Update app.py

app.py

@@ -468,92 +468,6 @@ def wrap_text_for_ass(text: str, max_width: int, font_family: str, font_size: in
     
     return "\\N".join(lines)
 
-
-def create_caption_ass(captions: List[CaptionSegment], style: CaptionStyle, work_dir: str, font_path: str) -> str:
-    """Create ASS subtitle file for captions with enhanced colors and outlines"""
-    
-    # Get actual font family name
-    font_family_name = get_font_family_name(font_path)
-    print(f"📝 Caption font family: {font_family_name}")
-    
-    # Get font color and convert from RRGGBB to BBGGRR for ASS
-    font_color_rgb = COLOR_MAP.get(style.color.lower(), "FFFFFF")
-    font_color_bgr = font_color_rgb[4:6] + font_color_rgb[2:4] + font_color_rgb[0:2]
-    print(f"🎨 Font color (text): {style.color} -> RGB={font_color_rgb} -> BGR={font_color_bgr}")
-    
-    # Get outline color and convert to BGR
-    outline_color_rgb = COLOR_MAP.get(style.outline_color.lower(), "000000")
-    outline_color_bgr = outline_color_rgb[4:6] + outline_color_rgb[2:4] + outline_color_rgb[0:2]
-    print(f"✏️ Outline color (border): {style.outline_color} -> RGB={outline_color_rgb} -> BGR={outline_color_bgr}, Width={style.outline_width}")
-    
-    # Parse background color with proper alpha calculation
-    bg_parts = style.bg_color.split('@')
-    bg_color_name = bg_parts[0].lower()
-    bg_opacity = float(bg_parts[1]) if len(bg_parts) > 1 else 0.5
-    bg_opacity = max(0, min(1, bg_opacity))
-    bg_color_rgb = COLOR_MAP.get(bg_color_name, "000000")
-    bg_color_bgr = bg_color_rgb[4:6] + bg_color_rgb[2:4] + bg_color_rgb[0:2]
-    bg_alpha = int((1 - bg_opacity) * 255)
-    print(f"🎨 BG color (rectangle): {bg_color_name} -> RGB={bg_color_rgb} -> BGR={bg_color_bgr}, Opacity={bg_opacity}, Alpha={bg_alpha}")
-    
-    # Map position to ASS alignment
-    pos_map = {
-        "bottom-left": 1, "bottom-center": 2, "bottom-right": 3,
-        "left": 4, "center": 5, "right": 6,
-        "top-left": 7, "top-center": 8, "top-right": 9
-    }
-    alignment = pos_map.get(style.position, 2)
-    
-    # Calculate margins
-    margin_l = style.margin if alignment in [1,4,7] else 0
-    margin_r = style.margin if alignment in [3,6,9] else 0
-    margin_v = style.margin
-    
-    # Create ASS header with proper border settings
-    # BorderStyle=3 gives solid box, Outline adds text border
-    # Format: &HAABBGGRR where AA is alpha, BB blue, GG green, RR red
-    ass_header = f"""[Script Info]
-; Script generated by Video Styling Space - Auto Caption
-ScriptType: v4.00+
-PlayResX: 1920
-PlayResY: 1080
-ScaledBorderAndShadow: yes
-WrapStyle: 1
-
-[V4+ Styles]
-Format: Name, Fontname, Fontsize, PrimaryColour, SecondaryColour, OutlineColour, BackColour, Bold, Italic, Underline, StrikeOut, ScaleX, ScaleY, Spacing, Angle, BorderStyle, Outline, Shadow, Alignment, MarginL, MarginR, MarginV, Encoding
-Style: Default,{font_family_name},{style.font_size},&H00{font_color_bgr},&H000000FF,&H00{outline_color_bgr},&H{bg_alpha:02X}{bg_color_bgr},0,0,0,0,100,100,0,0,3,{style.outline_width},0,{alignment},{margin_l},{margin_r},{margin_v},1
-
-[Events]
-Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV, Effect, Text
-"""
-    
-    # Add caption events
-    ass_events = []
-    for i, caption in enumerate(captions):
-        start = format_ass_time(caption.start_time)
-        end = format_ass_time(caption.end_time)
-        
-        text = caption.text
-        if style.max_width > 0:
-            text = wrap_text_for_ass(text, style.max_width, font_family_name, style.font_size)
-        
-        ass_events.append(f"Dialogue: 0,{start},{end},Default,,0,0,0,,{text}")
-    
-    # Write ASS file
-    ass_file = os.path.join(work_dir, "captions.ass")
-    with open(ass_file, 'w', encoding='utf-8') as f:
-        f.write(ass_header + "\n".join(ass_events))
-    
-    print("\n=== ASS FILE DEBUG ===")
-    with open(ass_file, 'r', encoding='utf-8') as f:
-        print(f.read())
-    print("=== END ASS DEBUG ===\n")
-    
-    print(f"📝 Created caption ASS file with {len(captions)} captions")
-    print(f"📝 Style line: BorderStyle=3, Outline={style.outline_width}, OutlineColor=&H00{outline_color_bgr}, BackColor=&H{bg_alpha:02X}{bg_color_bgr}")
-    return ass_file
-    
 # =============================================
 # SRT PARSING FUNCTIONS
 # =============================================
@@ -615,7 +529,7 @@ def srt_time_to_seconds(time_str: str) -> float:
 # =============================================
 
