sknw icon indicating copy to clipboard operation
sknw copied to clipboard

Does latest installed version include `iso`?

Open petermr opened this issue 4 years ago • 4 comments

I am very interested in using iso, ring and other arguments. Currently I use

pip install sknw

Installing collected packages: sknw
Successfully installed sknw-0.14

but this does not seem to include iso

>       nx_graph = sknw.build_sknw(skeleton_image, multi=True, iso=False, ring=True, full=True)
E       TypeError: build_sknw() got an unexpected keyword argument 'iso'

Do I have to pull and build it locally?

(sknw is very useful to me and these additional features are just what I need)

petermr avatar Dec 18 '21 11:12 petermr

I have now cloned sknw and the setup.py indicates it has an earlier version number

pm286macbook:workspace pm286$ git clone https://github.com/Image-Py/sknw.git
Cloning into 'sknw'...
remote: Enumerating objects: 80, done.
remote: Counting objects: 100% (27/27), done.
remote: Compressing objects: 100% (21/21), done.
remote: Total 80 (delta 10), reused 21 (delta 6), pack-reused 53
Receiving objects: 100% (80/80), 17.17 KiB | 1.32 MiB/s, done.
Resolving deltas: 100% (33/33), done.
(base) pm286macbook:workspace pm286$ cd sknw/
(base) pm286macbook:sknw pm286$ ls
LICENSE		README.md	setup.py	sknw
(base) pm286macbook:sknw pm286$ more setup.py 
from setuptools import setup

descr = """sknw: skeleton analysis in Python.
Inspired by Juan Nunez-Iglesias's skan.
"""

if __name__ == '__main__':
    setup(name='sknw',
        version='0.13',
        url='https://github.com/yxdragon/sknw',
        description='Analysis of object skeletons',
        long_description=descr,
        author='YXDragon',
        author_email='[email protected]',
        license='BSD 3-clause',
        packages=['sknw'],
        package_data={},
        install_requires=[
            'numpy',
            'networkx',
            'numba'
        ],
    )

petermr avatar Dec 18 '21 11:12 petermr

you can just copy the sknw.py beside your script. do a test for me. if it works, I would publish it to pypi later.

yxdragon avatar Dec 18 '21 12:12 yxdragon

Which version should I use? V0.14 is the installed version but v0.13 appears more recent Should the checkout have V0.15

petermr avatar Dec 18 '21 13:12 petermr

@petermr The last PyPI package (v0.14) does not include the latest changes.

Right now we are using a copy of the sknw.py in our code with your reference. It seems to work! We'll import sknw from the package when you update the PyPI package.

anuvc avatar Mar 15 '22 05:03 anuvc