author: steeve laquitaine
Go to documentation
Create virtual environment and install dependencies:
conda create -n bsfit python==3.6.13
conda activate bsfit
conda install --file bsfit/requirements.txt -y
Setup jupyter notebook:
conda install -n heuristic ipykernel --update-deps --force-reinstall
Unit-Test all the package's functions:
pytest bsfit/test.py
- Keep Doctstrings in Google Style Guide format.
- Edit docs/source/
- Go to "Build & deploy" section
Basic steps: develop (features) -> pre-prod (homologation) -> master (production)
-
On branch
develop
(to add features):- when: after having added a new working feature
bash deploy.sh
-
Pull request
pre-prod
branch (homologation, check docs rendering)- when: after deploying sphinx/ to docs/
-
Pull request to
master
(production branch for end-users)- when: the entire codebase is clean
Main paper to be cited (Laquitaine et al., 2018)
@article{laquitaine2018switching,
title={A switching observer for human perceptual estimation},
author={Laquitaine, Steeve and Gardner, Justin L},
journal={Neuron},
volume={97},
number={2},
pages={462--474},
year={2018},
publisher={Elsevier}
}