DeepProZyme not installable & legally usable?

GwennyGit commented 1 month ago

Hi,

Encountered installation issues

I tried to install your tool and encountered the following issues:

The environement.yml did not work for me, so I tried to install everything from the file by hand.
While doing that, I could not create a Conda environment with Python 3.6. Looking at the documentation for Conda, I discovered that the oldest supported Python version for Conda is 3.8.
With a conda environment with Python 3.8, I could not install the requested versions of Pytorch and transformers.

Licensing

Along with that, I realised that your tool is not licensed. Without a license, I am unsure if I have the right to use your software legally. On the website The Legal Side of Open Source, it is stated:

If you want others to use, distribute, modify, or contribute back to your project, you need to include an open source license.

This sounds like I have no right to use software without a license. 🤔 Adding a license with GitHub is very simple. See here for more information.

Maiya19724 commented 1 month ago

Hi,

Encountered installation issues

I tried to install your tool and encountered the following issues:

The environement.yml did not work for me, so I tried to install everything from the file by hand.

While doing that, I could not create a Conda environment with Python 3.6. Looking at the documentation for Conda, I discovered that the oldest supported Python version for Conda is 3.8.

With a conda environment with Python 3.8, I could not install the requested versions of Pytorch and transformers.

Licensing

Along with that, I realised that your tool is not licensed. Without a license, I am unsure if I have the right to use your software legally. On the website The Legal Side of Open Source, it is stated:

If you want others to use, distribute, modify, or contribute back to your project, you need to include an open source license.

This sounds like I have no right to use software without a license. 🤔 Adding a license with GitHub is very simple. See here for more information.

Which version of Conda are you using? Perhaps adding a channel that supports Python 3.6 in Conda might solve the problem. @GwennyGit

(deepectransformer) symlab@symlab:~/software/DeepProZyme$ python --version
Python 3.6.15
(deepectransformer) symlab@symlab:~/software/DeepProZyme$ conda --version
conda 23.7.4
(deepectransformer) symlab@symlab:~/software/DeepProZyme$ mamba list | grep "python"
biopython                 1.78             py36h8f6f2f9_2    conda-forge
python                    3.6.15          hb7a2778_0_cpython    conda-forge
python-dateutil           2.8.2              pyhd8ed1ab_0    conda-forge
python_abi                3.6                     2_cp36m    conda-forge

GwennyGit commented 1 month ago

Thank you for your fast reply.

My conda version is conda 24.3.0; I will probably update it to the newest version soon.

My main issue is that my colleague @cb-hades and I would like to use your tool within our tool refineGEMs. We are currently refactoring the code and adding new functionalities to the branch dev-2. To add a version of gap filling via genes where no data on the user’s organism is available in any database, we wanted to integrate/use your tool. However, refineGEMs will be working with Python 3.10 and higher only after our new release, making it impossible to integrate your tool without docker or a virtual environment. Moreover, without knowing if we can legally use your tool, we are hesitant to integrate it at all. (That is why we asked about the license in the first place.) 🤔

Maiya19724 commented 1 month ago

I'm very sorry, but I'm not the author of this tool, so I can't answer your question. RefineGEMs looks like a great tool. Additionally, in terms of EC number prediction, I personally believe that Clean has better capabilities than DeepProZyme. Given the response time issues with the DeepProZyme author, I would recommend using Clean in your tool. @GwennyGit

kaistsystemsbiology / DeepProZyme