metadata.py

import subprocess
from pathlib import Path

from vision import TagResult


# Formats that support IPTC embedding via exiftool
_IPTC_FORMATS = {".jpg", ".jpeg"}


def _set_list(args: list, flag: str, values: list[str]) -> None:
    """Append exiftool args to replace a list tag with the given values."""
    if values:
        args.append(f"-{flag}={values[0]}")
        for v in values[1:]:
            args.append(f"-{flag}+={v}")
    else:
        args.append(f"-{flag}=")


def write_metadata(path: str, result: TagResult) -> None:
    """Write bilingual tags and captions to image metadata in-place using exiftool.

    XMP is used for all formats (HEIC, HEIF, JPEG):
      - XMP:Subject         — all keywords (EN + DE combined)
      - XMP-dc:Description-en — English caption (language alternative)
      - XMP-dc:Description-de — German caption (language alternative)

    IPTC is added for JPEG only:
      - IPTC:Keywords            — all keywords (EN + DE combined)
      - IPTC:Caption-Abstract    — English caption (IPTC has no language tagging)
    """
    suffix = Path(path).suffix.lower()
    args = ["exiftool", "-overwrite_original", "-charset", "UTF8"]

    all_tags = result.tags_en + result.tags_de

    # --- XMP (universal) ---
    _set_list(args, "XMP:Subject", all_tags)

    if result.caption_en:
        args.append(f"-XMP-dc:Description-en={result.caption_en}")
    if result.caption_de:
        args.append(f"-XMP-dc:Description-de={result.caption_de}")

    # --- IPTC (JPEG only) ---
    if suffix in _IPTC_FORMATS:
        _set_list(args, "IPTC:Keywords", all_tags)
        if result.caption_en:
            args.append(f"-IPTC:Caption-Abstract={result.caption_en}")

    args.append(path)

    proc = subprocess.run(args, capture_output=True, text=True)
    if proc.returncode != 0:
        raise RuntimeError(f"exiftool error: {proc.stderr.strip()}")