 def create_text_overlay(input_video, output_video, text_style):
-    """Add text overlay using ASS subtitles with enhanced colors and outlines"""
+    """Add text overlay using drawtext method with background box and outline"""
     font_path = get_font_path(text_style.font_family)
     if not font_path:
         print(f"⚠️ Font not found: {text_style.font_family}")
@@ -623,116 +537,44 @@ def create_text_overlay(input_video, output_video, text_style):
     
     print(f"✅ Using font: {font_path}")
     
-    # Get the actual font family name from the font file
-    font_family_name = get_font_family_name(font_path)
-    print(f"📝 Font family name: {font_family_name}")
-    
-    # Create working directory for ASS file
-    work_dir = os.path.dirname(output_video)
-    ass_file = os.path.join(work_dir, "subtitle.ass")
-    
-    # Color mapping
-    LOCAL_COLOR_MAP = {
-        "white": "FFFFFF", "black": "000000", "red": "FF0000",
-        "green": "00FF00", "blue": "0000FF", "yellow": "FFFF00",
-        "gold": "FFD700", "purple": "800080", "magenta": "FF00FF",
-        "cyan": "00FFFF", "orange": "FFA500", "pink": "FFC0CB"
-    }
-    
-    # Get font color and convert from RRGGBB to BBGGRR for ASS
-    font_color_rgb = LOCAL_COLOR_MAP.get(text_style.color.lower(), "FFFFFF")
-    font_color_bgr = font_color_rgb[4:6] + font_color_rgb[2:4] + font_color_rgb[0:2]
-    print(f"🎨 Font color (text): {text_style.color} -> RGB={font_color_rgb} -> BGR={font_color_bgr}")
-    
-    # Get outline color and convert to BGR
-    outline_color_rgb = LOCAL_COLOR_MAP.get(text_style.outline_color.lower(), "000000")
-    outline_color_bgr = outline_color_rgb[4:6] + outline_color_rgb[2:4] + outline_color_rgb[0:2]
-    print(f"✏️ Outline color (border): {text_style.outline_color} -> RGB={outline_color_rgb} -> BGR={outline_color_bgr}, Width={text_style.outline_width}")
-    
     # Parse background color
     bg_parts = text_style.bg_color.split('@')
     bg_color_name = bg_parts[0].lower()
     bg_opacity = float(bg_parts[1]) if len(bg_parts) > 1 else 0.5
-    bg_opacity = max(0, min(1, bg_opacity))
-    bg_color_rgb = LOCAL_COLOR_MAP.get(bg_color_name, "000000")
-    bg_color_bgr = bg_color_rgb[4:6] + bg_color_rgb[2:4] + bg_color_rgb[0:2]
-    bg_alpha = int((1 - bg_opacity) * 255)
-    print(f"🎨 BG color (rectangle): {bg_color_name} -> RGB={bg_color_rgb} -> BGR={bg_color_bgr}, Opacity={bg_opacity}, Alpha={bg_alpha}")
     
