Add find_regressions_from_logs.py to long-term logs collection pipeline

      - long_term_regression.html

    expose_as: "ivas long term job logs overview"

ivas-long-term-regressions:

  rules:

    - if: $CI_PIPELINE_SOURCE == 'web' && $MANUAL_PIPELINE_TYPE == "long-term-logs"

    - if: $CI_PIPELINE_SOURCE == 'schedule' && $MANUAL_PIPELINE_TYPE == "long-term-logs"

  needs:

    - ivas-long-term-job-logs

  stage: maintenance

  timeout: "25 minutes"

  script:

    - !reference [ .job-linux, before_script ]

    - set -euxo pipefail

    - bash "${CI_PROJECT_DIR}"/ivas-codec-ci/snippets/print-common-info.sh

    - bash "${CI_PROJECT_DIR}"/ivas-codec-ci/snippets/basop/update-scripts-repo.sh

    # Find regressions

    - python3 scripts/find_regressions_from_logs.py logs regressions_mld.csv  --measure MLD --days 7 --ratio_thr 1.1

    - python3 scripts/find_regressions_from_logs.py logs regressions_max_abs_diff.csv  --measure MAX_ABS_DIFF --days 7 --ratio_thr 1.1

  artifacts:

    name: "$CI_JOB_NAME--sha-$CI_COMMIT_SHORT_SHA--results"

    expire_in: 4 weeks

    when: always

    paths:

      - regressions_mld.csv

      - regressions_max_abs_diff.csv

    expose_as: "ivas long term regressions"

# To store backup copy:

# - Prepare accessible folder for backup, e.g. /usr/local/backup

# - Set MANUAL_PIPELINE_TRIGGER to backup-long-term-logs