt.Fprintf(os.Stderr, "Error flushing "+format+" writer: %v\n", err)
 		os.Exit(1)
 	}
+
+	if out != os.Stdout {
+		fmt.Fprintf(os.Stderr, "wrote %s\n", *outputFile)
+	}
 }

io/transforms/gsheet2csv/fixtures/example.go

@@ -15,12 +15,12 @@ func main() {
 		fmt.Fprintf(os.Stderr, "Usage: %s <url>\n", os.Args[0])
 		os.Exit(1)
 	}
-	url := os.Args[1]
+	urlOrPath := os.Args[1]
 
-	gsr := gsheet2csv.NewReaderFromURL(url)
+	gsr := gsheet2csv.NewReaderFrom(urlOrPath)
 	records, err := gsr.ReadAll()
 	if err != nil {
-		fmt.Fprintf(os.Stderr, "Error reading from %s\n", gsr.URL)
+		fmt.Fprintf(os.Stderr, "Error reading from %s: %v\n", gsr.URL, err)
 		os.Exit(1)
 	}
 

io/transforms/gsheet2csv/fixtures/sheet.csv

@@ -0,0 +1,18 @@
+# this is a comment
+"# this is, well, a quoted comment"
+"# this is a ""super""-quoted comment"
+Key,Value,
+Name,55,
+Girlfriend's Age,55,
+,,
+My IQ,55,
+,55,
+"Key,with,Comma",,
+,"Value,with,Comma",
+"Quoted ""Key""",Normal Value,
+Normal Key,"Quoted ""Value""",
+"Quoted ""Key""",,
+,"Quoted ""Value""",
+x,y,z
+"# comment with trailing comma,"
+"#1",2,#3

io/transforms/gsheet2csv/gsheet2csv.go

@@ -17,6 +17,7 @@ import (
 	"fmt"
 	"io"
 	"net/http"
+	"os"
 	"strings"
 	"unicode/utf8"
 )
@@ -39,6 +40,22 @@ type Reader struct {
 	err            error
 }
 
+func NewReaderFrom(urlOrPath string) *Reader {
+	if strings.HasPrefix(urlOrPath, "https://") || strings.HasPrefix(urlOrPath, "http://") {
+		return NewReaderFromURL(urlOrPath)
+	}
+
+	urlOrPath = strings.TrimPrefix(urlOrPath, "file://")
+	f, err := os.Open(urlOrPath)
+	r := NewReader(f)
+	r.URL = urlOrPath
+	if err != nil {
+		r.err = err
+	}
+
+	return r
+}
+
 func NewReaderFromURL(url string) *Reader {
 	docid, gid := ParseIDs(url)
 
@@ -85,9 +102,9 @@ func GetSheet(docid, gid string) (*http.Response, error) {
 func NewReader(r io.Reader) *Reader {
 	csvr := csv.NewReader(r)
 	csvr.Comma = ','
-	csvr.Comment = 0         // to allow distinguishing between quoted comments and fields
-	csvr.FieldsPerRecord = 0 // Google Sheets is consistent
-	csvr.LazyQuotes = false  // fields that need quotes use them correctly
+	csvr.Comment = 0          // to allow distinguishing between quoted comments and fields
+	csvr.FieldsPerRecord = -1 // Google Sheets is consistent, but our commented files are not
+	csvr.LazyQuotes = false   // fields that need quotes use them correctly
 	csvr.TrimLeadingSpace = false
 	csvr.ReuseRecord = false
 	return &Reader{
@@ -186,8 +203,9 @@ func ParseIDs(urlStr string) (docid string, gid string) {
 
 type Writer struct {
 	*csv.Writer
-	Comment rune
-	w       io.Writer
+	Comment       rune
+	QuoteComments bool
+	w             io.Writer
 }
 
 func NewWriter(w io.Writer) *Writer {
@@ -201,25 +219,31 @@ func NewWriter(w io.Writer) *Writer {
 func (w *Writer) Write(record []string) error {
 	if len(record) > 1 {
 		if rv1, _ := utf8.DecodeRuneInString(record[0]); rv1 == w.Comment {
-			w.Flush()
-
 			lastNonEmpty := len(record) - 1
-			for len(record[lastNonEmpty]) == 0 {
-				lastNonEmpty -= 1
+			if lastNonEmpty > -1 {
+				for len(record[lastNonEmpty]) == 0 {
+					lastNonEmpty -= 1
+				}
 			}
 
+			first := 0
 			if lastNonEmpty == 0 {
 				record = record[:1]
-			} else {
-				for i, f := range record {
-					if i == 0 || strings.Contains(f, `"`) {
-						f = strings.ReplaceAll(f, `"`, `""`)
-						record[i] = `"` + f + `"`
-					}
+				// if !w.QuoteComments {
+				// 	return nil
+				// }
+				first = -1
+			}
+
+			for i, f := range record {
+				if i == first || strings.Contains(f, `"`) || strings.Contains(f, string(w.Comma)) {
+					f = strings.ReplaceAll(f, `"`, `""`)
+					record[i] = `"` + f + `"`
 				}
 			}
 
 			line := strings.Join(record, string(w.Comma))
+			w.Flush()
 			if _, err := w.w.Write([]byte(line + "\n")); err != nil {
 				return err
 			}
@@ -230,6 +254,19 @@ func (w *Writer) Write(record []string) error {
 	return w.Writer.Write(record)
 }
 
+func QuoteRecord(record []string, comma, comment string) {
+	const quote = `"`
+
+	for i, f := range record {
+		if (i == 0 && strings.HasPrefix(f, comment)) ||
+			(strings.Contains(f, quote) || strings.Contains(f, string(comma))) {
+
+			f = strings.ReplaceAll(f, `"`, `""`)
+			record[i] = `"` + f + `"`
+		}
+	}
+}
+
 func (w *Writer) WriteAll(records [][]string) error {
 	for _, r := range records {
 		if err := w.Write(r); err != nil {