executor/inf2ffmpeg/inf2ffmpeg-deprecated.go

//go:build id3v2

// Read icedax .INF files for each track and change the ID3 tags accordingly.
//
// Licensed under a MIT license (https://gwyneth-llewelyn.mit-license.org/)

package main

import (
	//	"bytes"
	"bufio"
	"fmt"

	//	"io"
	"log"
	"os"
	"os/exec"
	"path/filepath"
	"strconv"
	"strings"
	"time"

	"github.com/bogem/id3v2/v2"
	"github.com/karrick/godirwalk"
	"github.com/karrick/golf" // flag replacement library
)

/*
#.INF example file

#created by icedax 1.1.11 07/01/23 10:07:54
#
CDINDEX_DISCID= '8_8DwRM2rLmumKtqXp59zmg4jkk-'
CDDB_DISCID=    0x450fae05
MCN=
ISRC=
#
Albumperformer= 'CD005-Berliner Philharmoniker, Herbert von Karajan'
Performer=      'CD005-Berliner Philharmoniker, Herbert von Karajan'
Albumtitle=     'Beethoven: Symphony No.9'
Tracktitle=     'Symphony No.9 in d op.125 - Presto - Allegro assai -'
Tracknumber=    5
Trackstart=     220380
# track length in sectors (1/75 seconds each), rest samples
Tracklength=    80702, 0
Pre-emphasis=   no
Channels=       2
Copy_permitted= once (copyright protected)
Endianess=      little
# index list
Index=          0
Index0=         -1

// Tags on audio.cddb
"DISCID":			"",
"DTITLE":			"TALB",	// may have multiple lines!!
"DYEAR":			"TDRC",
"DGENRE":			"TCON",

// on each .inf file (one per track)
"CDINDEX_DISCID":	"",
"CDDB_DISCID":		"",
"MCN":				"",
"ISRC":				"TSRC",
"Albumperformer":	"TPE2",
"Performer":		"TPE3",
"Albumtitle":		"TALB",
"Tracktitle":		"TIT2",
"Tracknumber":		"TRCK",
"Trackstart": 		"",		// probably sectors
"Tracklength": 		"TLEN",	// sectors [1 sector = 1/75 sec], rest
"Pre-emphasis":		"",		// yes/no
"Channels": 		"",		// 1: mono; 2: stereo
"Copy_permitted":	"",		// once (copyright protected)
"Endianess":		"",		// little/big
"Index": 			"",		// 0
"Index0":			"",		// -1 */

// instantiate a few global variables, to avoid memory leaks.

var (
	dirPath string		// path where the files are
	extension string	// file extension (including dot); almost always ".inf"
	recursive bool		// if set, walks through directories recursively
	noConvert bool		// if set, will NOT launch ffmpeg
	dryRun bool			// if set, doesn't save any tags (and enter debug mode)
	justRead bool		// if set, just read tags from a single file (for debugging purposes)
	help bool			// if set, show usage
	ffmpegPath string	// path to ffmpeg executable
	genre string		// album genre; must be provided via command line, since .inf doesn't have it.
	picture string		// path to the album image, to set the .m4a to it
	pic id3v2.PictureFrame	// an encoded picture for inserting into a .m4a as album cover.
)

func main() {
	// read flags, if any
	golf.StringVarP(&dirPath,		'd', "dir", ".", "directory to convert from")
	golf.BoolVarP(&recursive,		'r', "recursive", false, "traverse directory structure recursively")
	golf.StringVarP(&extension,		'e', "extension", ".inf", "icedax track info file extension")
	golf.StringVarP(&ffmpegPath,	'm', "ffmpeg", "/usr/local/bin/ffmpeg", "path to ffmpeg executable")
	golf.BoolVarP(&noConvert,		'n', "no-conversion", false, "If true, does not convert from .wav to .m4a with ffmpeg")
	golf.StringVarP(&genre,			'g', "genre", "", "override album genre (sometimes called 'content type')")
	golf.StringVarP(&picture,		'p', "picture", "", "path to album picture")
	golf.BoolVarP(&dryRun,			'x', "dry-run", false, "if true, does not save any tags to file (and enters debug mode)")
	golf.BoolVarP(&justRead,		'j', "just-read", false, "just read a .m4a file for debugging purposes")
	golf.BoolVarP(&help,			'h', "help", false, "show command usage")
	golf.Parse()

