config.py

from __future__ import annotations

from dataclasses import dataclass, field
from pathlib import Path
from typing import Optional


SUPPORTED_FORMATS: frozenset[str] = frozenset({"avif", "jpg", "jpeg", "png", "webp"})

SUPPORTED_MIME_TYPES: frozenset[str] = frozenset({
    "image/avif",
    "image/jpeg",
    "image/jpg",
    "image/png",
    "image/webp",
})


@dataclass
class Config:
    """Runtime configuration for the Image Processor Pro pipeline."""

    output_root: Path = Path("output")
    logs_root: Path = Path("logs")

    jpeg_quality: int = 85
    max_width: Optional[int] = None
    max_height: Optional[int] = None

    download_timeout: int = 30
    max_retries: int = 3
    retry_backoff: float = 1.5
    max_workers: int = 8

    chunk_size: int = 16384
    user_agent: str = "ImageProcessorPro/1.0 (+https://example.local)"

    strip_metadata: bool = True
    background_color: tuple[int, int, int] = (255, 255, 255)

    # Watermark removal — color-agnostic text detection inside a fixed region.
    # Uses morphological top-hat + black-hat to isolate thin text strokes
    # (works for white OR black text on any background), filtered by
    # connected-component area so uniform backgrounds aren't masked.
    remove_watermark: bool = False
    watermark_engine: str = "lama"       # inpaint backend: "lama" (AI) or "classical" (cv2)
    watermark_device: str = "cpu"        # LaMa compute device: "cpu" or "mps"
    watermark_x_frac: float = 0.0        # search region left edge (fraction)
    watermark_y_frac: float = 0.93       # search region top edge (fraction)
    watermark_w_frac: float = 0.55       # search region width (fraction)
    watermark_h_frac: float = 0.07       # search region height (fraction)
    watermark_text_kernel: int = 9       # morphological kernel size (px, odd)
    watermark_text_threshold: int = 25   # contrast threshold (0-255)
    watermark_min_component_area: int = 4     # smallest CC kept (noise filter)
    watermark_max_area_frac: float = 0.6      # reject a blob bigger than this fraction of the ROI
    watermark_inpaint_radius: int = 3    # cv2.inpaint radius in pixels
    watermark_mask_feather: int = 2      # px of dilation on the mask edge (covers anti-aliased text halo)

    allowed_formats: frozenset[str] = field(default_factory=lambda: SUPPORTED_FORMATS)

    def output_dir_for_today(self) -> Path:
        from datetime import datetime

        day = datetime.now().strftime("%Y-%m-%d")
        path = self.output_root / day
        path.mkdir(parents=True, exist_ok=True)
        return path

    def ensure_dirs(self) -> None:
        self.output_root.mkdir(parents=True, exist_ok=True)
        self.logs_root.mkdir(parents=True, exist_ok=True)