Пример использования ML Space CLI в пайплайне

stages:
- start

start-learning-job:
stage: start
rules:
    - if: '$CI_COMMIT_TAG != null'
variables:
    MLS_VERSION: "0.4.0"
    APIKEY_ID: <escaped>
    APIKEY_SECRET: <escaped>
    WORKSPACE_ID: <escaped>
    X_API_KEY: <escaped>
    REGION: S000X
    OUTPUT: json
    PROFILE: PREPROD
script:
    - echo 'Сбор данных... done'
    - echo 'Пред обработка данных ... done'
    - echo 'Обучение модели...'
    - pip install "mls==$(MLS_VERSION)" --index-url https://gitverse.ru/api/packages/cloudru/pypi/simple/ --extra-index-url https://pypi.org/simple -U
    - printf "$(APIKEY_ID)\n$(APIKEY_SECRET)\n$(WORKSPACE_ID)\n$(X_API_KEY)\n$(REGION)\n$(OUTPUT)\n" | mls configure --profile $(PROFILE)
    - mls job run --config job.yaml --output json|jq -r ".job_name" > job_name.txt
artifacts:
    paths:
    - job_name.txt