Improve WebVTT to TTML conversion

subtitles.go

@@ -255,7 +255,7 @@ type StyleAttributes struct {
 	WebVTTItalics        bool
 	WebVTTLine           string
 	WebVTTLines          int
-	WebVTTPosition       string
+	WebVTTPosition       *WebVTTPosition
 	WebVTTRegionAnchor   string
 	WebVTTScroll         string
 	WebVTTSize           string
@@ -307,28 +307,28 @@ func (sa *StyleAttributes) propagateSRTAttributes() {
 	switch sa.SRTPosition {
 	case 7: // top-left
 		sa.WebVTTAlign = "left"
-		sa.WebVTTPosition = "10%"
+		sa.WebVTTPosition = newWebVTTPosition("10%")
 	case 8: // top-center
-		sa.WebVTTPosition = "10%"
+		sa.WebVTTPosition = newWebVTTPosition("10%")
 	case 9: // top-right
 		sa.WebVTTAlign = "right"
-		sa.WebVTTPosition = "10%"
+		sa.WebVTTPosition = newWebVTTPosition("10%")
 	case 4: // middle-left
 		sa.WebVTTAlign = "left"
-		sa.WebVTTPosition = "50%"
+		sa.WebVTTPosition = newWebVTTPosition("50%")
 	case 5: // middle-center
-		sa.WebVTTPosition = "50%"
+		sa.WebVTTPosition = newWebVTTPosition("50%")
 	case 6: // middle-right
 		sa.WebVTTAlign = "right"
-		sa.WebVTTPosition = "50%"
+		sa.WebVTTPosition = newWebVTTPosition("50%")
 	case 1: // bottom-left
 		sa.WebVTTAlign = "left"
-		sa.WebVTTPosition = "90%"
+		sa.WebVTTPosition = newWebVTTPosition("90%")
 	case 2: // bottom-center
-		sa.WebVTTPosition = "90%"
+		sa.WebVTTPosition = newWebVTTPosition("90%")
 	case 3: // bottom-right
 		sa.WebVTTAlign = "right"
-		sa.WebVTTPosition = "90%"
+		sa.WebVTTPosition = newWebVTTPosition("90%")
 	}
 
 	sa.WebVTTBold = sa.SRTBold
@@ -411,10 +411,10 @@ func (sa *StyleAttributes) propagateTTMLAttributes() {
 		coordinates := strings.Split(*sa.TTMLOrigin, " ")
 		if len(coordinates) > 1 {
 			sa.WebVTTLine = coordinates[0]
-			sa.WebVTTPosition = coordinates[1]
+			sa.WebVTTPosition = newWebVTTPosition(coordinates[1])
 			if sa.TTMLWritingMode != nil && strings.HasPrefix(*sa.TTMLWritingMode, "tb") {
 				sa.WebVTTLine = coordinates[1]
-				sa.WebVTTPosition = coordinates[0]
+				sa.WebVTTPosition = newWebVTTPosition(coordinates[0])
 			}
 		}
 	}
@@ -428,6 +428,77 @@ func (sa *StyleAttributes) propagateWebVTTAttributes() {
 	sa.SRTBold = sa.WebVTTBold
 	sa.SRTItalics = sa.WebVTTItalics
 	sa.SRTUnderline = sa.WebVTTUnderline
+
+	// may be overridden by position parsing later
+	switch sa.WebVTTAlign {
+	case "left", "right", "center", "start", "end":
+		sa.TTMLTextAlign = astikit.StrPtr(sa.WebVTTAlign)
+	}
+
+	for _, tag := range sa.WebVTTTags {
+		switch tag.Name {
+		case "c":
+			if len(tag.Classes) > 0 {
+				for _, color := range tag.Classes {
+					if strings.HasPrefix(color, "bg_") && len(color) > 3 {
+						if bgColor, err := newColorFromWebVTTString(color[3:]); err == nil {
+							sa.TTMLBackgroundColor = astikit.StrPtr("#" + bgColor.TTMLString())
+						}
+					} else {
+						if fgColor, err := newColorFromWebVTTString(color); err == nil {
+							sa.TTMLColor = astikit.StrPtr("#" + fgColor.TTMLString())
+						}
+					}
+				}
+			}
+		}
+	}
+
+	// Handle WebVTT position and alignment conversion
+	sa.propagateWebVTTPosition()
+
+	if sa.WebVTTSize != "" {
+		sa.TTMLExtent = astikit.StrPtr(fmt.Sprintf("%s 10%%", sa.WebVTTSize))
+	}
+}
+
+// propagateWebVTTPosition handles WebVTT position and line attributes conversion
+func (sa *StyleAttributes) propagateWebVTTPosition() {
+	var yPos string
+	var hasPosition, hasLine bool
+
+	// Parse position if available
+	if sa.WebVTTPosition != nil {
+		hasPosition = true
+
+		// Handle position alignment (takes precedence over WebVTTAlign)
+		switch sa.WebVTTPosition.Alignment {
+		case "line-left":
+			sa.TTMLTextAlign = astikit.StrPtr("left")
+		case "center":
+			sa.TTMLTextAlign = astikit.StrPtr("center")
+		case "line-right":
+			sa.TTMLTextAlign = astikit.StrPtr("right")
+		}
+	}
+
+	// Handle line if available
+	if sa.WebVTTLine != "" {
+		hasLine = true
+		yPos = sa.WebVTTLine
+	}
+
+	// Set TTMLOrigin based on available position and line data
+	if hasPosition && hasLine {
+		// Both position and line are available
+		sa.TTMLOrigin = astikit.StrPtr(fmt.Sprintf("%s %s", sa.WebVTTPosition.XPosition, yPos))
+	} else if hasPosition {
+		// Only position is available, use default Y position (80% for bottom)
+		sa.TTMLOrigin = astikit.StrPtr(fmt.Sprintf("%s 80%%", sa.WebVTTPosition.XPosition))
+	} else if hasLine {
+		// Only line is available, use default X position (10% for left)
+		sa.TTMLOrigin = astikit.StrPtr(fmt.Sprintf("10%% %s", yPos))
+	}
 }
 
 // merge - base on parent, override style attributes if defined in child

testdata/example-in.vtt

@@ -33,7 +33,7 @@ NOTE This a comment inside the VTT
 and this is the second line
 
 2
-00:02:04.08 --> 00:02:07.12  region:fred position:10%,start align:left size:35%
+00:02:04.08 --> 00:02:07.12  region:fred position:10%,line-left align:left size:35%
 MAN:
 How did we end up here?
 

testdata/example-out.vtt

@@ -29,7 +29,7 @@ NOTE This a comment inside the VTT
 and this is the second line
 
 2
-00:02:04.080 --> 00:02:07.120 align:left position:10%,start region:fred size:35%
+00:02:04.080 --> 00:02:07.120 align:left position:10%,line-left region:fred size:35%
 MAN:
 How did we end up here?
 

ttml_test.go

@@ -21,9 +21,9 @@ func TestTTML(t *testing.T) {
 	assert.Equal(t, &astisub.Metadata{Framerate: 25, Language: astisub.LanguageFrench, Title: "Title test", TTMLCopyright: "Copyright test"}, s.Metadata)
 	// Styles
 	assert.Equal(t, 3, len(s.Styles))
-	assert.Equal(t, astisub.Style{ID: "style_0", InlineStyle: &astisub.StyleAttributes{TTMLColor: astikit.StrPtr("white"), TTMLExtent: astikit.StrPtr("100% 10%"), TTMLFontFamily: astikit.StrPtr("sansSerif"), TTMLFontStyle: astikit.StrPtr("normal"), TTMLOrigin: astikit.StrPtr("0% 90%"), TTMLTextAlign: astikit.StrPtr("center"), WebVTTAlign: "center", WebVTTLine: "0%", WebVTTLines: 2, WebVTTPosition: "90%", WebVTTRegionAnchor: "0%,0%", WebVTTScroll: "up", WebVTTSize: "10%", WebVTTViewportAnchor: "0%,90%", WebVTTWidth: "100%"}, Style: s.Styles["style_2"]}, *s.Styles["style_0"])
-	assert.Equal(t, astisub.Style{ID: "style_1", InlineStyle: &astisub.StyleAttributes{TTMLColor: astikit.StrPtr("white"), TTMLExtent: astikit.StrPtr("100% 13%"), TTMLFontFamily: astikit.StrPtr("sansSerif"), TTMLFontStyle: astikit.StrPtr("normal"), TTMLOrigin: astikit.StrPtr("0% 87%"), TTMLTextAlign: astikit.StrPtr("center"), WebVTTAlign: "center", WebVTTLine: "0%", WebVTTLines: 2, WebVTTPosition: "87%", WebVTTRegionAnchor: "0%,0%", WebVTTScroll: "up", WebVTTSize: "13%", WebVTTViewportAnchor: "0%,87%", WebVTTWidth: "100%"}}, *s.Styles["style_1"])
-	assert.Equal(t, astisub.Style{ID: "style_2", InlineStyle: &astisub.StyleAttributes{TTMLColor: astikit.StrPtr("white"), TTMLExtent: astikit.StrPtr("100% 20%"), TTMLFontFamily: astikit.StrPtr("sansSerif"), TTMLFontStyle: astikit.StrPtr("normal"), TTMLOrigin: astikit.StrPtr("0% 80%"), TTMLTextAlign: astikit.StrPtr("center"), WebVTTAlign: "center", WebVTTLine: "0%", WebVTTLines: 4, WebVTTPosition: "80%", WebVTTRegionAnchor: "0%,0%", WebVTTScroll: "up", WebVTTSize: "20%", WebVTTViewportAnchor: "0%,80%", WebVTTWidth: "100%"}}, *s.Styles["style_2"])
