Migrate to Pages: public/meme assets, gallery and scripts; remove root meme/; tooling defaults

Made-with: Cursor

animated_webp_to_gif.py

@@ -0,0 +1,212 @@
+#!/usr/bin/env python3
+"""
+Recursively find *.webp under 原图/ (by default). Animated WebP -> same-stem .gif
+(then remove the .webp). Static WebP unchanged. Default 10 worker threads.
+"""
+from __future__ import annotations
+
+import argparse
+import io
+import re
+import sys
+import tempfile
+import threading
+from concurrent.futures import ThreadPoolExecutor
+from pathlib import Path
+
+try:
+    from PIL import Image, ImageSequence
+except ImportError as exc:  # pragma: no cover
+    raise SystemExit(
+        "This script requires Pillow. Install it with: python3 -m pip install pillow"
+    ) from exc
+
+
+# 只处理已编号的成品动图（避免误处理未编号的素材）
+OUTPUT_NAME_RE = re.compile(r"^(\d+)\.webp$", re.IGNORECASE)
+
+DEFAULT_JOBS = 10
+_log_lock = threading.Lock()
+
+
+def _log(*args, file=sys.stdout, **kwargs) -> None:
+    with _log_lock:
+        print(*args, file=file, **kwargs)
+
+
+def convert_one(path: Path, root: Path, dry_run: bool) -> str:
+    try:
+        if not is_animated_webp(path):
+            return "skip_static"
+    except OSError as e:
+        _log(f"[skip] {path}: {e}", file=sys.stderr)
+        return "error"
+    gif_path = path.with_suffix(".gif")
+    rel = path.relative_to(root)
+    if dry_run:
+        _log(f"would convert: {rel} -> {gif_path.name}")
+        return "ok"
+    try:
+        webp_to_gif(path, gif_path)
+        path.unlink()
+        _log(f"{rel} -> {gif_path.name} (animated)")
+        return "ok"
+    except Exception as e:
+        _log(f"[error] {path}: {e}", file=sys.stderr)
+        if gif_path.exists():
+            try:
+                gif_path.unlink()
+            except OSError:
+                pass
+        return "error"
+
+
+def is_animated_webp_image(im: Image.Image) -> bool:
+    if im.format != "WEBP":
+        return False
+    return bool(
+        getattr(im, "is_animated", False) or getattr(im, "n_frames", 1) > 1
+    )
+
+
+def is_animated_webp(path: Path) -> bool:
+    with Image.open(path) as im:
+        return is_animated_webp_image(im)
+
+
+def save_animated_webp_image_to_gif(im: Image.Image, dst: Path) -> None:
+    if im.format != "WEBP":
+        raise ValueError("expected WebP image")
+    loop = im.info.get("loop", 0)
+    frames: list[Image.Image] = []
+    durations: list[int] = []
+    for frame in ImageSequence.Iterator(im):
+        rgba = frame.convert("RGBA")
+        frames.append(rgba)
+        d = frame.info.get(
+            "duration", im.info.get("duration", im.info.get("gif_duration", 100))
+        )
+        try:
+            durations.append(int(d) if d is not None else 100)
+        except (TypeError, ValueError):
+            durations.append(100)
+
+    if len(frames) == 0:
+        raise ValueError("no frames")
+    if len(durations) < len(frames):
+        durations.extend([durations[-1]] * (len(frames) - len(durations)))
+
+    dst.parent.mkdir(parents=True, exist_ok=True)
+    with tempfile.NamedTemporaryFile(
+        suffix=".gif", delete=False, dir=dst.parent
+    ) as tmp:
+        tmp_path = Path(tmp.name)
+    try:
+        if len(frames) == 1:
+            frames[0].save(
+                tmp_path,
+                format="GIF",
+                duration=durations[0],
+                loop=loop,
+            )
+        else:
+            frames[0].save(
+                tmp_path,
+                format="GIF",
+                save_all=True,
+                append_images=frames[1:],
+                duration=durations,
+                loop=loop,
+                disposal=2,
+            )
+        tmp_path.replace(dst)
+    finally:
+        if tmp_path.exists():
+            try:
+                tmp_path.unlink()
+            except OSError:
+                pass
+
+
+def webp_bytes_to_gif(data: bytes, dst: Path) -> None:
+    with Image.open(io.BytesIO(data)) as im:
+        save_animated_webp_image_to_gif(im, dst)
+
+
+def webp_to_gif(src: Path, dst: Path) -> None:
+    with Image.open(src) as im:
+        save_animated_webp_image_to_gif(im, dst)
+
+
+def main(argv: list[str]) -> int:
+    parser = argparse.ArgumentParser(
+        description=(
+            "Convert animated WebP under 原图/ to GIF (in place); "
+            "static WebP unchanged. Only numbered N.webp files are considered."
+        )
+    )
+    parser.add_argument(
+        "root",
+        nargs="?",
+        default="原图",
+        type=Path,
+        help="Root directory to scan (default: 原图).",
+    )
+    parser.add_argument(
+        "--dry-run",
+        action="store_true",
+        help="Only print what would be converted.",
+    )
+    parser.add_argument(
+        "--all-webp",
+        action="store_true",
+        help="Process any *.webp, not only numbered N.webp.",
+    )
+    parser.add_argument(
+        "-j",
+        "--jobs",
+        type=int,
+        default=DEFAULT_JOBS,
+        metavar="N",
+        help=f"Parallel worker threads (default: {DEFAULT_JOBS}).",
+    )
+    args = parser.parse_args(argv)
+    root = args.root.expanduser().resolve()
+    if not root.is_dir():
+        print(f"Not a directory: {root}", file=sys.stderr)
+        return 1
+
+    webp_files = sorted(root.rglob("*.webp"))
+    converted = 0
+    skipped_static = 0
+    skipped_name = 0
+    errors = 0
+
+    candidates: list[Path] = []
+    for path in webp_files:
+        if not args.all_webp and not OUTPUT_NAME_RE.match(path.name):
+            skipped_name += 1
+            continue
+        candidates.append(path)
+
+    workers = max(1, args.jobs)
+    with ThreadPoolExecutor(max_workers=workers) as pool:
+        for outcome in pool.map(
+            lambda p: convert_one(p, root, args.dry_run), candidates
+        ):
+            if outcome == "ok":
+                converted += 1
+            elif outcome == "skip_static":
+                skipped_static += 1
+            else:
+                errors += 1
+
+    print(
+        f"Done. converted={converted}, skipped_static_webp={skipped_static}, "
+        f"skipped_unnumbered={skipped_name}, errors={errors} ({workers} workers)"
+    )
+    return 0 if errors == 0 else 1
+
+
+if __name__ == "__main__":
+    raise SystemExit(main(sys.argv[1:]))