	if help {
		golf.Usage()
		os.Exit(0)
	}
	if justRead {
		os.Exit(justReadM4A(golf.Args()[0]))
	}
//	fmt.Printf("[DEBUG] Skip conversion mode set to %v; Dry run set to %v; Recursive set to %v\n", noConvert, dryRun, recursive)

	if !noConvert {
		// check of we can run ffmpeg
		path, err := exec.LookPath(ffmpegPath)	// also search under the $PATH shell
		if err != nil {
			log.Panicf("couldn't find executable `ffmpeg` at %q, error was: %s", ffmpegPath, err)
		}
		fmt.Printf("`ffmpeg` is available at %s\n", path)
	}
	fmt.Printf("Now entering directory %q (recursive: %t)...\n\n", dirPath, recursive)

	// prepare image, if we got one:
	if picture != "" {
		artwork, err := os.ReadFile(picture)
		if err != nil {
			// abort if we cannot read the file, the user
			log.Fatalf("Error while reading artwork file %q: %s\nTry a different picture or check that the path is correct.\n", picture, err)
		}
		// see id3v2 documentation
		pic = id3v2.PictureFrame{
			Encoding:    id3v2.EncodingUTF8,
			MimeType:    "image/jpeg",
			PictureType: id3v2.PTFrontCover,
			Description: "Front cover",
			Picture:     artwork,
		}
	}

	// tags we managed to extract successfully, either via audio.cddb or the .INF files.
	tags := make(map[string]string)

	// Before transversing this directory, we will attempt to read the audio.cddb file.
	// If we cannot find it, we'll have to assume sort-of-reasonable defaults...
	audioCddb, ferr := os.ReadFile(filepath.Join(dirPath, "audio.cddb"))
	if ferr != nil {
		fmt.Println("[WARN] audio.cddb file not found; proceeding with defaults. Error was: ", ferr)
	} else {
		var line, key, value string // one input line
		sc := bufio.NewScanner(strings.NewReader(string(audioCddb)))
		for sc.Scan() {
			line = sc.Text()
			// Drop comments
			if strings.HasPrefix(line, "#") {
				continue	// skip comments
			}

			// key = value (see also below)
			// Value can be an empty string, of course!
			keyValue := strings.Split(line, "=")

			// missing "="?
			if len(keyValue) < 2 {
				fmt.Println("malformed line...")
				continue
			}
			key   = strings.ToUpper(strings.Trim(keyValue[0], " \t\"'"))
			value = strings.Trim(keyValue[1], " \t\"'")

			// extract the values for each key, and set it on tags
			switch key {
				case "DISCID":
					tags["DISCID"] = value
				case "DTITLE":
					tags["DTITLE"] += value		// note: multiple lines are possible!
				case "DYEAR":
					tags["DYEAR"] = value
				case "DGENRE":
					tags["DGENRE"] = value
			} // end switch key
		} // end for sc.Scan()
	} // end read file audio.cbbd

	// similar to executor code: transverse all files, check if we have .inf
	// then parse it, and eventually convert the files, too.
	// Note: this is done serially for now, but it *should+ become goroutines! (gwyneth 20230709)
	err := godirwalk.Walk(dirPath,
		&godirwalk.Options{
			Callback: func(osPathname string, de *godirwalk.Dirent) error {
				// skip directories/symlinks to directories (if not in recursive mode)
				isDir, dirErr := de.IsDirOrSymlinkToDir();
				if isDir && !recursive {
					if dirErr == nil {
						// return godirwalk.SkipThis
						return nil
					}
					fmt.Printf("error when trying to access directory/symlink %q: %s",
						osPathname, dirErr)
						return nil
				}

				// only read files ending in .inf
				// note: I was using HasSuffix here, but apparently filepath.Ext() is way faster! (gwyneth 20230706)
				if filepath.Ext(osPathname) != extension {
					return nil
				}