-    # Map position to ASS alignment
-    pos_map = {
-        "bottom-left": 1, "bottom-center": 2, "bottom-right": 3,
-        "left": 4, "center": 5, "right": 6,
-        "top-left": 7, "top-center": 8, "top-right": 9
+    # Map position to drawtext position with margin
+    margin = text_style.margin
+    drawtext_pos = {
+        "bottom-left": f"x={margin}:y=h-th-{margin}",
+        "bottom-center": f"x=(w-tw)/2:y=h-th-{margin}",
+        "bottom-right": f"x=w-tw-{margin}:y=h-th-{margin}",
+        "center": "x=(w-tw)/2:y=(h-th)/2",
+        "left": f"x={margin}:y=(h-th)/2",
+        "right": f"x=w-tw-{margin}:y=(h-th)/2",
+        "top-left": f"x={margin}:y={margin}",
+        "top-center": f"x=(w-tw)/2:y={margin}",
+        "top-right": f"x=w-tw-{margin}:y={margin}"
     }
-    alignment = pos_map.get(text_style.position, 5)
-    
-    # Calculate margins
-    margin_l = text_style.margin if alignment in [1,4,7] else 0
-    margin_r = text_style.margin if alignment in [3,6,9] else 0
-    margin_v = text_style.margin
-    
-    # Create fontconfig configuration
-    font_dir = os.path.dirname(font_path)
-    fc_config = f"""<?xml version="1.0"?>
-<!DOCTYPE fontconfig SYSTEM "fonts.dtd">
-<fontconfig>
-    <dir>{font_dir}</dir>
-</fontconfig>"""
+    position = drawtext_pos.get(text_style.position, "x=(w-tw)/2:y=(h-th)/2")
     
-    fc_file = os.path.join(work_dir, "fonts.conf")
-    with open(fc_file, 'w') as f:
-        f.write(fc_config)
-    os.environ['FONTCONFIG_FILE'] = fc_file
-    
-    # Create ASS file content with solid rectangle background and outline
-    # BorderStyle=3 gives solid box, Outline adds text border
-    # Format: &HAABBGGRR where AA is alpha, BB blue, GG green, RR red
-    ass_content = f"""[Script Info]
-; Script generated by Video Styling Space
-ScriptType: v4.00+
-PlayResX: 1920
-PlayResY: 1080
-ScaledBorderAndShadow: yes
-
-[V4+ Styles]
-Format: Name, Fontname, Fontsize, PrimaryColour, SecondaryColour, OutlineColour, BackColour, Bold, Italic, Underline, StrikeOut, ScaleX, ScaleY, Spacing, Angle, BorderStyle, Outline, Shadow, Alignment, MarginL, MarginR, MarginV, Encoding
-Style: Default,{font_family_name},{text_style.font_size},&H00{font_color_bgr},&H000000FF,&H00{outline_color_bgr},&H{bg_alpha:02X}{bg_color_bgr},0,0,0,0,100,100,0,0,3,{text_style.outline_width},0,{alignment},{margin_l},{margin_r},{margin_v},1
-
-[Events]
-Format: Layer, Start, End, Style, Name, MarginL, MarginR, MarginV, Effect, Text
-Dialogue: 0,0:00:00.00,0:00:10.00,Default,,0,0,0,,{text_style.text}"""
-    
-    # Write ASS file with UTF-8 encoding
-    with open(ass_file, 'w', encoding='utf-8') as f:
-        f.write(ass_content)
-    
-    # Debug: Print the ASS file content
-    print("\n=== ASS FILE DEBUG ===")
-    with open(ass_file, 'r', encoding='utf-8') as f:
-        print(f.read())
-    print("=== END ASS DEBUG ===\n")
-    
-    print(f"📝 Created ASS subtitle file with font family: {font_family_name}")
-    print(f"📝 Style line: BorderStyle=3, Outline={text_style.outline_width}, OutlineColor=&H00{outline_color_bgr}, BackColor=&H{bg_alpha:02X}{bg_color_bgr}")
-    
-    # Run FFmpeg with ASS filter
-    cmd = [
+    # Build drawtext command with background box and outline
+    # borderw = outline width, bordercolor = outline color
+    drawtext_cmd = [
         'ffmpeg', '-y',
         '-i', input_video,
-        '-vf', f"ass={ass_file}",
+        '-vf', f"drawtext=text='{text_style.text}':fontfile={font_path}:fontsize={text_style.font_size}:fontcolor={text_style.color}:{position}:box=1:boxcolor={bg_color_name}@{bg_opacity}:boxborderw={text_style.padding}:borderw={text_style.outline_width}:bordercolor={text_style.outline_color}",
         '-c:a', 'copy',
         output_video
     ]
     
