Reproducible Research

• Author: MinDong Sung
• Date: 2022-10-21
• For: DHLab

---

VSCode for Data Science

• Remote SSH
  • Ctrl + R: 최근 workspace 목록 확인
  • [참고] 비밀번호 입력하기 힘들면 SSH key 생성하기
• Shortcut
  • F5: Run
  • Ctrl or CMD + F5: Run without debugging
  • Ctrl or CMD + D: multi cursor
  • Ctrl or CMD + Shift + P: palette
• Duplicate workspace
• Jupyter notebook in VScode
  • #%%
  • Shift + Enter로 jupyter notebook 실행 : (setting.json)

    "jupyter.sendSelectionToInteractiveWindow": true
    

• Path set
  • Python script

    #.env
    PYTHONPATH=.
    

    # setting.json
    "python.envFile": "${workspaceFolder}/.env"
    

  • Python notebook

    # setting.json
    "jupyter.notebookFileRoot": "${workspaceFolder}"
    

Project Management

• envioronment 관리

  • pyenv: python version 관리
  • poetry: python library 관리
    • poetry init -n
    • poetry add
    • poetry install
  • using R in vscode - VSCode-R
  • renv: R library 관리
  • library(here)

• project structure

  • data/
    • raw/
    • processed/
  • src/
  • figure/
  • README.md
  • Snakefile/Makefile: for workflow

• git - code backup and version control

  • git init
  • git add
  • git commit
  • git push

• DVC - Data backup and version control
  • AWS-cli prerequisite
  • poetry add 'dvc[s3]'
  • dvc init
  • dvc remote add -d data s3://data-folder
  • dvc remote modify --local data access_key_id '[mykey]'
  • dvc remote modify --local bikes secret_access_key '[mysecret]'
  • dvc add
  • dvc push
  • dvc commit
  • dvc pull