convert_to_webp.py

@@ -3,7 +3,7 @@ from __future__ import annotations
 
 import argparse
 import io
-import os
+import re
 import sys
 import zipfile
 from concurrent.futures import ThreadPoolExecutor
@@ -18,6 +18,7 @@ except ImportError as exc:  # pragma: no cover
         "This script requires Pillow. Install it with: python3 -m pip install pillow"
     ) from exc
 
+from animated_webp_to_gif import webp_bytes_to_gif
 
 IMAGE_EXTENSIONS = {
     ".avif",
@@ -30,20 +31,46 @@ IMAGE_EXTENSIONS = {
     ".png",
     ".tif",
     ".tiff",
+    ".webp",
 }
 
+# 已编号的成品：1.webp、42.gif（不参与转换；用于各子文件夹内续号）
+OUTPUT_NAME_RE = re.compile(r"^(\d+)\.(webp|gif)$", re.IGNORECASE)
+
+DEFAULT_SOURCE_DIR = Path("原图")
+
 
 @dataclass(frozen=True)
 class Job:
     source_label: str
-    target_path: Path
+    target_dir: Path
+    index: int
     loader: Callable[[], bytes]
+    # 磁盘上的源文件；转换成功后删除（zip 成员为 None）。与输出路径相同时不会删。
+    source_file: Path | None = None
 
 
 def is_image_path(path: Path) -> bool:
     return path.is_file() and path.suffix.lower() in IMAGE_EXTENSIONS
 
 
+def is_numbered_asset_file(path: Path) -> bool:
+    return path.is_file() and bool(OUTPUT_NAME_RE.match(path.name))
+
+
+def max_index_in_directory(directory: Path) -> int:
+    highest = 0
+    if not directory.is_dir():
+        return 0
+    for p in directory.iterdir():
+        if not p.is_file():
+            continue
+        m = OUTPUT_NAME_RE.match(p.name)
+        if m:
+            highest = max(highest, int(m.group(1)))
+    return highest
+
+
 def convert_to_webp(data: bytes) -> tuple[bytes, bool]:
     with Image.open(io.BytesIO(data)) as im:
         is_animated = bool(
@@ -76,137 +103,237 @@ def convert_to_webp(data: bytes) -> tuple[bytes, bool]:
         return output.getvalue(), False
 
 
+def is_animated_gif(data: bytes) -> bool:
+    with Image.open(io.BytesIO(data)) as im:
+        if im.format != "GIF":
+            return False
+        n_frames = getattr(im, "n_frames", 1)
+        return bool(getattr(im, "is_animated", False) or n_frames > 1)
+
+
+def is_animated_webp_bytes(data: bytes) -> bool:
+    with Image.open(io.BytesIO(data)) as im:
+        if im.format != "WEBP":
+            return False
+        return bool(
+            getattr(im, "is_animated", False) or getattr(im, "n_frames", 1) > 1
+        )
+
+
 def _read_zip_member(zip_path: Path, member_name: str) -> bytes:
     with zipfile.ZipFile(zip_path) as zf:
         return zf.read(member_name)
 
 
-def collect_directory_jobs(root: Path, output_base: Path) -> list[Job]:
+def collect_directory_jobs(
+    source_root: Path, output_root: Path, *, delete_sources: bool
+) -> list[Job]:
     grouped: dict[Path, list[Path]] = {}
-    for path in sorted(root.rglob("*")):
+    for path in sorted(source_root.rglob("*")):
         if not is_image_path(path):
             continue
+        if is_numbered_asset_file(path):
+            continue
         grouped.setdefault(path.parent, []).append(path)
 
     jobs: list[Job] = []
-    for directory in sorted(grouped):
+    for directory in sorted(grouped, key=lambda p: str(p)):
         files = sorted(grouped[directory], key=lambda p: p.name)
-        relative_dir = directory.relative_to(root)
-        target_dir = output_base / relative_dir
-        for index, path in enumerate(files, start=1):
+        target_dir = output_root / directory.relative_to(source_root)
+        next_i = max_index_in_directory(target_dir) + 1
+        for path in files:
             jobs.append(
                 Job(
-                    source_label=path.relative_to(root).as_posix(),
-                    target_path=target_dir / f"{index}.webp",
+                    source_label=path.relative_to(source_root).as_posix(),
+                    target_dir=target_dir,
+                    index=next_i,
                     loader=lambda p=path: p.read_bytes(),
+                    source_file=path if delete_sources else None,
                 )
             )
+            next_i += 1
     return jobs
 
 
-def collect_zip_jobs(zip_path: Path, output_base: Path) -> list[Job]:
+def collect_zip_jobs(zip_path: Path, output_root: Path) -> list[Job]:
     grouped: dict[PurePosixPath, list[PurePosixPath]] = {}
     with zipfile.ZipFile(zip_path) as zf:
         for info in zf.infolist():
             if info.is_dir():
                 continue
             member_path = PurePosixPath(info.filename)
-            if member_path.suffix.lower() not in IMAGE_EXTENSIONS:
+            suf = member_path.suffix.lower()
+            if suf not in IMAGE_EXTENSIONS:
+                continue
+            if OUTPUT_NAME_RE.match(member_path.name):
                 continue
             grouped.setdefault(member_path.parent, []).append(member_path)
 