				// open INF file for reading, and scan it line by line into
				// an array of strings
				rawInfFile, err := os.Open(osPathname)
				if err != nil {
					log.Fatal(err)
				}

				fileScanner := bufio.NewScanner(rawInfFile)
				fileScanner.Split(bufio.ScanLines)
				var (
					fileLines []string	// array of strings
					infFileLength int	// length of array
				)
				for fileScanner.Scan() {
					fileLines = append(fileLines, fileScanner.Text())
					infFileLength++
				}

				rawInfFile.Close()

				var key, value string
				// now go through the array
				for i := 0; i < infFileLength; i++ {
					if strings.HasPrefix(fileLines[i], "#") {
						continue	// skip comments
					}

					// .INF files have each line with key=value, with lots of space in-between (or not!)
					// Value often has quotes around but we'll ignore it for now (gwyneth 20230706)
					keyValue := strings.Split(fileLines[i], "=")

					// missing "="?
					if len(keyValue) < 2 {
						fmt.Printf("malformed line %d\n", i)
						continue
					}
					key   = strings.Trim(keyValue[0], " \t\"'")
					value = strings.Trim(keyValue[1], " \t\"'")
					tags[key] = value
				} // end for i loop

				fmt.Printf("[DEBUG] Matched tags: %v\n", tags)

				fullFileNameBase := osPathname[:len(osPathname) - len(filepath.Ext(osPathname))]
				fullDir := filepath.Dir(osPathname)
				fileNameM4A	:= fullFileNameBase + ".m4a"
				var track int	// track number, which will be used elsewhere
				// Deal with conversion
				if !noConvert {
					// icedax generates .wav and .inf with the same basename:
					fileNameWAV := fullFileNameBase + ".wav"

					// Do we know how the track is called? (from .inf file)
					// This should almost always be the case, unless the original CD
					// was completely borked and icedax couldn't find the track names at all.
					if tags["Tracktitle"] != "" && tags["Tracknumber"] != "" {
						track, err = strconv.Atoi(tags["Tracknumber"])
						if err != nil {
							fmt.Printf("invalid track number: %q, conversion error: %s", track, err)
						}
						fileNameM4A = filepath.Join(fullDir, fmt.Sprintf("%02d %s.m4a", track, tags["Tracktitle"]))
					}

					cmd := exec.Command(ffmpegPath, "-i", fileNameWAV, fileNameM4A)
					stdoutStderr, err := cmd.CombinedOutput()
					if err != nil {
						fmt.Printf("error while running %q: %q \n", ffmpegPath, err)
						return nil
					}
					fmt.Printf("✅ %s\n", stdoutStderr)
				}
				// open .m4a file
				m4aFile, err := id3v2.Open(fileNameM4A, id3v2.Options{Parse: true})
				if err != nil {
					if !dryRun {
						fmt.Println("[ERROR] ", err)	// err includes filename
					} else {
						fmt.Println("[WARN] Couldn't open .m4a file (not generated?); cannot proceed with tag processing.")
					}
					return nil
				}
				defer m4aFile.Close()

				// See what ffmpeg already wrote
				fmt.Printf("[INFO] %d tag(s) found in existing .m4a file, taking %d bytes.\n",
					m4aFile.Count(), m4aFile.Size())

				m4aFile.SetDefaultEncoding(id3v2.EncodingUTF8)
				m4aFile.SetArtist(tags["Performer"])	// TPE3
				m4aFile.SetTitle(tags["Tracktitle"])	// TIT2
				// are we overriding the genre from the command line?
				if genre == "" {
					// no? then use whatever was on audio.cddb; could also be empty, of course.
					genre = tags["DGENRE"]
				}
				m4aFile.SetGenre(genre)
				m4aFile.SetVersion(4)	// ID3v2 version; v4 is the latest
				m4aFile.SetYear(tags["DYEAR"])

				// Set comment frame.
				comment := id3v2.CommentFrame{
					Encoding:		id3v2.EncodingUTF8,
					Language:		"eng",
					Description:	"Additional song IDs",
					Text:			fmt.Sprintf("DISCID=%q\tCDINDEX_DISCID=%q\tCDDB_DISCID=%q\tMCN=%q",
										tags["DISCID"],
										tags["CDINDEX_DISCID"],
										tags["CDDB_DISCID"],
										tags["MCN"],
									),
				}
				m4aFile.AddCommentFrame(comment)

				// if we have a picture, read it and attach it to .m4a file
				if picture != "" {
					m4aFile.AddAttachedPicture(pic)
				}

