Source code for askemblaex.config

"""
askemblaex/config.py

Runtime configuration helpers.

The set of image extensions recognised by the pipeline can be overridden via
the ``ASKEMBLAEX_IMG_EXTS`` environment variable (comma-separated list,
e.g. ``".png,.jpg,.tiff"``).
"""

from __future__ import annotations

import os


DEFAULT_IMG_EXTS = {
    ".png",
    ".jpg",
    ".jpeg",
    ".tiff",
    ".tif",
}




def get_img_exts() -> set[str]:
    """
    Return the set of image file extensions the pipeline should process.

    Reads ``ASKEMBLAEX_IMG_EXTS`` from the environment. If unset, returns
    :data:`DEFAULT_IMG_EXTS`. Extensions are normalised to lowercase and
    prefixed with a dot if missing.

    Returns:
        Set of dot-prefixed lowercase extension strings e.g.
        ``{".png", ".jpg", ".tiff"}``.
    """
    raw = os.getenv("GENIEXTRACT_IMG_EXTS")
    if not raw:
        return DEFAULT_IMG_EXTS

    exts = set()
    for part in raw.split(","):
        part = part.strip().lower()
        if not part:
            continue
        if not part.startswith("."):
            part = "." + part
        exts.add(part)

    return exts