Preserve br tag in ttml (#106)

* Preserve br tag in ttml * [ttml] Change to more appropriate name and optimize ttml xml decoder
asticode · Jun 17, 2024 · 35e8bad · 35e8bad
1 parent 2f42372
commit 35e8bad
Show file tree

Hide file tree

Showing 4 changed files with 108 additions and 2 deletions.
diff --git a/testdata/example-in-breaklines.ttml b/testdata/example-in-breaklines.ttml
@@ -0,0 +1,24 @@
+<tt>
+  <head>
+  </head>
+  <body>
+    <div>
+      <p xml:id="1" begin="00:00:00.000" end="00:00:01.000">
+        <span>First line<br/>
+Second line</span>
+      </p>
+      <p xml:id="2" begin="00:00:01.000" end="00:00:02.000">
+        <span>Third line<br></br>Fourth line</span>
+      </p>
+      <p xml:id="3" begin="00:00:02.000" end="00:00:03.000">
+                Fifth line
+        <br/>
+                Sixth <span>middle</span> line
+      </p>
+      <p xml:id="4" begin="00:00:03.000" end="00:00:04.000">
+        Seventh line 
+        <br></br>Eighth <span>middle</span> line
+      </p>
+    </div>
+  </body>
+</tt>
diff --git a/testdata/example-out-breaklines.ttml b/testdata/example-out-breaklines.ttml
@@ -0,0 +1,32 @@
+<tt xmlns="http://www.w3.org/ns/ttml" xmlns:ttm="http://www.w3.org/ns/ttml#metadata" xmlns:tts="http://www.w3.org/ns/ttml#styling">
+    <head>
+        <styling></styling>
+        <layout></layout>
+    </head>
+    <body>
+        <div>
+            <p begin="00:00:00.000" end="00:00:01.000">
+                <span>First line</span>
+                <br></br>
+                <span>Second line</span>
+            </p>
+            <p begin="00:00:01.000" end="00:00:02.000">
+                <span>Third lineFourth line</span>
+            </p>
+            <p begin="00:00:02.000" end="00:00:03.000">
+                <span>Fifth line</span>
+                <br></br>
+                <span>Sixth </span>
+                <span>middle </span>
+                <span>line</span>
+            </p>
+            <p begin="00:00:03.000" end="00:00:04.000">
+                <span>Seventh line</span>
+                <br></br>
+                <span>Eighth </span>
+                <span>middle </span>
+                <span>line</span>
+            </p>
+        </div>
+    </body>
+</tt>
diff --git a/ttml.go b/ttml.go
@@ -197,6 +197,38 @@ func (i *TTMLInItems) UnmarshalXML(d *xml.Decoder, start xml.StartElement) (err
 	return nil
 }
 
+type ttmlXmlDecoder 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 *ttmlXmlDecoder) 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 newTTMLXmlDecoder(ts TTMLInSubtitle) *ttmlXmlDecoder {
+	return &ttmlXmlDecoder{Decoder: *xml.NewDecoder(strings.NewReader("<p>" + ts.Items + "</p>")), holdingToken: nil}
+}
+
 // TTMLInItem represents an input TTML item
 type TTMLInItem struct {
 	Style string `xml:"style,attr,omitempty"`
@@ -380,7 +412,7 @@ 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 {
+		if err = newTTMLXmlDecoder(ts).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 TestTTMLBreakLines(t *testing.T) {
+	// Open
+	s, err := astisub.OpenFile("./testdata/example-in-breaklines.ttml")
+	assert.NoError(t, err)
+
+	// Write
+	w := &bytes.Buffer{}
+	err = s.WriteToTTML(w)
+	assert.NoError(t, err)
+
+	c, err := ioutil.ReadFile("./testdata/example-out-breaklines.ttml")
+	assert.NoError(t, err)
+
+	assert.Equal(t, strings.TrimSpace(string(c)), strings.TrimSpace(w.String()))
+}