large-image
raw JSON → 1.34.1 verified Fri May 01 auth: no python
A Python library for working with large, multiresolution images, supporting tiles, annotations, and warping. Current version is 1.34.1, with regular releases every few weeks.
pip install large-image Common errors
error ModuleNotFoundError: No module named 'openslide' ↓
cause Missing OpenSlide native library or Python binding.
fix
Install OpenSlide system package (e.g., apt install openslide-tools) and then
pip install openslide-python. error AttributeError: module 'large_image' has no attribute 'LargeImageSource' ↓
cause Incorrect import of legacy class name.
fix
Use
from large_image import getTileSource instead of instantiating a class. Warnings
gotcha Tile coordinates (x, y, z) are integer tile indices, not pixel coordinates. Use getRegion for pixel-based cropping. ↓
fix Use getRegion with pixel coordinates if you need arbitrary rectangular crops.
deprecated The `large_image_source_tiff` and other per-format modules are deprecated in favor of the unified `source` API. ↓
fix Use `from large_image import source` or `getTileSource` directly.
gotcha When installed without optional dependencies (e.g., OpenSlide, tifffile), unsupported image formats will raise an ImportError. ↓
fix Install with the appropriate extras: `pip install large-image[openslide,tiff]`.
Install
pip install large-image[sources] Imports
- source wrong
import large_image_source_tiffcorrectfrom large_image import source - getTileSource wrong
from large_image_source_tiff import getTileSourcecorrectfrom large_image import getTileSource
Quickstart
from large_image import getTileSource
# Open an image (replace with a path to a large image)
source = getTileSource('path/to/large_image.svs')
print('Image metadata:', source.getMetadata())
# Get a tile (x, y, z) - e.g., tile at top-left of level 0
tile = source.getTile(0, 0, 0)
print('Tile shape:', tile.shape)
# Get a region (format: {'left', 'top', 'right', 'bottom'})
region = source.getRegion(region={'left': 0, 'top': 0, 'right': 256, 'bottom': 256})
print('Region shape:', region.shape)