Fix comment extractor args separator (, → ;) and use comment_sort=top

yt-dlp separates extractor args with ; not ,. The malformed arg was
causing max_comments to parse as a garbage string, fetching ~1 comment.
Also swap max_comment_depth (not a real YouTube extractor arg) for
comment_sort=top to get highest-engagement comments first.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

backend/database.py

@@ -84,6 +84,7 @@ def init_db():
     try:
         # Column migrations — safe to run on every startup
         _add_column_if_missing(raw_conn, "videos", "view_count", "INTEGER")
+        _add_column_if_missing(raw_conn, "videos", "like_count", "INTEGER")
         raw_conn.commit()
         # executescript handles multi-statement SQL including trigger BEGIN...END blocks
         raw_conn.executescript(FTS_SETUP_SQL)

backend/models.py

@@ -61,6 +61,7 @@ class Video(Base):
     category = Column(String)
     chapters = Column(Text)  # JSON array of {start_time, end_time, title}
     view_count = Column(Integer)
+    like_count = Column(Integer)
 
 
 class UserVideo(Base):

backend/routers/videos.py

@@ -66,6 +66,7 @@ class VideoDetail(BaseModel):
     local_file_url: Optional[str] = None
     is_recommended: bool = False
     view_count: Optional[int] = None
+    like_count: Optional[int] = None
 
     model_config = {"from_attributes": True}
 
@@ -422,7 +423,7 @@ def surprise_me(
 
 _VIDEO_SELECT = """
     SELECT v.id, v.youtube_video_id, v.title, v.description, v.thumbnail_url,
-           v.duration_seconds, v.published_at, v.tags, v.category, v.view_count,
+           v.duration_seconds, v.published_at, v.tags, v.category, v.view_count, v.like_count,
            c.id AS channel_id, c.name AS channel_name, c.youtube_channel_id AS channel_youtube_id,
            COALESCE(uv.watched, 0) AS watched,
            COALESCE(uv.watch_progress_seconds, 0) AS watch_progress_seconds,

backend/services/ytdlp.py

@@ -75,6 +75,7 @@ def _normalize_video(info: dict) -> dict:
         "category": info.get("category") or (info.get("categories") or [None])[0],
         "chapters": json.dumps(chapters) if chapters else None,
         "view_count": info.get("view_count"),
+        "like_count": info.get("like_count"),
         "channel": {
             "youtube_channel_id": info.get("channel_id"),
             "name": info.get("channel") or info.get("uploader", ""),
@@ -390,7 +391,7 @@ def fetch_video_comments(youtube_video_id: str, max_comments: int = 20) -> list[
         "yt-dlp", url,
         "--dump-json",
         "--write-comments",
-        "--extractor-args", f"youtube:max_comments={max_comments},max_comment_depth=1",
+        "--extractor-args", f"youtube:max_comments={max_comments};comment_sort=top",
         "--no-download",
         "--no-playlist",
         "--quiet",

frontend/src/pages/Watch.jsx

@@ -894,6 +894,7 @@ export default function Watch() {
             <div className="flex items-center gap-2 text-sm text-zinc-500 flex-wrap">
               {date && <span>{date}</span>}
               {video?.view_count > 0 && <><span>·</span><span>{formatViews(video.view_count)}</span></>}
+              {video?.like_count > 0 && <><span>·</span><span>👍 {formatViews(video.like_count).replace(" views", "")}</span></>}
               {video?.category && <><span>·</span><span>{video.category}</span></>}
               {video?.duration_seconds && (
                 <><span>·</span>