-    print(f"🎬 Running FFmpeg with ASS filter...")
-    result = subprocess.run(cmd, capture_output=True, text=True)
+    print(f"🎬 Running drawtext command with outline={text_style.outline_width}, bordercolor={text_style.outline_color}")
+    result = subprocess.run(drawtext_cmd, capture_output=True, text=True)
     
     if result.returncode == 0:
-        print(f"✅ ASS method succeeded with background and outline")
+        print(f"✅ Drawtext method succeeded with background and outline")
         return True
     else:
-        print(f"❌ ASS method failed: {result.stderr}")
+        print(f"❌ Drawtext failed: {result.stderr}")
         return False
 
 # =============================================
@@ -1141,6 +983,7 @@ async def add_captions_to_video(request: CaptionRequest):
         
         current_video = video_path
         
+        # Get font path
         font_path = get_font_path(request.caption_style.font_family)
         if not font_path:
             return CaptionResponse(
@@ -1149,58 +992,77 @@ async def add_captions_to_video(request: CaptionRequest):
                 error=f"Caption font not found: {request.caption_style.font_family}"
             )
         
-        font_dir = os.path.dirname(font_path)
-        fc_config = f"""<?xml version="1.0"?>
-<!DOCTYPE fontconfig SYSTEM "fonts.dtd">
-<fontconfig>
-    <dir>{font_dir}</dir>
-</fontconfig>"""
-        
-        fc_file = os.path.join(work_dir, "fonts.conf")
-        with open(fc_file, 'w') as f:
-            f.write(fc_config)
-        os.environ['FONTCONFIG_FILE'] = fc_file
-        
-        ass_file = create_caption_ass(
-            request.captions, 
-            request.caption_style, 
-            work_dir, 
-            font_path
-        )
-        
-        captioned_path = os.path.join(work_dir, "captioned.mp4")
-        
-        cmd = [
-            'ffmpeg', '-y',
-            '-i', current_video,
-            '-vf', f"ass={ass_file}",
-            '-c:a', 'copy',
-            captioned_path
-        ]
+        # Get margin value
+        margin = request.caption_style.margin
+        
+        # Map position to drawtext position with margin
+        drawtext_pos = {
+            "bottom-left": f"x={margin}:y=h-th-{margin}",
+            "bottom-center": f"x=(w-tw)/2:y=h-th-{margin}", 
+            "bottom-right": f"x=w-tw-{margin}:y=h-th-{margin}",
+            "center": "x=(w-tw)/2:y=(h-th)/2",
+            "left": f"x={margin}:y=(h-th)/2",
+            "right": f"x=w-tw-{margin}:y=(h-th)/2",
+            "top-left": f"x={margin}:y={margin}",
+            "top-center": f"x=(w-tw)/2:y={margin}",
+            "top-right": f"x=w-tw-{margin}:y={margin}"
+        }
+        position = drawtext_pos.get(request.caption_style.position, f"x=(w-tw)/2:y=h-th-{margin}")
         
-        print(f"🎬 Applying captions with FFmpeg...")
-        result = subprocess.run(cmd, capture_output=True, text=True)
+        # Process each caption with its own drawtext filter
+        current_input = video_path
         
