# retoor <retoor@molodetz.nl>
"""Validate gallery source files against the catalogue contract.

Usage:
    python tools/gallery/check.py [media_dir]

Checks every flyer-*.jpg and meme-*.jpg in the media directory (default:
the repo static/media tree) plus every entry of molodetz.gallery.GALLERY:

- name is lowercase, starts with flyer- or meme-, ends in .jpg
- flyers are 1080x1350, memes are 1080x1080 (1280x720 legacy memes warn)
- every source file is catalogued or retired, every catalogue entry has a file (repo tree only; scratch dirs get warnings)
- catalogue captions are non-empty English without em or en dashes
- no two catalogued works are perceptual near-duplicates (sync would skip one)

Exit code 0 when clean, 1 when problems are found. Warnings do not fail.
"""

import sys
from pathlib import Path

REPO_ROOT = Path(__file__).resolve().parent.parent.parent
sys.path.insert(0, str(REPO_ROOT))

from PIL import Image  # noqa: E402

import imagehash  # noqa: E402

from molodetz.gallery import DUPLICATE_DISTANCE, GALLERY, RETIRED_GALLERY  # noqa: E402

DIMENSIONS = {"flyer": (1080, 1350), "meme": (1080, 1080)}
LEGACY_MEME = (1280, 720)


def check(media_dir, strict=True):
    errors, warnings = [], []
    catalogued = {filename: (kind, caption) for filename, kind, caption in GALLERY}
    retired = set(RETIRED_GALLERY)
    files = sorted(path for path in media_dir.iterdir() if path.suffix.lower() == ".jpg" and path.stem.split("-")[0].lower() in ("flyer", "meme"))
    seen = []
    for path in files:
        name = path.name
        if name != name.lower():
            errors.append(f"{name}: name must be lowercase")
        kind = "flyer" if name.lower().startswith("flyer-") else "meme"
        if name not in catalogued and name not in retired:
            if strict:
                errors.append(f"{name}: present on disk but missing from GALLERY")
                continue
            warnings.append(f"{name}: not catalogued yet, add a GALLERY entry on handoff")
            caption = ""
        elif name in retired:
            continue
        else:
            expected_kind, caption = catalogued[name]
            if expected_kind != kind:
                errors.append(f"{name}: catalogued as {expected_kind}, prefix says {kind}")
        if strict and (not caption or not caption.strip()):
            errors.append(f"{name}: catalogue caption is empty")
        elif "\u2014" in caption or "\u2013" in caption:
            errors.append(f"{name}: caption contains an em or en dash")
        with Image.open(path) as image:
            size = image.size
            digest = imagehash.phash(image)
        expected = DIMENSIONS[kind]
        if size != expected:
            if kind == "meme" and size == LEGACY_MEME:
                warnings.append(f"{name}: legacy 1280x720 size, new memes must be 1080x1080")
            else:
                errors.append(f"{name}: size {size[0]}x{size[1]}, expected {expected[0]}x{expected[1]}")
        for other_name, other_digest in seen:
            if digest - other_digest <= DUPLICATE_DISTANCE:
                errors.append(f"{name}: near-duplicate of {other_name}, sync would skip it")
        seen.append((name, digest))
    on_disk = {path.name for path in files}
    if strict:
        for filename in catalogued:
            if filename not in on_disk:
                errors.append(f"{filename}: catalogued in GALLERY but missing from {media_dir}")
    return errors, warnings


def main(argv):
    default_dir = REPO_ROOT / "molodetz" / "static" / "media"
    media_dir = Path(argv[1]) if len(argv) > 1 else default_dir
    if not media_dir.is_dir():
        print(f"missing directory: {media_dir}")
        return 1
    strict = media_dir.resolve() == default_dir.resolve()
    errors, warnings = check(media_dir, strict=strict)
    for line in warnings:
        print(f"warn: {line}")
    for line in errors:
        print(f"error: {line}")
    if errors:
        print(f"{len(errors)} error(s), {len(warnings)} warning(s)")
        return 1
    print(f"clean: {media_dir} ({len(warnings)} warning(s))")
    return 0


if __name__ == "__main__":
    raise SystemExit(main(sys.argv))