Initial implementation of `taglib` MetadataExtractor

This commit is contained in:
Deluan 2020-09-04 16:59:45 -04:00 committed by Deluan Quintão
parent b6aa6eb7b2
commit df4328819d
6 changed files with 227 additions and 129 deletions

3
go.mod
View File

@ -23,6 +23,9 @@ require (
github.com/mattn/go-sqlite3 v2.0.3+incompatible
github.com/microcosm-cc/bluemonday v1.0.4
github.com/mitchellh/mapstructure v1.3.2 // indirect
github.com/nicksellen/audiotags v0.0.0-20160226222119-94015fa599bd
github.com/onsi/ginkgo v1.14.0
github.com/onsi/gomega v1.10.1
github.com/onsi/ginkgo v1.14.1
github.com/onsi/gomega v1.10.2
github.com/pelletier/go-toml v1.8.0 // indirect

2
go.sum
View File

@ -253,6 +253,8 @@ github.com/modern-go/concurrent v0.0.0-20180306012644-bacd9c7ef1dd/go.mod h1:6dJ
github.com/modern-go/reflect2 v0.0.0-20180701023420-4b7aa43c6742/go.mod h1:bx2lNnkwVCuqBIxFjflWJWanXIb3RllmbCylyMrvgv0=
github.com/modern-go/reflect2 v1.0.1/go.mod h1:bx2lNnkwVCuqBIxFjflWJWanXIb3RllmbCylyMrvgv0=
github.com/mwitkow/go-conntrack v0.0.0-20161129095857-cc309e4a2223/go.mod h1:qRWi+5nqEBWmkhHvq77mSJWrCKwh8bxhgT7d/eI7P4U=
github.com/nicksellen/audiotags v0.0.0-20160226222119-94015fa599bd h1:xKn/gU8lZupoZt/HE7a/R3aH93iUO6JwyRsYelQUsRI=
github.com/nicksellen/audiotags v0.0.0-20160226222119-94015fa599bd/go.mod h1:B6icauz2l4tkYQxmDtCH4qmNWz/evSW5CsOqp6IE5IE=
github.com/niemeyer/pretty v0.0.0-20200227124842-a10e7caefd8e h1:fD57ERR4JtEqsWbfPhv4DMiApHyliiK5xCTNVSPiaAs=
github.com/niemeyer/pretty v0.0.0-20200227124842-a10e7caefd8e/go.mod h1:zD1mROLANZcx1PVRCS0qkT7pwLkGfwJo4zjcN/Tysno=
github.com/nxadm/tail v1.4.4 h1:DQuhQpB1tVlglWS2hLQ5OV6B5r8aGxSrPc5Qo6uTN78=

