Skip to content

Add Harbor benchmark runner scripts and document k=1/k=3 sweeps #3

Add Harbor benchmark runner scripts and document k=1/k=3 sweeps

Add Harbor benchmark runner scripts and document k=1/k=3 sweeps #3

Workflow file for this run

name: ci
on:
pull_request:
push:
branches: [main]
# Only the lint job runs in GitHub Actions. Building the `bioterm-*:v1`
# Docker images pulls multi-GB of bioinformatics tool chains and can
# chew through a private account's runner quota in a single push, so
# image builds and the oracle-all sweep are intentionally kept local
# (see scripts/build_images.sh and scripts/run_oracle_all.sh).
jobs:
lint:
name: lint tasks
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- uses: actions/setup-python@v5
with:
python-version: "3.12"
- name: install helper deps
run: pip install click
- name: syntax-check shell scripts
run: |
bash -n scripts/build_images.sh
bash -n scripts/run_oracle_all.sh
bash -n scripts/refresh_db_golds.sh
bash -n scripts/new_task.sh
- name: validate task directories
run: python scripts/validate_task.py
- name: scaffolder smoke test
run: |
bash scripts/new_task.sh ci-smoke-stub --category metadata
# scaffolded tasks intentionally fail validation until filled in
if python scripts/validate_task.py --only ci-smoke-stub; then
echo "scaffolder regression: placeholder task passed validation" >&2
exit 1
fi
rm -rf tasks/ci-smoke-stub