tokebe / niclassify

A combined GUI toolkit for predictively classifying arthropod sequences as Native/Introduced
GNU General Public License v3.0
0 stars 0 forks source link

Installation issue at scikit-learn #37

Open nr-graham opened 11 months ago

nr-graham commented 11 months ago

Hi Jackson. I'm installing the newest niclassify 1.5.1. I'm running Python 3.9 on Windows 10. During installation, I ran into a problem with the virtual env set up. Specifically at scikit-learn. See error message below. Note I was able to pip install in a venv in a separate console no problem using: python -m venv sklearn-venv sklearn-venv\Scripts\activate pip install -U scikit-learn

See error message: Installed c:\users\graha\desktop\desktop\niclassify-1.5.1-beta\niclassify-1.5.1-beta\niclassifyenv\lib\site-packages\scipy-1.5.4-py3.9-win-amd64.egg Searching for scikit-learn==0.24.0rc1 Reading https://pypi.org/simple/scikit-learn/ No local packages or working download links found for scikit-learn==0.24.0rc1 error: Could not find suitable distribution for Requirement.parse('scikit-learn==0.24.0rc1')

Could it just be a problem with scikit-learn==0.24.0rc1 ? I can see that seaborn and scipi were added to the easy install path before my error occurred. Adding seaborn 0.11.0 to easy-install.pth file Adding scipy 1.5.4 to easy-install.pth file

tokebe commented 11 months ago

Hi Natalie, it looks like at some point that specific version was made unavailable. I'll test removing the 'rc1' part of the version, and if that fixes installation I'll push the change and let you know.

tokebe commented 11 months ago

@nr-graham I've published a new release which should address this issue.

Please note that niclassify has become wildly out of date in a number of ways (for instance, python 3.9 binaries are no longer provided), which I will be addressing in a rewrite I had previously mentioned. Unfortunately this rewrite has spent a lot of time on the backburner due to other obligations, but progress is being made.

nr-graham commented 11 months ago

Hi Jackson! Thanks for the update. I’ll try it again. Yes I was astounded to see how many new python versions were out since you started this project! Best, Natalie

On Tue, Oct 31, 2023 at 11:28 AM Jackson Callaghan @.***> wrote:

@nr-graham https://github.com/nr-graham I've published a new release which should address this issue.

Please note that niclassify has become wildly out of date in a number of ways (for instance, python 3.9 binaries are no longer provided), which I will be addressing in a rewrite I had previously mentioned. Unfortunately this rewrite has spent a lot of time on the backburner due to other obligations, but progress is being made.

— Reply to this email directly, view it on GitHub https://github.com/tokebe/niclassify/issues/37#issuecomment-1788061993, or unsubscribe https://github.com/notifications/unsubscribe-auth/AL3EOK3Q6Z5BHLGQRXYYDGTYCFUOZAVCNFSM6AAAAAA6XHMBGKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTOOBYGA3DCOJZGM . You are receiving this because you were mentioned.Message ID: @.***>

-- Natalie R. Graham, Ph.D. (she/her) Research Scientist Strategic Environmental Research and Development Program (SERDP) Dept. Environmental Science Policy and Management University of California, Berkeley Berkeley sits on Ohlone land https://cejce.berkeley.edu/ohloneland

Need a science expert? You can request a woman in STEMM https://gage.500womenscientists.org/ Google scholar: https://scholar.google.com/citations?hl=en&user=UaFEoCEAAAAJ Linked-in: www.linkedin.com/in/natalie-graham-ecological-data