View File

@ -3,59 +3,29 @@ package metadata
import (
"bufio"
"errors"
"fmt"
"os"
"os/exec"
"path"
"regexp"
"strconv"
"strings"
"time"
"github.com/deluan/navidrome/conf"
"github.com/deluan/navidrome/log"
)
type ffmpegMetadata struct {
filePath string
suffix string
fileInfo os.FileInfo
tags map[string]string
baseMetadata
}
func (m *ffmpegMetadata) Title() string { return m.getTag("title", "sort_name") }
func (m *ffmpegMetadata) Album() string { return m.getTag("album", "sort_album") }
func (m *ffmpegMetadata) Artist() string { return m.getTag("artist", "sort_artist") }
func (m *ffmpegMetadata) AlbumArtist() string { return m.getTag("album_artist", "albumartist") }
func (m *ffmpegMetadata) SortTitle() string { return m.getSortTag("", "title", "name") }
func (m *ffmpegMetadata) SortAlbum() string { return m.getSortTag("", "album") }
func (m *ffmpegMetadata) SortArtist() string { return m.getSortTag("", "artist") }
func (m *ffmpegMetadata) SortAlbumArtist() string {
return m.getSortTag("tso2", "albumartist", "album_artist")
}
func (m *ffmpegMetadata) Composer() string { return m.getTag("composer", "tcm", "sort_composer") }
func (m *ffmpegMetadata) Genre() string { return m.getTag("genre") }
func (m *ffmpegMetadata) Year() int { return m.parseYear("date") }
func (m *ffmpegMetadata) TrackNumber() (int, int) { return m.parseTuple("track") }
func (m *ffmpegMetadata) DiscNumber() (int, int) { return m.parseTuple("tpa", "disc") }
func (m *ffmpegMetadata) DiscSubtitle() string {
return m.getTag("tsst", "discsubtitle", "setsubtitle")
}
func (m *ffmpegMetadata) Duration() float32 { return m.parseDuration("duration") }
func (m *ffmpegMetadata) BitRate() int { return m.parseInt("bitrate") }
func (m *ffmpegMetadata) HasPicture() bool {
return m.getTag("has_picture", "metadata_block_picture") != ""
}
func (m *ffmpegMetadata) Comment() string { return m.getTag("comment") }
func (m *ffmpegMetadata) Compilation() bool { return m.parseBool("compilation") }
func (m *ffmpegMetadata) Duration() float32 { return m.parseDuration("duration") }
func (m *ffmpegMetadata) BitRate() int { return m.parseInt("bitrate") }
func (m *ffmpegMetadata) ModificationTime() time.Time { return m.fileInfo.ModTime() }
func (m *ffmpegMetadata) FilePath() string { return m.filePath }
func (m *ffmpegMetadata) Suffix() string { return m.suffix }
func (m *ffmpegMetadata) Size() int64 { return m.fileInfo.Size() }
func (m *ffmpegMetadata) DiscNumber() (int, int) { return m.parseTuple("tpa", "disc") }
type ffmpegMetadataExtractor struct{}
type ffmpegExtractor struct{}
func (e *ffmpegMetadataExtractor) Extract(files ...string) (map[string]Metadata, error) {
func (e *ffmpegExtractor) Extract(files ...string) (map[string]Metadata, error) {
args := createProbeCommand(files)
log.Trace("Executing command", "args", args)
@ -116,8 +86,9 @@ func parseOutput(output string) map[string]string {
}
func extractMetadata(filePath, info string) (*ffmpegMetadata, error) {
m := &ffmpegMetadata{filePath: filePath, tags: map[string]string{}}
m.suffix = strings.ToLower(strings.TrimPrefix(path.Ext(filePath), "."))
m := &ffmpegMetadata{}
m.filePath = filePath
m.tags = map[string]string{}
var err error
m.fileInfo, err = os.Stat(filePath)
if err != nil {
@ -175,88 +146,6 @@ func (m *ffmpegMetadata) parseInfo(info string) {
}
}
func (m *ffmpegMetadata) parseInt(tagName string) int {
if v, ok := m.tags[tagName]; ok {
i, _ := strconv.Atoi(v)
return i
}
return 0
}
var dateRegex = regexp.MustCompile(`^([12]\d\d\d)`)
func (m *ffmpegMetadata) parseYear(tagName string) int {
if v, ok := m.tags[tagName]; ok {
match := dateRegex.FindStringSubmatch(v)
if len(match) == 0 {
log.Warn("Error parsing year from ffmpeg date field", "file", m.filePath, "date", v)
return 0
}
year, _ := strconv.Atoi(match[1])
return year
}
return 0
}
func (m *ffmpegMetadata) getTag(tags ...string) string {
for _, t := range tags {
if v, ok := m.tags[t]; ok {
return v
}
}
return ""
}
func (m *ffmpegMetadata) getSortTag(originalTag string, tags ...string) string {
formats := []string{"sort%s", "sort_%s", "sort-%s", "%ssort", "%s_sort", "%s-sort"}
all := []string{originalTag}
for _, tag := range tags {
for _, format := range formats {
name := fmt.Sprintf(format, tag)
all = append(all, name)
}
}
return m.getTag(all...)
}
func (m *ffmpegMetadata) parseTuple(tags ...string) (int, int) {
for _, tagName := range tags {
if v, ok := m.tags[tagName]; ok {
tuple := strings.Split(v, "/")
t1, t2 := 0, 0
t1, _ = strconv.Atoi(tuple[0])
if len(tuple) > 1 {
t2, _ = strconv.Atoi(tuple[1])
} else {
t2, _ = strconv.Atoi(m.tags[tagName+"total"])
}
return t1, t2
}
}
return 0, 0
}
func (m *ffmpegMetadata) parseBool(tagName string) bool {
if v, ok := m.tags[tagName]; ok {
i, _ := strconv.Atoi(strings.TrimSpace(v))
return i == 1
}
return false
}
var zeroTime = time.Date(0000, time.January, 1, 0, 0, 0, 0, time.UTC)
func (m *ffmpegMetadata) parseDuration(tagName string) float32 {
if v, ok := m.tags[tagName]; ok {
d, err := time.Parse("15:04:05", v)
if err != nil {
return 0
}
return float32(d.Sub(zeroTime).Seconds())
}
return 0
}
// Inputs will always be absolute paths
func createProbeCommand(inputs []string) []string {
split := strings.Split(conf.Server.ProbeCommand, " ")

View File

@ -9,7 +9,7 @@ var _ = Describe("ffmpegMetadata", func() {
// TODO Need to mock `ffmpeg`
XContext("ExtractAllMetadata", func() {
It("correctly parses metadata from all files in folder", func() {
e := &ffmpegMetadataExtractor{}
e := &ffmpegExtractor{}
mds, err := e.Extract("tests/fixtures/test.mp3", "tests/fixtures/test.ogg")
Expect(err).NotTo(HaveOccurred())
Expect(mds).To(HaveLen(2))
@ -224,13 +224,15 @@ Input #0, mp3, from '/Users/deluan/Downloads/椎名林檎 - 加爾基 精液 栗
"May 12, 2016": 0,
}
for tag, expected := range examples {
md := &ffmpegMetadata{tags: map[string]string{"date": tag}}
md := &ffmpegMetadata{}
md.tags = map[string]string{"date": tag}
Expect(md.Year()).To(Equal(expected))
}
})
It("returns 0 if year is invalid", func() {
md := &ffmpegMetadata{tags: map[string]string{"date": "invalid"}}
md := &ffmpegMetadata{}
md.tags = map[string]string{"date": "invalid"}
Expect(md.Year()).To(Equal(0))
})
})

View File

@ -1,6 +1,25 @@
package metadata
import "time"
import (
"fmt"
"os"
"path"
"regexp"
"strconv"
"strings"
"time"
"github.com/deluan/navidrome/log"
)
type Extractor interface {
Extract(files ...string) (map[string]Metadata, error)
}
func Extract(files ...string) (map[string]Metadata, error) {
e := &taglibExtractor{}
return e.Extract(files...)
}
type Metadata interface {
Title() string
@ -28,11 +47,122 @@ type Metadata interface {
Size() int64
}
type Extractor interface {
Extract(files ...string) (map[string]Metadata, error)
type baseMetadata struct {
filePath string
fileInfo os.FileInfo
tags map[string]string
}
func Extract(files ...string) (map[string]Metadata, error) {
e := &ffmpegMetadataExtractor{}
return e.Extract(files...)
func (m *baseMetadata) Title() string { return m.getTag("title", "sort_name") }
func (m *baseMetadata) Album() string { return m.getTag("album", "sort_album") }
func (m *baseMetadata) Artist() string { return m.getTag("artist", "sort_artist") }
func (m *baseMetadata) AlbumArtist() string { return m.getTag("album_artist", "albumartist") }
func (m *baseMetadata) SortTitle() string { return m.getSortTag("", "title", "name") }
func (m *baseMetadata) SortAlbum() string { return m.getSortTag("", "album") }
func (m *baseMetadata) SortArtist() string { return m.getSortTag("", "artist") }
func (m *baseMetadata) SortAlbumArtist() string {
return m.getSortTag("tso2", "albumartist", "album_artist")
}
func (m *baseMetadata) Composer() string { return m.getTag("composer", "tcm", "sort_composer") }
func (m *baseMetadata) Genre() string { return m.getTag("genre") }
func (m *baseMetadata) Year() int { return m.parseYear("date") }
func (m *baseMetadata) Comment() string { return m.getTag("comment") }
func (m *baseMetadata) Compilation() bool { return m.parseBool("compilation") }
func (m *baseMetadata) TrackNumber() (int, int) { return m.parseTuple("track", "tracknumber") }
func (m *baseMetadata) DiscNumber() (int, int) { return m.parseTuple("disc", "discnumber") }
func (m *baseMetadata) DiscSubtitle() string {
return m.getTag("tsst", "discsubtitle", "setsubtitle")
}
func (m *baseMetadata) ModificationTime() time.Time { return m.fileInfo.ModTime() }
func (m *baseMetadata) Size() int64 { return m.fileInfo.Size() }
func (m *baseMetadata) FilePath() string { return m.filePath }
func (m *baseMetadata) Suffix() string {
return strings.ToLower(strings.TrimPrefix(path.Ext(m.FilePath()), "."))
}
func (m *baseMetadata) Duration() float32 { panic("not implemented") }
func (m *baseMetadata) BitRate() int { panic("not implemented") }
func (m *baseMetadata) HasPicture() bool { panic("not implemented") }
func (m *baseMetadata) parseInt(tagName string) int {
if v, ok := m.tags[tagName]; ok {
i, _ := strconv.Atoi(v)
return i
}
return 0
}
var dateRegex = regexp.MustCompile(`^([12]\d\d\d)`)
func (m *baseMetadata) parseYear(tagName string) int {
if v, ok := m.tags[tagName]; ok {
match := dateRegex.FindStringSubmatch(v)
if len(match) == 0 {
log.Warn("Error parsing year from ffmpeg date field", "file", m.filePath, "date", v)
return 0
}
year, _ := strconv.Atoi(match[1])
return year
}
return 0
}
func (m *baseMetadata) getTag(tags ...string) string {
for _, t := range tags {
if v, ok := m.tags[t]; ok {
return v
}
}
return ""
}
func (m *baseMetadata) getSortTag(originalTag string, tags ...string) string {
formats := []string{"sort%s", "sort_%s", "sort-%s", "%ssort", "%s_sort", "%s-sort"}
all := []string{originalTag}
for _, tag := range tags {
for _, format := range formats {
name := fmt.Sprintf(format, tag)
all = append(all, name)
}
}
return m.getTag(all...)
}
func (m *baseMetadata) parseTuple(tags ...string) (int, int) {
for _, tagName := range tags {
if v, ok := m.tags[tagName]; ok {
tuple := strings.Split(v, "/")
t1, t2 := 0, 0
t1, _ = strconv.Atoi(tuple[0])
if len(tuple) > 1 {
t2, _ = strconv.Atoi(tuple[1])
} else {
t2, _ = strconv.Atoi(m.tags[tagName+"total"])
}
return t1, t2
}
}
return 0, 0
}
func (m *baseMetadata) parseBool(tagName string) bool {
if v, ok := m.tags[tagName]; ok {
i, _ := strconv.Atoi(strings.TrimSpace(v))
return i == 1
}
return false
}
var zeroTime = time.Date(0000, time.January, 1, 0, 0, 0, 0, time.UTC)
func (m *baseMetadata) parseDuration(tagName string) float32 {
if v, ok := m.tags[tagName]; ok {
d, err := time.Parse("15:04:05", v)
if err != nil {
return 0
}
return float32(d.Sub(zeroTime).Seconds())
}
return 0
}

View File

@ -0,0 +1,72 @@
package metadata
import (
"errors"
"os"
"github.com/deluan/navidrome/log"
"github.com/dhowden/tag"
"github.com/nicksellen/audiotags"
)
type taglibMetadata struct {
baseMetadata
props *audiotags.AudioProperties
hasPicture bool
}
func (m *taglibMetadata) Duration() float32 { return float32(m.props.Length) }
func (m *taglibMetadata) BitRate() int { return m.props.Bitrate }
func (m *taglibMetadata) HasPicture() bool { return m.hasPicture }
type taglibExtractor struct{}
func (e *taglibExtractor) Extract(paths ...string) (map[string]Metadata, error) {
mds := map[string]Metadata{}
var err error
for _, path := range paths {
md, err := e.extractMetadata(path)
if err == nil {
mds[path] = md
}
}
return mds, err
}
func (e *taglibExtractor) extractMetadata(filePath string) (*taglibMetadata, error) {
var err error
md := &taglibMetadata{}
md.filePath = filePath
md.fileInfo, err = os.Stat(filePath)
if err != nil {
log.Warn("Error stating file. Skipping", "filePath", filePath, err)
return nil, errors.New("error stating file")
}
md.tags, md.props, err = audiotags.Read(filePath)
if err != nil {
log.Warn("Error reading metadata from file. Skipping", "filePath", filePath, err)
return nil, errors.New("error reading tags")
}
md.hasPicture = hasEmbeddedImage(filePath)
return md, nil
}
func hasEmbeddedImage(path string) bool {
f, err := os.Open(path)
if err != nil {
log.Warn("Error opening file", "filePath", path, err)
return false
}
defer f.Close()
m, err := tag.ReadFrom(f)
if err != nil {
log.Warn("Error reading tags from file", "filePath", path, err)
return false
}
return m.Picture() != nil
}
var _ Metadata = (*taglibMetadata)(nil)
var _ Extractor = (*taglibExtractor)(nil)