Models steel industry with the SEDOS reference dataset.
To install steel_industry, follow these steps:
- git-clone steel_industry into local folder:
git clone https://github.com/sedos-project/steel_industry.git
- enter folder
cd steel_industry
- create virtual environment using conda:
conda env create -f environment.yml
- activate environment:
conda activate steel_industry
- install data_adapter_oemof package using poetry, via:
poetry install
- If you have trouble try
poetry update
.
https://sedos-project.github.io/organization/scenarios/ https://sedos-project.github.io/organization/results/
In this template, 3 possible linters are proposed:
- flake8 only sends warnings and error about linting (PEP8)
- pylint sends warnings and error about linting (PEP8) and also allows warning about imports order
- black sends warning but can also fix the files for you
You can perfectly use the 3 of them or subset, at your preference. Don't forget to edit .travis.yml
if you want to deactivate the automatic testing of some linters!