[ttml] Change to more appropriate name and optimize ttml xml decoder

asticode · Jun 17, 2024 · 2f60105 · 2f60105
1 parent e403fea
commit 2f60105
Show file tree

Hide file tree

Showing 6 changed files with 68 additions and 86 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/testdata/example-with-breaklines-in.ttml b/testdata/example-with-breaklines-in.ttml
diff --git a/testdata/example-with-breaklines-out.ttml b/testdata/example-with-breaklines-out.ttml
diff --git a/ttml.go b/ttml.go
@@ -197,16 +197,16 @@ 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 {
+// ttmlXmlDecoder is used only for decoding TTMLInItems, do not use it anywhere else
+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 *handleBrTokenReader) Token() (xml.Token, error) {
+func (r *ttmlXmlDecoder) Token() (xml.Token, error) {
 	if r.holdingToken != nil {
 		returnToken := r.holdingToken
 		r.holdingToken = nil
@@ -226,8 +226,8 @@ func (r *handleBrTokenReader) Token() (xml.Token, error) {
 	return t, nil
 }
 
-func newHandleBrTokenReader(r io.Reader) xml.TokenReader {
-	return &handleBrTokenReader{Decoder: *xml.NewDecoder(r), holdingToken: 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
@@ -413,8 +413,7 @@ func ReadFromTTML(i io.Reader) (o *Subtitles, err error) {
 
 		// Unmarshal items
 		var items = TTMLInItems{}
-		decoder := xml.NewTokenDecoder(newHandleBrTokenReader(strings.NewReader("<p>" + ts.Items + "</p>")))
-		if err = decoder.Decode(&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
@@ -2,6 +2,7 @@ package astisub_test
 
 import (
 	"bytes"
+	"fmt"
 	"io/ioutil"
 	"strings"
 	"testing"
@@ -53,18 +54,20 @@ func TestTTML(t *testing.T) {
 	assert.Equal(t, string(c), w.String())
 }
 
-func TestBreakLineHandling(t *testing.T) {
+func TestTTMLBreakLines(t *testing.T) {
 	// Open
-	s, err := astisub.OpenFile("./testdata/example-with-breaklines-in.ttml")
+	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-with-breaklines-out.ttml")
+	c, err := ioutil.ReadFile("./testdata/example-out-breaklines.ttml")
 	assert.NoError(t, err)
 
+	fmt.Println(w.String())
+
 	assert.Equal(t, strings.TrimSpace(string(c)), strings.TrimSpace(w.String()))
 }