-        if result.returncode != 0:
-            print(f"❌ FFmpeg error: {result.stderr}")
-            return CaptionResponse(
-                status="error",
-                project_id=request.project_id,
-                error=f"Failed to add captions: {result.stderr[:200]}"
+        for i, caption in enumerate(request.captions):
+            temp_output = os.path.join(work_dir, f"step_{i:03d}.mp4")
+            
+            # Build drawtext filter with background box and outline
+            drawtext_filter = (
+                f"drawtext=text='{caption.text}':"
+                f"fontfile={font_path}:"
+                f"fontsize={request.caption_style.font_size}:"
+                f"fontcolor={request.caption_style.color}:"
+                f"{position}:"
+                f"box=1:"
+                f"boxcolor={request.caption_style.bg_color}:"
+                f"boxborderw={request.caption_style.padding}:"
+                f"borderw={request.caption_style.outline_width}:"
+                f"bordercolor={request.caption_style.outline_color}:"
+                f"enable='between(t,{caption.start_time},{caption.end_time})'"
             )
+            
+            cmd = [
+                'ffmpeg', '-y',
+                '-i', current_input,
+                '-vf', drawtext_filter,
+                '-c:a', 'copy',
+                temp_output
+            ]
+            
+            print(f"🎬 Adding caption {i+1}/{len(request.captions)}: '{caption.text}' with outline={request.caption_style.outline_width}")
+            result = subprocess.run(cmd, capture_output=True, text=True)
+            
+            if result.returncode != 0:
+                print(f"❌ FFmpeg error at caption {i+1}: {result.stderr}")
+                return CaptionResponse(
+                    status="error",
+                    project_id=request.project_id,
+                    error=f"Failed to add caption {i+1}: {result.stderr[:200]}"
+                )
+            
+            current_input = temp_output
         
-        current_video = captioned_path
+        final_video = current_input
         
         if request.title_overlay:
             titled_path = os.path.join(work_dir, "titled.mp4")
-            if create_text_overlay(current_video, titled_path, request.title_overlay):
-                current_video = titled_path
+            if create_text_overlay(final_video, titled_path, request.title_overlay):
+                final_video = titled_path
             else:
                 print("⚠️ Title overlay failed, continuing with captioned video")
         
         timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
         final_filename = f"captioned_{timestamp}.mp4"
-        final_url = upload_to_dataset(current_video, request.project_id, final_filename, "videos")
+        final_url = upload_to_dataset(final_video, request.project_id, final_filename, "videos")
         
         shutil.rmtree(work_dir, ignore_errors=True)
         
@@ -1274,6 +1136,7 @@ async def transcribe_and_caption(request: TranscriptionRequest):
                 error="No captions generated from audio"
             )
         
+        # Get font path
         font_path = get_font_path(request.caption_style.font_family)
         if not font_path:
             return TranscriptionResponse(
@@ -1282,56 +1145,75 @@ async def transcribe_and_caption(request: TranscriptionRequest):
                 error=f"Caption font not found: {request.caption_style.font_family}"
             )
         
-        font_dir = os.path.dirname(font_path)
-        fc_config = f"""<?xml version="1.0"?>
-<!DOCTYPE fontconfig SYSTEM "fonts.dtd">
-<fontconfig>
-    <dir>{font_dir}</dir>
-</fontconfig>"""
-        
-        fc_file = os.path.join(work_dir, "fonts.conf")
-        with open(fc_file, 'w') as f:
-            f.write(fc_config)
-        os.environ['FONTCONFIG_FILE'] = fc_file
-        
-        ass_file = create_caption_ass(
-            captions,
-            request.caption_style,
-            work_dir,
-            font_path
-        )
-        
-        captioned_path = os.path.join(work_dir, "captioned.mp4")
-        
-        cmd = [
-            'ffmpeg', '-y',
-            '-i', video_path,
-            '-vf', f"ass={ass_file}",
-            '-c:a', 'copy',
-            captioned_path
-        ]
+        # Get margin value
+        margin = request.caption_style.margin
+        
+        # Map position to drawtext position with margin
+        drawtext_pos = {
+            "bottom-left": f"x={margin}:y=h-th-{margin}",
+            "bottom-center": f"x=(w-tw)/2:y=h-th-{margin}", 
+            "bottom-right": f"x=w-tw-{margin}:y=h-th-{margin}",
+            "center": "x=(w-tw)/2:y=(h-th)/2",
+            "left": f"x={margin}:y=(h-th)/2",
+            "right": f"x=w-tw-{margin}:y=(h-th)/2",
+            "top-left": f"x={margin}:y={margin}",
+            "top-center": f"x=(w-tw)/2:y={margin}",
+            "top-right": f"x=w-tw-{margin}:y={margin}"
+        }
+        position = drawtext_pos.get(request.caption_style.position, f"x=(w-tw)/2:y=h-th-{margin}")
         
