ibis-benchmarks.yml

name: Benchmarks

on:
  push:
    branches:
      - main
      - "*.x.x"
  merge_group:

# since we're writing to cloud storage, we don't want to have multiple
# instances of this job running at one time
concurrency: benchmarks-${{ github.repository }}

permissions:
  # increase the rate limit for github operations, but limit token permissions
  # to read-only
  contents: read

jobs:
  benchmarks:
    runs-on: ubuntu-latest
    steps:
      - name: checkout
        uses: actions/checkout@v4

      - name: install python
        uses: actions/setup-python@v5
        id: install_python
        with:
          python-version: "3.11"

      - name: install uv
        uses: astral-sh/setup-uv@v5.0.1

      - name: install system dependencies
        run: sudo apt-get install -qq -y build-essential libgeos-dev freetds-dev unixodbc-dev

      - name: make benchmark output dir
        run: mkdir .benchmarks

      - name: benchmark
        run: uv run --all-extras --group tests pytest --benchmark-enable --benchmark-json .benchmarks/output.json ibis/tests/benchmarks

      - uses: google-github-actions/auth@v2
        with:
          credentials_json: ${{ secrets.GCP_CREDENTIALS }}

      - uses: google-github-actions/setup-gcloud@v2

      - name: show gcloud info
        run: gcloud info

      - name: download the latest duckdb release
        env:
          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
        run: |
          set -euo pipefail

          gh release download -R duckdb/duckdb --pattern 'duckdb_cli-linux-amd64.zip'
          unzip duckdb_cli-linux-amd64.zip

      - name: convert json data to parquet
        run: |
          set -euo pipefail

          # sort json keys
          jq --sort-keys -rcM < "$PWD/.benchmarks/output.json" > output.json

          # connect to a file to allow spilling to disk
          ./duckdb json2parquet.ddb <<EOF
            COPY (
              SELECT * FROM read_ndjson_auto('output.json', maximum_object_size=2**27)
            ) TO 'output.parquet' (FORMAT PARQUET, COMPRESSION ZSTD)
          EOF

      - name: copy data to gcs
        run: |
          set -euo pipefail

          timestamp="$(date --iso-8601=ns --utc | tr ','  '.')"
          gsutil cp output.parquet "gs://ibis-benchmark-data/ci/${timestamp}.parquet"