Preserve br tag in ttml

asticode · May 31, 2024 · e403fea · e403fea
1 parent 2d03157
commit e403fea
Show file tree

Hide file tree

Showing 4 changed files with 130 additions and 2 deletions.
diff --git a/testdata/example-with-breaklines-in.ttml b/testdata/example-with-breaklines-in.ttml
@@ -0,0 +1,34 @@
+<tt xmlns:xs="http://www.w3.org/2001/XMLSchema" xmlns="http://www.w3.org/2006/10/ttaf1" xmlns:tt="http://www.w3.org/2006/10/ttaf1" xmlns:tts="http://www.w3.org/2006/10/ttaf1#styling" xmlns:ttp="http://www.w3.org/2006/10/ttaf1#parameter" xmlns:ttm="http://www.w3.org/2006/10/ttaf1#metadata" xml:lang="fr-FR" ttp:timeBase="smpte" ttp:frameRate="25" ttp:frameRateMultiplier="1:1" ttp:markerMode="discontinuous">
+    <head>
+        <metadata>
+          <ttm:title>Title test</ttm:title>
+          <ttm:copyright>Copyright test</ttm:copyright>
+        </metadata>
+        <styling>
+            <style xml:id="style_0" tts:fontFamily="sansSerif" tts:color="white" tts:fontStyle="normal" tts:textAlign="center" tts:origin="0% 90%" tts:extent="100% 10%"/>
+        </styling>
+        <layout>
+            <region xml:id="region_0" tt:style="style_0" tt:color="blue"/>
+        </layout>
+    </head>
+    <body>
+        <div>
+            <p xml:id="sub_1" begin="00:00:00.000" end="00:00:01.000" region="region_0" style="style_0" color="red">
+                <span style="style_0" color="black">First line<br/>Second line</span>
+            </p>
+            <p xml:id="sub_2" begin="00:00:01.000" end="00:00:02.000" region="region_0" style="style_0" color="red">
+                <span style="style_0" color="black">Third line<br></br>Fourth line</span>
+            </p>
+            <p xml:id="sub_3" begin="00:00:02.000" end="00:00:03.000" region="region_0">
+                Fifth line
+                <br/>
+                Sixth <span style="style_0" color="green">middle</span> line
+            </p>
+            <p xml:id="sub_4" begin="00:00:03.000" end="00:00:04.000" region="region_0">
+                Seventh line
+                <br></br>
+                Eighth <span style="style_0" color="green">middle</span> line
+            </p>
+        </div>
+    </body>
+</tt>
diff --git a/testdata/example-with-breaklines-out.ttml b/testdata/example-with-breaklines-out.ttml
@@ -0,0 +1,42 @@
+<tt xmlns="http://www.w3.org/ns/ttml" xml:lang="fr" xmlns:ttm="http://www.w3.org/ns/ttml#metadata" xmlns:tts="http://www.w3.org/ns/ttml#styling">
+    <head>
+        <metadata>
+            <ttm:copyright>Copyright test</ttm:copyright>
+            <ttm:title>Title test</ttm:title>
+        </metadata>
+        <styling>
+            <style xml:id="style_0" tts:color="white" tts:extent="100% 10%" tts:fontFamily="sansSerif" tts:fontStyle="normal" tts:origin="0% 90%" tts:textAlign="center"></style>
+        </styling>
+        <layout>
+            <region xml:id="region_0" style="style_0" tts:color="blue"></region>
+        </layout>
+    </head>
+    <body>
+        <div>
+            <p begin="00:00:00.000" end="00:00:01.000" region="region_0" style="style_0" tts:color="red">
+                <span style="style_0" tts:color="black">First line</span>
+                <br></br>
+                <span style="style_0" tts:color="black">Second line</span>
+            </p>
+            <p begin="00:00:01.000" end="00:00:02.000" region="region_0" style="style_0" tts:color="red">
+                <span style="style_0" tts:color="black">Third line</span>
+                <br></br>
+                <span style="style_0" tts:color="black">Fourth line</span>
+            </p>
+            <p begin="00:00:02.000" end="00:00:03.000" region="region_0">
+                <span>Fifth line</span>
+                <br></br>
+                <span>Sixth </span>
+                <span style="style_0" tts:color="green">middle </span>
+                <span>line</span>
+            </p>
+            <p begin="00:00:03.000" end="00:00:04.000" region="region_0">
+                <span>Seventh line</span>
+                <br></br>
+                <span>Eighth </span>
+                <span style="style_0" tts:color="green">middle </span>
+                <span>line</span>
+            </p>
+        </div>
+    </body>
+</tt>
diff --git a/ttml.go b/ttml.go
@@ -197,6 +197,39 @@ func (i *TTMLInItems) UnmarshalXML(d *xml.Decoder, start xml.StartElement) (err
 	return nil
 }
 
