Switch EXIF library

Closes gohugoio#10855
Closes gohugoio#8586

go.mod

@@ -15,6 +15,7 @@ require (
 	github.com/bep/golibsass v1.1.1
 	github.com/bep/gowebp v0.3.0
 	github.com/bep/helpers v0.4.0
+	github.com/bep/imagemeta v0.0.0-20240707104023-e2e8168238fe
 	github.com/bep/lazycache v0.4.0
 	github.com/bep/logg v0.4.0
 	github.com/bep/mclib v1.20400.20402
@@ -60,7 +61,6 @@ require (
 	github.com/pbnjay/memory v0.0.0-20210728143218-7b4eea64cf58
 	github.com/pelletier/go-toml/v2 v2.2.2
 	github.com/rogpeppe/go-internal v1.12.0
-	github.com/rwcarlsen/goexif v0.0.0-20190401172101-9e8deecbddbd
 	github.com/sanity-io/litter v1.5.5
 	github.com/spf13/afero v1.11.0
 	github.com/spf13/cast v1.6.0
@@ -162,3 +162,5 @@ require (
 )
 
 go 1.20
+
+replace github.com/bep/imagemeta => /Users/bep/dev/go/bep/imagemeta

go.sum

@@ -405,7 +405,6 @@ github.com/rogpeppe/go-internal v1.12.0/go.mod h1:E+RYuTGaKKdloAfM02xzb0FW3Paa99
 github.com/russross/blackfriday/v2 v2.1.0 h1:JIOH55/0cWyOuilr9/qlrm0BSXldqnqwMsf35Ld67mk=
 github.com/russross/blackfriday/v2 v2.1.0/go.mod h1:+Rmxgy9KzJVeS9/2gXHxylqXiyQDYRxCVz55jmeOWTM=
 github.com/rwcarlsen/goexif v0.0.0-20190401172101-9e8deecbddbd h1:CmH9+J6ZSsIjUK3dcGsnCnO41eRBOnY12zwkn5qVwgc=
-github.com/rwcarlsen/goexif v0.0.0-20190401172101-9e8deecbddbd/go.mod h1:hPqNNc0+uJM6H+SuU8sEs5K5IQeKccPqeSjfgcKGgPk=
 github.com/sanity-io/litter v1.5.5 h1:iE+sBxPBzoK6uaEP5Lt3fHNgpKcHXc/A2HGETy0uJQo=
 github.com/sanity-io/litter v1.5.5/go.mod h1:9gzJgR2i4ZpjZHsKvUXIRQVk7P+yM3e+jAF7bU2UI5U=
 github.com/shogo82148/go-shuffle v0.0.0-20180218125048-27e6095f230d/go.mod h1:2htx6lmL0NGLHlO8ZCf+lQBGBHIbEujyywxJArf+2Yc=

resources/image.go

@@ -82,8 +82,9 @@ func (i *imageResource) Exif() *exif.ExifInfo {
 
 func (i *imageResource) getExif() *exif.ExifInfo {
 	i.metaInit.Do(func() {
-		supportsExif := i.Format == images.JPEG || i.Format == images.TIFF
-		if !supportsExif {
+		mf := i.Format.ToImageMetaImageFormatFormat()
+		if mf == -1 {
+			// No Exif support for this format.
 			return
 		}
 
@@ -114,7 +115,7 @@ func (i *imageResource) getExif() *exif.ExifInfo {
 			}
 			defer f.Close()
 
-			x, err := i.getSpec().imaging.DecodeExif(f)
+			x, err := i.getSpec().imaging.DecodeExif(mf, f)
 			if err != nil {
 				i.getSpec().Logger.Warnf("Unable to decode Exif metadata from image: %s", i.Key())
 				return nil

resources/images/exif/exif.go

@@ -14,21 +14,14 @@
 package exif
 
 import (
-	"bytes"
 	"fmt"
 	"io"
-	"math"
-	"math/big"
 	"regexp"
 	"strings"
 	"time"
-	"unicode"
-	"unicode/utf8"
 
+	"github.com/bep/imagemeta"
 	"github.com/bep/tmc"
-
-	_exif "github.com/rwcarlsen/goexif/exif"
-	"github.com/rwcarlsen/goexif/tiff"
 )
 
 const exifTimeLayout = "2006:01:02 15:04:05"
@@ -115,143 +108,62 @@ func NewDecoder(options ...func(*Decoder) error) (*Decoder, error) {
 	return d, nil
 }
 
-func (d *Decoder) Decode(r io.Reader) (ex *ExifInfo, err error) {
+func (d *Decoder) Decode(format imagemeta.ImageFormat, r io.Reader) (ex *ExifInfo, err error) {
 	defer func() {
 		if r := recover(); r != nil {
 			err = fmt.Errorf("exif failed: %v", r)
 		}
 	}()
 
