benchmarks.yml

name: ANN benchmarks

on:
  push:
    branches:
      - main
  pull_request:
  workflow_dispatch:

# Cancel the workflow for the previous commit when the new commit is pushed.
concurrency:
  group: ${{ github.workflow }}-${{ github.ref }}
  cancel-in-progress: True

jobs:
  unit-tests:
    runs-on: ubuntu-22.04
    steps:
      - uses: actions/checkout@v3 # Pull the repository
      - run: sudo apt-get update && sudo apt-get install -y libhdf5-dev python3-numpy python3-scipy python3-matplotlib python3-sklearn
      - run: pip3 install --quiet -r requirements.txt
      - run: pytest

  run-benchmarks:
    runs-on: ubuntu-22.04
    timeout-minutes: 30
    strategy:
      fail-fast: false
      matrix:
        dataset: [random-xs-20-angular]
        library:
          - annoy
          - balltree
          - bruteforce
          - ckdtree
          - descartes
          - diskann
          - dolphinnpy
          - elasticsearch
          - elastiknn
          - faiss
          - flann
          - glass
          - hnswlib
          - kdtree
          - luceneknn
          - milvus
          - mrpt
          - nndescent
          - n2
          - nmslib
          - onng_ngt
          - opensearchknn
          - panng_ngt
          - pg_embedding
          - pgvector
          - pgvecto_rs
          - pynndescent
          - redisearch
          - qdrant
          - qg_ngt
          - qsg_ngt
          - scann
          - sptag
          - tinyknn
          - vald
          - vearch
          - vespa
          - voyager
          - weaviate
        include:
          - library: pynndescent
            dataset: random-xs-16-hamming
          - library: datasketch
            dataset: random-s-jaccard
          - library: pynndescent
            dataset: random-s-jaccard

    name: ${{ matrix.library }} (${{ matrix.dataset }})

    env:
      LIBRARY: ${{ matrix.library }}
      DATASET: ${{ matrix.dataset }}

    steps:
    - uses: actions/checkout@v3 # Pull the repository

    - name: Install OS Dependencies
      run: sudo apt-get update && sudo apt-get install -y libhdf5-dev python3-numpy python3-scipy python3-matplotlib python3-sklearn

    - name: Install Project Dependencies
      run: pip3 install --quiet -r requirements.txt

    - name: Build Library Docker Image
      run: python3 install.py

    - name: Run the benchmark
      run: |
        python3 run.py --docker-tag ann-benchmarks-${LIBRARY} --max-n-algorithms 3 --runs 2 --dataset $DATASET --run-disabled --timeout 300
        python3 run.py --docker-tag ann-benchmarks-${LIBRARY} --max-n-algorithms 3 --runs 2 --dataset $DATASET --run-disabled --batch --timeout 300
        sudo chmod -R 777 results/
        python3 plot.py --dataset $DATASET --output plot.png
        python3 plot.py --dataset $DATASET --output plot-batch.png --batch
        python3 data_export.py --out test.csv
        python3 create_website.py --outputdir . --scatter --latex