+	assert.Equal(t, astisub.Style{ID: "style_0", InlineStyle: &astisub.StyleAttributes{TTMLColor: astikit.StrPtr("white"), TTMLExtent: astikit.StrPtr("100% 10%"), TTMLFontFamily: astikit.StrPtr("sansSerif"), TTMLFontStyle: astikit.StrPtr("normal"), TTMLOrigin: astikit.StrPtr("0% 90%"), TTMLTextAlign: astikit.StrPtr("center"), WebVTTAlign: "center", WebVTTLine: "0%", WebVTTLines: 2, WebVTTPosition: &astisub.WebVTTPosition{XPosition: "90%"}, WebVTTRegionAnchor: "0%,0%", WebVTTScroll: "up", WebVTTSize: "10%", WebVTTViewportAnchor: "0%,90%", WebVTTWidth: "100%"}, Style: s.Styles["style_2"]}, *s.Styles["style_0"])
+	assert.Equal(t, astisub.Style{ID: "style_1", InlineStyle: &astisub.StyleAttributes{TTMLColor: astikit.StrPtr("white"), TTMLExtent: astikit.StrPtr("100% 13%"), TTMLFontFamily: astikit.StrPtr("sansSerif"), TTMLFontStyle: astikit.StrPtr("normal"), TTMLOrigin: astikit.StrPtr("0% 87%"), TTMLTextAlign: astikit.StrPtr("center"), WebVTTAlign: "center", WebVTTLine: "0%", WebVTTLines: 2, WebVTTPosition: &astisub.WebVTTPosition{XPosition: "87%"}, WebVTTRegionAnchor: "0%,0%", WebVTTScroll: "up", WebVTTSize: "13%", WebVTTViewportAnchor: "0%,87%", WebVTTWidth: "100%"}}, *s.Styles["style_1"])
+	assert.Equal(t, astisub.Style{ID: "style_2", InlineStyle: &astisub.StyleAttributes{TTMLColor: astikit.StrPtr("white"), TTMLExtent: astikit.StrPtr("100% 20%"), TTMLFontFamily: astikit.StrPtr("sansSerif"), TTMLFontStyle: astikit.StrPtr("normal"), TTMLOrigin: astikit.StrPtr("0% 80%"), TTMLTextAlign: astikit.StrPtr("center"), WebVTTAlign: "center", WebVTTLine: "0%", WebVTTLines: 4, WebVTTPosition: &astisub.WebVTTPosition{XPosition: "80%"}, WebVTTRegionAnchor: "0%,0%", WebVTTScroll: "up", WebVTTSize: "20%", WebVTTViewportAnchor: "0%,80%", WebVTTWidth: "100%"}}, *s.Styles["style_2"])
 	// Regions
 	assert.Equal(t, 3, len(s.Regions))
 	assert.Equal(t, astisub.Region{ID: "region_0", Style: s.Styles["style_0"], InlineStyle: &astisub.StyleAttributes{TTMLColor: astikit.StrPtr("blue")}}, *s.Regions["region_0"])

webvtt.go

@@ -37,6 +37,39 @@ var (
 	webVTTRegexpTag                    = regexp.MustCompile(`(</*\s*([^\.\s]+)(\.[^\s/]*)*\s*([^/]*)\s*/*>)`)
 )
 