-	var x *_exif.Exif
-	x, err = _exif.Decode(r)
-	if err != nil {
-		if err.Error() == "EOF" {
-			// Found no Exif
-			return nil, nil
-		}
-		return
+	tagInfos := make(imagemeta.Tags)
+	handleTag := func(ti imagemeta.TagInfo) error {
+		tagInfos[ti.Tag] = ti
+		return nil
 	}
 
+	// var format imagemeta.ImageFormat
+
+	err = imagemeta.Decode(
+		imagemeta.Options{
+			R:           r.(imagemeta.Reader),
+			ImageFormat: format,
+			HandleTag:   handleTag,
+			Sources:     imagemeta.TagSourceEXIF | imagemeta.TagSourceXMP,
+		},
+	)
+
 	var tm time.Time
 	var lat, long float64
 
 	if !d.noDate {
-		tm, _ = x.DateTime()
+		tm, _ = tagInfos.GetDateTime()
 	}
 
 	if !d.noLatLong {
-		lat, long, _ = x.LatLong()
-		if math.IsNaN(lat) {
-			lat = 0
-		}
-		if math.IsNaN(long) {
-			long = 0
-		}
-	}
-
-	walker := &exifWalker{x: x, vals: make(map[string]any), includeMatcher: d.includeFieldsRe, excludeMatcher: d.excludeFieldsrRe}
-	if err = x.Walk(walker); err != nil {
-		return
+		lat, long, _ = tagInfos.GetLatLong()
 	}
 
-	ex = &ExifInfo{Lat: lat, Long: long, Date: tm, Tags: walker.vals}
-
-	return
-}
-
-func decodeTag(x *_exif.Exif, f _exif.FieldName, t *tiff.Tag) (any, error) {
-	switch t.Format() {
-	case tiff.StringVal, tiff.UndefVal:
-		s := nullString(t.Val)
-		if strings.Contains(string(f), "DateTime") {
-			if d, err := tryParseDate(x, s); err == nil {
-				return d, nil
-			}
+	tags := make(map[string]any)
+	for k, v := range tagInfos {
+		if d.excludeFieldsrRe != nil && d.excludeFieldsrRe.MatchString(k) {
+			continue
 		}
-		return s, nil
-	case tiff.OtherVal:
-		return "unknown", nil
-	}
-
-	var rv []any
-
-	for i := 0; i < int(t.Count); i++ {
-		switch t.Format() {
-		case tiff.RatVal:
-			n, d, _ := t.Rat2(i)
-			rat := big.NewRat(n, d)
-			// if t is int or t > 1, use float64
-			if rat.IsInt() || rat.Cmp(big.NewRat(1, 1)) == 1 {
-				f, _ := rat.Float64()
-				rv = append(rv, f)
-			} else {
-				rv = append(rv, rat)
-			}
-
-		case tiff.FloatVal:
-			v, _ := t.Float(i)
-			rv = append(rv, v)
-		case tiff.IntVal:
-			v, _ := t.Int(i)
-			rv = append(rv, v)
-		}
-	}
-
-	if t.Count == 1 {
-		if len(rv) == 1 {
-			return rv[0], nil
+		if d.includeFieldsRe != nil && !d.includeFieldsRe.MatchString(k) {
+			continue
 		}
+		tags[k] = v.Value
 	}
 
-	return rv, nil
-}
-
-// Code borrowed from exif.DateTime and adjusted.
-func tryParseDate(x *_exif.Exif, s string) (time.Time, error) {
-	dateStr := strings.TrimRight(s, "\x00")
-	// TODO(bep): look for timezone offset, GPS time, etc.
-	timeZone := time.Local
-	if tz, _ := x.TimeZone(); tz != nil {
-		timeZone = tz
-	}
-	return time.ParseInLocation(exifTimeLayout, dateStr, timeZone)
-}
+	ex = &ExifInfo{Lat: lat, Long: long, Date: tm, Tags: tags}
 
-type exifWalker struct {
-	x              *_exif.Exif
-	vals           map[string]any
-	includeMatcher *regexp.Regexp
-	excludeMatcher *regexp.Regexp
-}
-
-func (e *exifWalker) Walk(f _exif.FieldName, tag *tiff.Tag) error {
-	name := string(f)
-	if e.excludeMatcher != nil && e.excludeMatcher.MatchString(name) {
-		return nil
-	}
-	if e.includeMatcher != nil && !e.includeMatcher.MatchString(name) {
-		return nil
-	}
-	val, err := decodeTag(e.x, f, tag)
-	if err != nil {
-		return err
-	}
-	e.vals[name] = val
-	return nil
+	return
 }
 
-func nullString(in []byte) string {
-	var rv bytes.Buffer
-	for len(in) > 0 {
-		r, size := utf8.DecodeRune(in)
-		if unicode.IsGraphic(r) {
-			rv.WriteRune(r)
-		}
-		in = in[size:]
-	}
-	return rv.String()
+func getDateTime(tags map[string]imagemeta.TagInfo) time.Time {
+	return time.Time{}
 }
 
