variables:
  GIT_STRATEGY: fetch
  GIT_SSL_NO_VERIFY: "true"
  GIT_SUBMODULE_STRATEGY: recursive
  GIT_LFS_SKIP_SMUDGE: 1

.docker-before-script: &docker_before_script  # This is an anchor
  before_script:
    - docker login -u gitlab-ci-token -p $CI_JOB_TOKEN http://$CI_REGISTRY

.git-before-script: &git_before_script
  before_script:
    - git config --global user.name "CR (covid cron)"
    - git config --global user.email "beepbop@example.com"
    - url_host=`git remote get-url origin | sed -e "s/https:\/\/gitlab-ci-token:.*@//g"`
    - cd /tmp/
    - git clone "https://gitlab-ci-token:${CI_PUSH_TOKEN}@${url_host}"
    - cd covid-19-public-data
    - git lfs install --local
    - PATH=$PATH:~/.local/bin

stages:
  - build
  - datasets-update
  - notebooks-update
  - finalize-update

image_build:
  stage: build
  image: docker:stable
  except:
    - /^auto-update.*$/
    - schedules
  <<: *docker_before_script

  script:
    - CI_COMMIT_SHA_7=$(echo $CI_COMMIT_SHA | cut -c1-7)
    - docker build --tag $CI_REGISTRY_IMAGE:$CI_COMMIT_SHA_7 .
    - docker push $CI_REGISTRY_IMAGE:$CI_COMMIT_SHA_7
  tags:
    - image-build

update-build:
  stage: build
  image: docker:stable
  only:
    - schedules
  <<: *docker_before_script

  script:
    - docker build --tag $CI_REGISTRY_IMAGE:$CI_COMMIT_SHORT_SHA-update .
    - docker push $CI_REGISTRY_IMAGE:$CI_COMMIT_SHORT_SHA-update
  tags:
    - image-build

datasets:
  stage: datasets-update
  image:
    name: $CI_REGISTRY_IMAGE:$CI_COMMIT_SHORT_SHA-update
    entrypoint: [""]

  only:
    - schedules
  <<: *git_before_script
  script:
    - export BRANCH_NAME=auto-update-datasets-${CI_COMMIT_SHORT_SHA}
    - git checkout -b $BRANCH_NAME
    - renku dataset update covid-19_jhu-csse
    - renku dataset add covid-19_jhu-csse --force --source csse_covid_19_data/csse_covid_19_daily_reports https://github.com/CSSEGISandData/COVID-19.git
    - renku dataset update covid-19-tweet-ids
    - renku dataset update openzh-covid-19
    - renku dataset update covid-19-italy
    - renku dataset update covid-19-spain
    - renku dataset update covid-19-us-nyt
    - renku dataset add covid-19-ecdc -d covid-19-ecdc.csv https://opendata.ecdc.europa.eu/covid19/casedistribution/csv --force || true
    - renku dataset add covid-19-chile -s data/*.csv  -s data/covid19_chile.rds https://github.com/itoledor/coronavirus.git --force || true
    - renku dataset add covidtracker https://ocgptweb.azurewebsites.net/CSVDownload -d covidtracker.csv --force || true
    - renku rerun data/covidtracking/states-metadata.json data/covidtracking/states-daily.json
    - renku run papermill -p out_folder ./data/distancing-metrics/ --inject-paths notebooks/process/download-distancing-data.ipynb runs/download-distancing-data.runs.ipynb
    - git push --set-upstream origin $BRANCH_NAME

notebooks:
  stage: notebooks-update
  image:
    name: $CI_REGISTRY_IMAGE:$CI_COMMIT_SHORT_SHA-update
    entrypoint: [""]
  only:
    refs:
      - schedules
  <<: *git_before_script
  script:
    - export BRANCH_NAME=auto-update-notebooks-${CI_COMMIT_SHORT_SHA}
    - git checkout auto-update-datasets-${CI_COMMIT_SHORT_SHA}
    - git checkout -b $BRANCH_NAME
    - renku update --with-siblings
    - git push --set-upstream origin $BRANCH_NAME

finalize:
  stage: finalize-update
  image:
    name: $CI_REGISTRY_IMAGE:$CI_COMMIT_SHORT_SHA-update
    entrypoint: [""]
  only:
    - schedules
  <<: *git_before_script
  when: always
  script:
    - export BRANCH_NAME="auto-update"_$(date -u +"%Y-%m-%d_%H-%M")
    - git checkout auto-update-notebooks-${CI_COMMIT_SHORT_SHA} || git checkout auto-update-datasets-${CI_COMMIT_SHORT_SHA}
    - git rev-parse --abbrev-ref HEAD
    - git checkout -b $BRANCH_NAME
    - git push origin --delete auto-update-notebooks-${CI_COMMIT_SHORT_SHA} || true
    - git push origin --delete auto-update-datasets-${CI_COMMIT_SHORT_SHA} || true
    - git push origin $BRANCH_NAME