-        print(f"🎬 Applying captions...")
-        result = subprocess.run(cmd, capture_output=True, text=True)
+        # Process each caption with its own drawtext filter
+        current_input = video_path
         
-        if result.returncode != 0:
-            print(f"❌ FFmpeg error: {result.stderr}")
-            return TranscriptionResponse(
-                status="error",
-                project_id=request.project_id,
-                error=f"Failed to add captions: {result.stderr[:200]}"
+        for i, caption in enumerate(captions):
+            temp_output = os.path.join(work_dir, f"step_{i:03d}.mp4")
+            
+            # Build drawtext filter with background box and outline
+            drawtext_filter = (
+                f"drawtext=text='{caption.text}':"
+                f"fontfile={font_path}:"
+                f"fontsize={request.caption_style.font_size}:"
+                f"fontcolor={request.caption_style.color}:"
+                f"{position}:"
+                f"box=1:"
+                f"boxcolor={request.caption_style.bg_color}:"
+                f"boxborderw={request.caption_style.padding}:"
+                f"borderw={request.caption_style.outline_width}:"
+                f"bordercolor={request.caption_style.outline_color}:"
+                f"enable='between(t,{caption.start_time},{caption.end_time})'"
             )
+            
+            cmd = [
+                'ffmpeg', '-y',
+                '-i', current_input,
+                '-vf', drawtext_filter,
+                '-c:a', 'copy',
+                temp_output
+            ]
+            
+            print(f"🎬 Adding caption {i+1}/{len(captions)}: '{caption.text}'")
+            result = subprocess.run(cmd, capture_output=True, text=True)
+            
+            if result.returncode != 0:
+                print(f"❌ FFmpeg error at caption {i+1}: {result.stderr}")
+                return TranscriptionResponse(
+                    status="error",
+                    project_id=request.project_id,
+                    error=f"Failed to add caption {i+1}: {result.stderr[:200]}"
+                )
+            
+            current_input = temp_output
         
-        current_video = captioned_path
+        final_video = current_input
         
         if request.title_overlay:
             titled_path = os.path.join(work_dir, "titled.mp4")
-            if create_text_overlay(current_video, titled_path, request.title_overlay):
-                current_video = titled_path
+            if create_text_overlay(final_video, titled_path, request.title_overlay):
+                final_video = titled_path
         
         timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
         final_filename = f"transcribed_{timestamp}.mp4"
-        final_url = upload_to_dataset(current_video, request.project_id, final_filename, "videos")
+        final_url = upload_to_dataset(final_video, request.project_id, final_filename, "videos")
         
         transcript = " ".join([c.text for c in captions])
         
@@ -1409,58 +1291,79 @@ async def add_captions_from_srt(request: SrtCaptionRequest):
                 error=f"Caption font not found: {request.caption_style.font_family}"
             )
         
-        font_dir = os.path.dirname(font_path)
-        fc_config = f"""<?xml version="1.0"?>
-<!DOCTYPE fontconfig SYSTEM "fonts.dtd">
-<fontconfig>
-    <dir>{font_dir}</dir>
-</fontconfig>"""
-        
-        fc_file = os.path.join(work_dir, "fonts.conf")
-        with open(fc_file, 'w') as f:
-            f.write(fc_config)
-        os.environ['FONTCONFIG_FILE'] = fc_file
-        
-        ass_file = create_caption_ass(
-            captions,
-            request.caption_style,
-            work_dir,
-            font_path
-        )
-        
-        captioned_path = os.path.join(work_dir, "captioned.mp4")
-        
-        cmd = [
-            'ffmpeg', '-y',
-            '-i', video_path,
-            '-vf', f"ass={ass_file}",
-            '-c:a', 'copy',
-            captioned_path
-        ]
+        # Get margin value
+        margin = request.caption_style.margin
+        print(f"📏 Using margin: {margin} pixels")
+        
+        # Map position to drawtext position with margin
+        drawtext_pos = {
+            "bottom-left": f"x={margin}:y=h-th-{margin}",
+            "bottom-center": f"x=(w-tw)/2:y=h-th-{margin}", 
+            "bottom-right": f"x=w-tw-{margin}:y=h-th-{margin}",
+            "center": "x=(w-tw)/2:y=(h-th)/2",
+            "left": f"x={margin}:y=(h-th)/2",
+            "right": f"x=w-tw-{margin}:y=(h-th)/2",
+            "top-left": f"x={margin}:y={margin}",
+            "top-center": f"x=(w-tw)/2:y={margin}",
+            "top-right": f"x=w-tw-{margin}:y={margin}"
+        }
+        position = drawtext_pos.get(request.caption_style.position, f"x=(w-tw)/2:y=h-th-{margin}")
+        print(f"📍 Position string: {position}")
         
