Quick tourΒΆ
The simplest dogsled installation can be done via Conda using environment.yml file for Linux and macOS and win_environment.yml for Windows
Hint
It is highly recommended not to install dogsled (or any other package) in the base environment, create a new one using Conda or venv
user@arch:~$ git clone https://github.com/RhDm/dogsled.git
user@arch:~$ cd dogsled
user@arch:~$ conda env create -f environment.yml
Next, in your Python code, import dogsled and specify the folder in which the slides are located (source_path), the folder which should contain the normalised slides (norm_path) and the names of the slides to be normalised (slide_names):
>>> from dogsled.normaliser import NormaliseSlides >>> normaliser = NormaliseSlides(source_path = '/Users/uname/slides/', ... norm_path = '/Users/uname/slides/normalised', ... slide_names = 'CMU-1-Small-Region.svs')
Double check if everything is correct:
>>> normaliser.slide_paths
Rev up your engines and start normalisation!
>>> normaliser.start()
Now dogsled will normalise the slide, save the corresponding JPEGs in the specified folder and additionally generate the thumbnails of the pre-normalised slide and of the normalised image (with a thumbnail_ prefix)
user@arch:~$ cd /Users/uname/slides/normalised
user@arch:normalised$ ls -1
norm_SAS_21883_001.jpeg
thumbnail_norm_SAS_21883_001.jpeg
thumbnail_SAS_21883_001.jpeg
Note
Depending on your requirements and installation method, you may need to install QuPath as well. Please see installation for further details.