+type WebVTTPosition struct {
+	XPosition string
+	Alignment string
+}
+
+// newWebVTTPosition creates a new WebVTTPosition from a string.
+// The string can be in the format "XPosition,Alignment" or just "XPosition".
+func newWebVTTPosition(s string) *WebVTTPosition {
+	if s == "" {
+		return nil
+	}
+
+	parts := strings.Split(s, ",")
+	if len(parts) != 2 {
+		return &WebVTTPosition{XPosition: strings.TrimSpace(s)}
+	}
+
+	return &WebVTTPosition{
+		XPosition: strings.TrimSpace(parts[0]),
+		Alignment: strings.TrimSpace(parts[1]),
+	}
+}
+
+func (p *WebVTTPosition) String() string {
+	if p == nil {
+		return ""
+	}
+	if p.Alignment != "" {
+		return fmt.Sprintf("%s,%s", p.XPosition, p.Alignment)
+	}
+	return p.XPosition
+}
+
 // parseDurationWebVTT parses a .vtt duration
 func parseDurationWebVTT(i string) (time.Duration, error) {
 	return parseDuration(i, ".", 3)
@@ -272,7 +305,7 @@ func ReadFromWebVTT(i io.Reader) (o *Subtitles, err error) {
 					case "line":
 						item.InlineStyle.WebVTTLine = split[1]
 					case "position":
-						item.InlineStyle.WebVTTPosition = split[1]
+						item.InlineStyle.WebVTTPosition = newWebVTTPosition(split[1])
 					case "region":
 						if _, ok := o.Regions[split[1]]; !ok {
 							err = fmt.Errorf("astisub: line %d: Unknown region %s", lineNum, split[1])
@@ -574,12 +607,12 @@ func (s Subtitles) WriteToWebVTT(o io.Writer) (err error) {
 				c = append(c, bytesSpace...)
 				c = append(c, []byte("line:"+item.Style.InlineStyle.WebVTTLine)...)
 			}
-			if item.InlineStyle.WebVTTPosition != "" {
+			if item.InlineStyle.WebVTTPosition != nil {
 				c = append(c, bytesSpace...)
-				c = append(c, []byte("position:"+item.InlineStyle.WebVTTPosition)...)
-			} else if item.Style != nil && item.Style.InlineStyle != nil && item.Style.InlineStyle.WebVTTPosition != "" {
+				c = append(c, []byte("position:"+item.InlineStyle.WebVTTPosition.String())...)
+			} else if item.Style != nil && item.Style.InlineStyle != nil && item.Style.InlineStyle.WebVTTPosition != nil {
 				c = append(c, bytesSpace...)
-				c = append(c, []byte("position:"+item.Style.InlineStyle.WebVTTPosition)...)
+				c = append(c, []byte("position:"+item.Style.InlineStyle.WebVTTPosition.String())...)
 			}
 			if item.Region != nil {
 				c = append(c, bytesSpace...)
@@ -648,10 +681,21 @@ func (li LineItem) webVTTBytes(previous, next *LineItem) (c []byte) {
 		c = append(c, []byte("<"+formatDurationWebVTT(li.StartAt)+">")...)
 	}
 
-	// Get color
+	// Get color - only add TTMLColor-based tag if there are no WebVTT color tags
 	var color string
-	if li.InlineStyle != nil && li.InlineStyle.TTMLColor != nil {
-		color = cssColor(*li.InlineStyle.TTMLColor)
+	var hasColorTags bool
+	if li.InlineStyle != nil {
+		// Check if we have WebVTT color tags
+		for _, tag := range li.InlineStyle.WebVTTTags {
+			if tag.Name == "c" {
+				hasColorTags = true
+				break
+			}
+		}
+		// Only use TTMLColor if we don't have WebVTT color tags
+		if !hasColorTags && li.InlineStyle.TTMLColor != nil {
+			color = cssColor(*li.InlineStyle.TTMLColor)
+		}
 	}
 
 	// Append
@@ -692,3 +736,62 @@ func cssColor(rgb string) string {
 	}
 	return colors[strings.ToLower(rgb)] // returning the empty string is ok
 }
+
+// webVTTColorMap maps WebVTT color class names to Color instances
+var webVTTColorMap = map[string]*Color{
+	"black":   ColorBlack,
+	"red":     ColorRed,
+	"green":   ColorGreen,
+	"yellow":  ColorYellow,
+	"blue":    ColorBlue,
+	"magenta": ColorMagenta,
+	"cyan":    ColorCyan,
+	"white":   ColorWhite,
+	"silver":  ColorSilver,
+	"gray":    ColorGray,
+	"maroon":  ColorMaroon,
+	"olive":   ColorOlive,
+	"lime":    ColorLime,
+	"teal":    ColorTeal,
+	"navy":    ColorNavy,
+	"purple":  ColorPurple,
+}
+
+func newColorFromWebVTTString(color string) (*Color, error) {
+	switch color {
+	case "black":
+		return ColorBlack, nil
+	case "red":
+		return ColorRed, nil
+	case "green":
+		return ColorGreen, nil
+	case "yellow":
+		return ColorYellow, nil
+	case "blue":
+		return ColorBlue, nil
+	case "magenta":
+		return ColorMagenta, nil
+	case "cyan":
+		return ColorCyan, nil
+	case "white":
+		return ColorWhite, nil
+	case "silver":
+		return ColorSilver, nil
+	case "gray":
+		return ColorGray, nil
+	case "maroon":
+		return ColorMaroon, nil
+	case "olive":
+		return ColorOlive, nil
+	case "lime":
+		return ColorLime, nil
+	case "teal":
+		return ColorTeal, nil
+	case "navy":
+		return ColorNavy, nil
+	case "purple":
+		return ColorPurple, nil
+	default:
+		return nil, fmt.Errorf("unknown color class %s", color)
+	}
+}