stlearn.pp.tiling

stlearn.pp.tiling(adata: AnnData, out_path: Path | str = './tiling', library_id: str | None = None, crop_size: int = 40, target_size: int = 299, img_fmt: str = 'JPEG', quality: int = 75, verbose: bool = False, copy: bool = False) → AnnData | None[source]

Tiling H&E images to small tiles based on spot spatial location.

Parameters:

  • adata (AnnData) – Annotated data matrix containing spatial information.

  • out_path (Path or str, default "./tiling") – Path to save spot image tiles.

  • library_id (str, optional) – Library id stored in AnnData. If None, uses first available library.

  • crop_size (int, default 40) – Size of tiles to crop from original image.

  • target_size (int, default 299) – Target size for resized tiles (input size for CNN).

  • img_fmt (str, default "JPEG") – Image format (‘JPEG’ or ‘PNG’).

  • quality (int, default 75) – JPEG quality (1-100). Only used for JPEG format.

  • verbose (bool, default False) – Enable verbose output.

  • copy (bool, default False) – Return a copy instead of modifying adata in-place.

Returns:

  • Depending on copy, returns or updates adata with the following fields.

  • **tile_path** (adata.obs field) – Saved path for each spot image tiles