-        print(f"🎬 Applying captions from SRT...")
-        result = subprocess.run(cmd, capture_output=True, text=True)
+        # Process each caption with its own drawtext filter
+        current_input = video_path
         
-        if result.returncode != 0:
-            print(f"❌ FFmpeg error: {result.stderr}")
-            return CaptionResponse(
-                status="error",
-                project_id=request.project_id,
-                error=f"Failed to add captions: {result.stderr[:200]}"
+        for i, caption in enumerate(captions):
+            temp_output = os.path.join(work_dir, f"step_{i:03d}.mp4")
+            
+            # Build drawtext filter with background box and outline
+            drawtext_filter = (
+                f"drawtext=text='{caption.text}':"
+                f"fontfile={font_path}:"
+                f"fontsize={request.caption_style.font_size}:"
+                f"fontcolor={request.caption_style.color}:"
+                f"{position}:"
+                f"box=1:"
+                f"boxcolor={request.caption_style.bg_color}:"
+                f"boxborderw={request.caption_style.padding}:"
+                f"borderw={request.caption_style.outline_width}:"
+                f"bordercolor={request.caption_style.outline_color}:"
+                f"enable='between(t,{caption.start_time},{caption.end_time})'"
             )
+            
+            cmd = [
+                'ffmpeg', '-y',
+                '-i', current_input,
+                '-vf', drawtext_filter,
+                '-c:a', 'copy',
+                temp_output
+            ]
+            
+            print(f"🎬 Adding caption {i+1}/{len(captions)}: '{caption.text}' at margin {margin} with outline={request.caption_style.outline_width}")
+            result = subprocess.run(cmd, capture_output=True, text=True)
+            
+            if result.returncode != 0:
+                print(f"❌ FFmpeg error at caption {i+1}: {result.stderr}")
+                return CaptionResponse(
+                    status="error",
+                    project_id=request.project_id,
+                    error=f"Failed to add caption {i+1}: {result.stderr[:200]}"
+                )
+            
+            current_input = temp_output
         
-        current_video = captioned_path
+        final_video = current_input
         
         if request.title_overlay:
             titled_path = os.path.join(work_dir, "titled.mp4")
-            if create_text_overlay(current_video, titled_path, request.title_overlay):
-                current_video = titled_path
+            if create_text_overlay(final_video, titled_path, request.title_overlay):
+                final_video = titled_path
             else:
                 print("⚠️ Title overlay failed, continuing with captioned video")
         
         timestamp = datetime.now().strftime("%Y%m%d_%H%M%S")
         final_filename = f"captioned_from_srt_{timestamp}.mp4"
-        final_url = upload_to_dataset(current_video, request.project_id, final_filename, "videos")
+        final_url = upload_to_dataset(final_video, request.project_id, final_filename, "videos")
         
         shutil.rmtree(work_dir, ignore_errors=True)
         
@@ -1492,7 +1395,7 @@ async def add_captions_from_srt(request: SrtCaptionRequest):
 async def root():
     return {
         "name": "Text Styling API with Auto Caption",
-        "version": "3.6.0",
+        "version": "4.0.0",
         "features": {
             "title_overlay": "✅",
             "manual_captions": "✅",
@@ -1500,7 +1403,8 @@ async def root():
             "srt_support": "✅",
             "enhanced_colors": f"✅ ({len(COLOR_MAP)} colors)",
             "text_outline": "✅",
-            "background_boxes": "✅"
+            "background_boxes": "✅",
+            "margin_control": "✅"
         },
         "endpoints": {
             "style": "POST /api/style - Add title overlay",