     jobs: list[Job] = []
     for directory in sorted(grouped, key=lambda p: str(p)):
         files = sorted(grouped[directory], key=lambda p: str(p.name))
-        for index, member_path in enumerate(files, start=1):
-            target_dir = output_base / Path(str(directory))
+        target_dir = output_root / Path(str(directory))
+        next_i = max_index_in_directory(target_dir) + 1
+        for member_path in files:
             jobs.append(
                 Job(
                     source_label=member_path.as_posix(),
-                    target_path=target_dir / f"{index}.webp",
+                    target_dir=target_dir,
+                    index=next_i,
                     loader=lambda name=member_path.as_posix(), zp=zip_path: _read_zip_member(
                         zp, name
                     ),
                 )
             )
+            next_i += 1
     return jobs
 
 
-def collect_jobs(source: Path, output_base: Path | None) -> list[Job]:
-    if source.is_dir():
-        base = output_base if output_base is not None else source.parent / f"{source.name}-out"
-        base.mkdir(parents=True, exist_ok=True)
-        return collect_directory_jobs(source, base)
-
-    if source.is_file() and source.suffix.lower() == ".zip":
-        base = output_base if output_base is not None else source.parent / source.stem
-        base.mkdir(parents=True, exist_ok=True)
-        return collect_zip_jobs(source, base)
-
-    if is_image_path(source):
-        return [
-            Job(
-                source_label=source.name,
-                target_path=source.parent / "1.webp",
-                loader=lambda p=source: p.read_bytes(),
+def _remove_source_if_any(job: Job, out: Path, kind: str) -> str:
+    if job.source_file is None:
+        return kind
+    try:
+        src = job.source_file.resolve()
+        if src == out.resolve():
+            return kind
+        src.unlink()
+        return f"{kind}; removed-source"
+    except OSError as exc:
+        print(
+            f"[warn] ok: {out} ({kind}) but could not delete {job.source_file}: {exc}",
+            file=sys.stderr,
         )
-        ]
-
-    return []
+        return kind
 
 
-def run_job(job: Job) -> tuple[bool, str, Path, bool | None, str | None]:
+def run_job(job: Job) -> tuple[bool, str, Path, str | None, str | None]:
     try:
         data = job.loader()
+        job.target_dir.mkdir(parents=True, exist_ok=True)
+
+        if is_animated_gif(data):
+            out = job.target_dir / f"{job.index}.gif"
+            out.write_bytes(data)
+            kind = _remove_source_if_any(job, out, "gif-animated")
+            return True, job.source_label, out, kind, None
+
+        if is_animated_webp_bytes(data):
+            out = job.target_dir / f"{job.index}.gif"
+            webp_bytes_to_gif(data, out)
+            kind = _remove_source_if_any(job, out, "animated-webp->gif")
+            return True, job.source_label, out, kind, None
+
         webp_data, animated = convert_to_webp(data)
-        job.target_path.parent.mkdir(parents=True, exist_ok=True)
-        job.target_path.write_bytes(webp_data)
-        return True, job.source_label, job.target_path, animated, None
+        out = job.target_dir / f"{job.index}.webp"
+        out.write_bytes(webp_data)
+        base = "animated-webp" if animated else "static-webp"
+        kind = _remove_source_if_any(job, out, base)
+        return True, job.source_label, out, kind, None
     except Exception as exc:
-        return False, job.source_label, job.target_path, None, str(exc)
+        return False, job.source_label, job.target_dir / f"{job.index}.?", None, str(exc)
 