+// Code borrowed f
 var tcodec *tmc.Codec
 
 func init() {

resources/images/exif/exif_test.go

@@ -21,7 +21,7 @@ import (
 	"testing"
 	"time"
 
-	"github.com/gohugoio/hugo/htesting/hqt"
+	"github.com/bep/imagemeta"
 	"github.com/google/go-cmp/cmp"
 
 	qt "github.com/frankban/quicktest"
@@ -35,11 +35,12 @@ func TestExif(t *testing.T) {
 
 	d, err := NewDecoder(IncludeFields("Lens|Date"))
 	c.Assert(err, qt.IsNil)
-	x, err := d.Decode(f)
+	x, err := d.Decode(imagemeta.ImageFormatJPEG, f)
 	c.Assert(err, qt.IsNil)
 	c.Assert(x.Date.Format("2006-01-02"), qt.Equals, "2017-10-27")
 
 	// Malaga: https://goo.gl/taazZy
+
 	c.Assert(x.Lat, qt.Equals, float64(36.59744166666667))
 	c.Assert(x.Long, qt.Equals, float64(-4.50846))
 
@@ -51,7 +52,7 @@ func TestExif(t *testing.T) {
 
 	v, found = x.Tags["DateTime"]
 	c.Assert(found, qt.Equals, true)
-	c.Assert(v, hqt.IsSameType, time.Time{})
+	c.Assert(v, qt.Equals, "2017:11:23 09:56:54")
 
 	// Verify that it survives a round-trip to JSON and back.
 	data, err := json.Marshal(x)
@@ -72,8 +73,8 @@ func TestExifPNG(t *testing.T) {
 
 	d, err := NewDecoder()
 	c.Assert(err, qt.IsNil)
-	_, err = d.Decode(f)
-	c.Assert(err, qt.Not(qt.IsNil))
+	_, err = d.Decode(imagemeta.ImageFormatPNG, f)
+	c.Assert(err, qt.IsNil)
 }
 
 func TestIssue8079(t *testing.T) {
@@ -85,28 +86,11 @@ func TestIssue8079(t *testing.T) {
 
 	d, err := NewDecoder()
 	c.Assert(err, qt.IsNil)
-	x, err := d.Decode(f)
+	x, err := d.Decode(imagemeta.ImageFormatJPEG, f)
 	c.Assert(err, qt.IsNil)
 	c.Assert(x.Tags["ImageDescription"], qt.Equals, "Città del Vaticano #nanoblock #vatican #vaticancity")
 }
 
-func TestNullString(t *testing.T) {
-	c := qt.New(t)
-
-	for _, test := range []struct {
-		in     string
-		expect string
-	}{
-		{"foo", "foo"},
-		{"\x20", "\x20"},
-		{"\xc4\x81", "\xc4\x81"}, // \u0101
-		{"\u0160", "\u0160"},     // non-breaking space
-	} {
-		res := nullString([]byte(test.in))
-		c.Assert(res, qt.Equals, test.expect)
-	}
-}
-
 func BenchmarkDecodeExif(b *testing.B) {
 	c := qt.New(b)
 	f, err := os.Open(filepath.FromSlash("../../testdata/sunset.jpg"))
@@ -118,7 +102,7 @@ func BenchmarkDecodeExif(b *testing.B) {
 
 	b.ResetTimer()
 	for i := 0; i < b.N; i++ {
-		_, err = d.Decode(f)
+		_, err = d.Decode(imagemeta.ImageFormatJPEG, f)
 		c.Assert(err, qt.IsNil)
 		f.Seek(0, 0)
 	}
@@ -145,7 +129,7 @@ func TestIssue10738(t *testing.T) {
 
 		d, err := NewDecoder(IncludeFields(include))
 		c.Assert(err, qt.IsNil)
-		x, err := d.Decode(f)
+		x, err := d.Decode(imagemeta.ImageFormatJPEG, f)
 		c.Assert(err, qt.IsNil)
 
 		// Verify that it survives a round-trip to JSON and back.
@@ -194,7 +178,7 @@ func TestIssue10738(t *testing.T) {
 				include: "Lens|Date|ExposureTime",
 			}, want{
 				10,
-				0,
+				1,
 			},
 		},
 		{
@@ -221,7 +205,7 @@ func TestIssue10738(t *testing.T) {
 				include: "Lens|Date|ExposureTime",
 			}, want{
 				1,
-				0,
+				1,
 			},
 		},
 		{
@@ -266,7 +250,7 @@ func TestIssue10738(t *testing.T) {
 				include: "Lens|Date|ExposureTime",
 			}, want{
 				30,
-				0,
+				1,
 			},
 		},
 		{
@@ -293,7 +277,7 @@ func TestIssue10738(t *testing.T) {
 				include: "Lens|Date|ExposureTime",
 			}, want{
 				4,
-				0,
+				1,
 			},
 		},
 	}