				// other tags from .INF file
				switch true {
					case tags["Albumtitle"] != "":
						m4aFile.AddTextFrame(m4aFile.CommonID("Album/Movie/Show title"),
							id3v2.EncodingUTF8, tags["Albumtitle"])
					case tags["Albumperformer"] != "":
							m4aFile.AddTextFrame(m4aFile.CommonID("Band/Orchestra/Accompaniment"),
								id3v2.EncodingUTF8, tags["Albumperformer"])
					case tags["ISRC"] != "":
						m4aFile.AddTextFrame(m4aFile.CommonID("ISRC"), id3v2.EncodingUTF8, tags["ISRC"])
					// other tags we have from the .inf file
					case tags["Tracknumber"] != "":
						m4aFile.AddTextFrame(m4aFile.CommonID("Track number/Position in set"),
							id3v2.EncodingUTF8, tags["Tracknumber"])
					case tags["Tracklength"] != "":
						var tlen, ignore int	// Track length, in sectors
						_, err := fmt.Sscanf(tags["Tracklength"], "%d, %d", tlen, ignore)
						if err != nil {
							fmt.Printf("[WARN] Track length not found! Error was: %s\n", err)
						} else {
							m4aFile.AddTextFrame(m4aFile.CommonID("Length"),
								id3v2.EncodingUTF8,
								strconv.FormatFloat(float64(tlen) / 75, 'f', 2, 64))
						}
				}

				m4aFile.AddTextFrame(m4aFile.CommonID("Software/Hardware and settings used for encoding"),
					id3v2.EncodingUTF8,
					"ffmpeg version 6.0")

				// basically an experiment. TOFN is the original file name, in case someone borks the
				// filename and needs to get it back somehow
				m4aFile.AddTextFrame("TOFN", id3v2.EncodingUTF8,
					fmt.Sprintf("%02d %s.m4a", track, tags["Tracktitle"]))

				// capture a timestamp to embed as the date the tags were generated
				m4aFile.AddTextFrame(m4aFile.CommonID("Tagging time"),
					id3v2.EncodingUTF8,
					time.Now().UTC().Format(time.RFC3339))

				fmt.Printf("[INFO] After reading .inf file, we now have %d tag(s), taking %d bytes.\n",
					m4aFile.Count(), m4aFile.Size())
				fmt.Printf("[DEBUG] All the frames added by inf2ffmpeg: %#v\n",
					m4aFile.AllFrames())

				// Write everything to file, unless --dry-run is set
				if !dryRun {
					m4err := m4aFile.Save()
					if m4err != nil {
						log.Printf("[ERROR] Couldn't save tags to %q: %s\n", fileNameM4A, m4err)
					} else {
						fmt.Println("✅ All tags successfully saved to ", fileNameM4A)
					}
				}

				return nil
			},	// ends Callback
			Unsorted: false, // (optional) set true for faster yet non-deterministic enumeration (see godoc)
		})	// end options for dirwalk
		if err != nil {
			log.Printf("sorry, got error %s\n", err)
		}
}

// justReadM4A is a debugging tool to read a .m4a and show what tags it contains.
func justReadM4A(fileNameM4A string) int {
	m4aFile, err := id3v2.Open(fileNameM4A, id3v2.Options{Parse: true})
	if err != nil {
		log.Println(err)
		os.Exit(66)	// proper exit status code for file not found
	}
	defer m4aFile.Close()

	log.Printf("[INFO] %d tag(s) found in %q, taking %d bytes.\n",
		m4aFile.Count(), fileNameM4A, m4aFile.Size())

	fmt.Println("Title:",	m4aFile.Title())
	fmt.Println("Artist:",	m4aFile.Artist())
	fmt.Println("Year:",	m4aFile.Year())
	fmt.Println("Genre:",	m4aFile.Genre())
	fmt.Println("ID3 version:",	m4aFile.Version())

	fmt.Println("Other frames found:")

	allTags := m4aFile.AllFrames()
	for i, aFrame := range allTags {
		fmt.Printf("%s: %#v\nTags for this frame:\n", i, aFrame)
		for j, aTag := range aFrame {
			fmt.Printf("\t%s: %02d - %v\n", i, j, aTag.UniqueIdentifier())
		}
	}
	log.Printf("[INFO] Finished parsing %q.\n",
		fileNameM4A)

	return 0	// 1 is the status error for 'generic' failure
}