 
 def main(argv: list[str]) -> int:
     parser = argparse.ArgumentParser(
-        description="Convert images to WebP in place, with independent numbering per folder."
+        description=(
+            "Read unnumbered images from 原图/ (by default), write numbered *.webp / *.gif "
+            "into the same folder tree (in-place). Subfolders mirror 原图. "
+            "Animated GIFs and animated WebPs become .gif; static images become lossless .webp. "
+            "Next index per folder = max existing N in that folder + 1. "
+            "By default unnumbered source files are deleted after a successful conversion."
+        )
     )
     parser.add_argument(
         "source",
         nargs="?",
-        default=".",
-        help="Source folder, image file, or .zip archive. Default: current directory.",
+        default=None,
+        type=Path,
+        help=f"Source root folder, single image, or .zip (default: {DEFAULT_SOURCE_DIR.as_posix()}/).",
     )
     parser.add_argument(
-        "--output-base",
+        "-o",
+        "--output",
         default=None,
-        help="Output root. Default for folders: <source>-out next to the source. Default for zip: a folder next to the zip with the same name.",
+        dest="output_root",
+        type=Path,
+        help="Output root (default: same as source for folders/single file; for .zip default: <zip-stem>/).",
     )
     parser.add_argument(
         "-j",
         "--jobs",
         type=int,
-        default=max(1, min(32, os.cpu_count() or 4)),
-        help="Number of worker threads. Default: number of CPU cores.",
+        default=10,
+        help="Number of worker threads (default: 10).",
+    )
+    parser.add_argument(
+        "--keep-sources",
+        action="store_true",
+        help="Keep unnumbered source files after conversion (default: delete them).",
     )
     args = parser.parse_args(argv)
 
-    source = Path(args.source).expanduser().resolve()
-    output_base = Path(args.output_base).expanduser().resolve() if args.output_base else None
+    cwd = Path.cwd()
+    source = (
+        args.source.expanduser().resolve()
+        if args.source is not None
+        else (cwd / DEFAULT_SOURCE_DIR).resolve()
+    )
+    jobs: list[Job] = []
 
-    jobs = collect_jobs(source, output_base)
-    jobs.sort(key=lambda job: str(job.target_path))
+    if source.is_dir():
+        out_root = (
+            args.output_root.expanduser().resolve()
+            if args.output_root is not None
+            else source
+        )
+        out_root.mkdir(parents=True, exist_ok=True)
+        jobs = collect_directory_jobs(
+            source, out_root, delete_sources=not args.keep_sources
+        )
+    elif source.is_file() and source.suffix.lower() == ".zip":
+        out_root = (
+            args.output_root.expanduser().resolve()
+            if args.output_root is not None
+            else (source.parent / source.stem).resolve()
+        )
+        out_root.mkdir(parents=True, exist_ok=True)
+        jobs = collect_zip_jobs(source, out_root)
+    elif is_image_path(source):
+        if is_numbered_asset_file(source):
+            print(f"Skip already-numbered file: {source}", file=sys.stderr)
+            return 0
+        target_dir = (
+            args.output_root.expanduser().resolve()
+            if args.output_root is not None
+            else source.parent.resolve()
+        )
+        target_dir.mkdir(parents=True, exist_ok=True)
+        idx = max_index_in_directory(target_dir) + 1
+        jobs = [
+            Job(
+                source_label=source.name,
+                target_dir=target_dir,
+                index=idx,
+                loader=lambda p=source: p.read_bytes(),
+                source_file=source if not args.keep_sources else None,
+            )
+        ]
+    else:
+        print(f"Not a folder, image, or zip: {source}", file=sys.stderr)
+        return 1
+
+    jobs.sort(key=lambda job: (str(job.target_dir), job.index))
 
     if not jobs:
-        print(f"No supported images found in: {source}")
-        return 1
+        print(f"No new images to convert under: {source}")
+        return 0
 
     converted = 0
     skipped = 0
 
     with ThreadPoolExecutor(max_workers=max(1, args.jobs)) as pool:
-        for ok, label, target, animated, error in pool.map(run_job, jobs):
+        for ok, label, target, kind, error in pool.map(run_job, jobs):
             if ok:
-                kind = "animated" if animated else "static"
                 print(f"{label} -> {target} ({kind})")
                 converted += 1
             else: