Graph Embedding Evaluation / Code and Datasets for "Graph Embedding on Biomedical Networks: Methods, Applications, and Evaluations" (Bioinformatics 2020)
This PR does a couple things, but mainly it closes #2. I did my best to leave helpful commit messages - I copied the most important ones here. I'd be happy to explain anything I did or give more insight on how you can continue to make your package more accessible for others!
Reorganize repository and clean up imports
Switched to the /src/bionev structure because it's the community standard
Added encoding line to beginning of each python file
Made all imports relative to bionev
Used PyCharm's "optimize imports" function to reorganize the order and remove unused imports
Make pip installable and relax requirements
requirements.txt is only for non-packages, so now the requirements live inside setup.cfg
The old requirements were pinned to very old versions of each library. I relaxed those requirements and let it install the newest versions of everything except tensorflow, and it seemed to work still.
Updated the README with installation instructions
Add CLI entrypoint
Used the entry_points to make it possible to run bionev from anywhere instead of needing python src/main.py
Updated the running example in the README to use the new CLI
Adds More Evaluation Metrics
I wanted to see the AUC-PR and MCC presented, so I added them to the output.
I changed the eval_results.txt output to make the output in structured JSON. this will make programmatic analysis of results much easier in the future
Thanks so much! It's my first time to build such repo and I don't have much experience. I appreciate your help for making it better. Feel free to let me know if you have further suggestions.
This PR does a couple things, but mainly it closes #2. I did my best to leave helpful commit messages - I copied the most important ones here. I'd be happy to explain anything I did or give more insight on how you can continue to make your package more accessible for others!
Reorganize repository and clean up imports
Make pip installable and relax requirements
Add CLI entrypoint
Adds More Evaluation Metrics