+// handleBrTokenReader is used only for decoding TTMLInItems, do not use it anywhere else
+type handleBrTokenReader struct {
+	xml.Decoder
+	holdingToken xml.Token
+}
+
+// Token implements the TokenReader interface, when it meets the "br" tag, it will hold the token and return a newline
+// instead. This is to work around the fact that the go xml unmarshaler will ignore the "br" tag if it's within a
+// character data field.
+func (r *handleBrTokenReader) Token() (xml.Token, error) {
+	if r.holdingToken != nil {
+		returnToken := r.holdingToken
+		r.holdingToken = nil
+		return returnToken, nil
+	}
+
+	t, err := r.Decoder.Token()
+	if err != nil {
+		return nil, err
+	}
+
+	if se, ok := t.(xml.StartElement); ok && strings.ToLower(se.Name.Local) == "br" {
+		r.holdingToken = t
+		return xml.CharData("\n"), nil
+	}
+
+	return t, nil
+}
+
+func newHandleBrTokenReader(r io.Reader) xml.TokenReader {
+	return &handleBrTokenReader{Decoder: *xml.NewDecoder(r), holdingToken: nil}
+}
+
 // TTMLInItem represents an input TTML item
 type TTMLInItem struct {
 	Style string `xml:"style,attr,omitempty"`
@@ -380,7 +413,8 @@ func ReadFromTTML(i io.Reader) (o *Subtitles, err error) {
 
 		// Unmarshal items
 		var items = TTMLInItems{}
-		if err = xml.Unmarshal([]byte("<span>"+ts.Items+"</span>"), &items); err != nil {
+		decoder := xml.NewTokenDecoder(newHandleBrTokenReader(strings.NewReader("<p>" + ts.Items + "</p>")))
+		if err = decoder.Decode(&items); err != nil {
 			err = fmt.Errorf("astisub: unmarshaling items failed: %w", err)
 			return
 		}

diff --git a/ttml_test.go b/ttml_test.go
@@ -2,10 +2,12 @@ package astisub_test
 
 import (
 	"bytes"
-	"github.com/asticode/go-astikit"
 	"io/ioutil"
+	"strings"
 	"testing"
 
+	"github.com/asticode/go-astikit"
+
 	"github.com/asticode/go-astisub"
 	"github.com/stretchr/testify/assert"
 )
@@ -50,3 +52,19 @@ func TestTTML(t *testing.T) {
 	assert.NoError(t, err)
 	assert.Equal(t, string(c), w.String())
 }
+
+func TestBreakLineHandling(t *testing.T) {
+	// Open
+	s, err := astisub.OpenFile("./testdata/example-with-breaklines-in.ttml")
+	assert.NoError(t, err)
+
+	// Write
+	w := &bytes.Buffer{}
+	err = s.WriteToTTML(w)
+	assert.NoError(t, err)
+
+	c, err := ioutil.ReadFile("./testdata/example-with-breaklines-out.ttml")
+	assert.NoError(t, err)
+
+	assert.Equal(t, strings.TrimSpace(string(c)), strings.TrimSpace(w.String()))
+}