histo_patch : Patch extraction library for Whole Slide Images

Overview

histo_patch is a python library to extract patches from annotated region in WSI files (Hamamatsu .ndpi or Aperio .svs). Hamamatsu .ndpa and QuPath .geojson annotation files are supported.

Installation

The package can be installed with pip:

$ pip install histo-patch

To run histo_patch, simply run below from console:

$ histopatch [something]

Usage:

Usage: histopatch [OPTIONS] PATH_TO_WSI PATH_TO_SAVE_DIRECTORY

Options:
-a, --annotation_type [ndpa|qupath]
                                 file type of annotation (ndpa or qupath
                                 (GeoJSON(Pretty JSON)))
-s, --annotation_shape [all|line|area]
                                 annotation shape (all, line, area)
-c, --color TEXT                used colors (white,black,red,green,blue,cyan
                                 ,magenta,yellow) in annotation for ndpa or (
                                 None,Tumor,Stroma,Immune_cells,Necrosis,Othe
                                 r,Region,Ignore,Positive,Negative) for
                                 QuPath
--line_as_area                  line annotation in ndpa file is treated as
                                 closed area
--src_size FLOAT                patch size in the original WSI
-m, --micrometer                specify src_size in micrometer
                                 (default:pixels)
--patch_size INTEGER            output patch size
--num_patch INTEGER             number of patches for each annotation shape
--nparent INTEGER               number of parent directories kept in the
                                 output
--help                          Show this message and exit.

Usage

histopatch [OPTIONS] PATH_TO_WSI PATH_TO_SAVE_DIRECTORY:

Options

-a, --annotation_type [ndpa|qupath] : File type of annotation (ndpa or qupath (GeoJSON (Pretty JSON)))

-s, --annotation_shape [all|line|area] : Annotation shape (all, line, area)

-c, --color TEXT : Colors used in annotation for ndpa (white, black, red, green, blue, cyan, magenta, yellow) or for QuPath (None, Tumor, Stroma, Immune_cells, Necrosis, Other, Region, Ignore, Positive, Negative)

--line_as_area : Line annotation in ndpa file is treated as closed area

--src_size FLOAT : Patch size in the original WSI

-m, --micrometer : Specify src_size in micrometer (default: pixels)

--patch_size INTEGER : Output patch size

--num_patch INTEGER : Number of patches for each annotation shape

--nparent INTEGER : Number of parent directories kept in the output

--help : Show this message and exit.

Prerequisites

Python version 3.6 or newer.

• tifffile >=2022.5.4
• numpy >=1.20.3
• tripy >=1.0.0
• pyclipper >=1.3.0
• opencv-python >= 4.6.0
• zarr >=2.11.3
• magecodecs >=2022.2.22
• click >=8.1.3

Recommended Environment

• OS

  • Linux
  • Mac