added video support

2026-06-22 23:36:20 +02:00
parent 2d93109831
commit 544d6c866c
7 changed files with 178 additions and 5 deletions
@@ -0,0 +1,128 @@
+from __future__ import annotations
+
+import re
+import subprocess
+import sys
+from pathlib import Path
+from urllib.parse import unquote
+
+
+VIDEO_TAG_RE = re.compile(
+    r"<video(?P<attrs>[^>]*)>(?P<body>.*?)</video>",
+    re.IGNORECASE | re.DOTALL,
+    )
+
+SOURCE_RE = re.compile(
+    r'<source[^>]+src=["\'](?P<src>[^"\']+)["\'][^>]*>',
+    re.IGNORECASE,
+)
+
+
+def extract_src(video_attrs: str, video_body: str) -> str | None:
+    """Extract video path from either <video src=""> or nested <source src="">."""
+
+    video_src = re.search(r'src=["\']([^"\']+)["\']', video_attrs, re.IGNORECASE)
+    if video_src:
+        return video_src.group(1)
+
+    source_src = SOURCE_RE.search(video_body)
+    if source_src:
+        return source_src.group("src")
+
+    return None
+
+
+def make_poster(video_path: Path, poster_path: Path) -> None:
+    """Extract a poster frame from a video using ffmpeg."""
+
+    poster_path.parent.mkdir(parents=True, exist_ok=True)
+
+    subprocess.run(
+        [
+            "ffmpeg",
+            "-y",
+            "-ss",
+            "00:00:01",
+            "-i",
+            str(video_path),
+            "-frames:v",
+            "1",
+            "-q:v",
+            "3",
+            str(poster_path),
+        ],
+        check=True,
+        stdout=subprocess.DEVNULL,
+        stderr=subprocess.DEVNULL,
+    )
+
+
+def process_file(html_file: Path, root: Path) -> None:
+    """Replace video preload mode and add poster attributes."""
+
+    content = html_file.read_text(encoding="utf-8")
+    changed = False
+
+    def replace_video(match: re.Match[str]) -> str:
+        nonlocal changed
+
+        attrs = match.group("attrs")
+        body = match.group("body")
+
+        if 'preload="auto"' not in attrs and "preload='auto'" not in attrs:
+            return match.group(0)
+
+        src = extract_src(attrs, body)
+        if not src:
+            return match.group(0)
+
+        src_clean = unquote(src.split("#")[0].split("?")[0])
+
+        if src_clean.startswith(("http://", "https://", "//")):
+            return match.group(0)
+
+        video_path = (html_file.parent / src_clean).resolve()
+
+        try:
+            video_path.relative_to(root.resolve())
+        except ValueError:
+            return match.group(0)
+
+        if not video_path.exists():
+            print(f"Missing video: {video_path}")
+            return match.group(0)
+
+        poster_path = video_path.with_suffix(".jpg")
+        poster_src = src_clean.rsplit(".", 1)[0] + ".jpg"
+
+        if not poster_path.exists():
+            make_poster(video_path, poster_path)
+
+        attrs_new = attrs
+        attrs_new = attrs_new.replace('preload="auto"', 'preload="none"')
+        attrs_new = attrs_new.replace("preload='auto'", 'preload="none"')
+
+        if "poster=" not in attrs_new.lower():
+            attrs_new += f' poster="{poster_src}"'
+
+        changed = True
+        return f"<video{attrs_new}>{body}</video>"
+
+    new_content = VIDEO_TAG_RE.sub(replace_video, content)
+
+    if changed:
+        html_file.write_text(new_content, encoding="utf-8")
+        print(f"Updated: {html_file}")
+
+
+def main() -> None:
+    """Process all HTML files below the given root directory."""
+
+    root = Path(sys.argv[1]).resolve()
+
+    for html_file in root.rglob("*.html"):
+        process_file(html_file, root)
+
+
+if __name__ == "__main__":
+    main()