Automate running benchmarks for all engines

.github/workflows/actions/run-engine-benchmark/action.yaml

@@ -0,0 +1,38 @@
+name: Run Engine Benchmark
+description: "Run benchmark with specified params"
+inputs:
+  engine:
+    description: "engine (i.e qdrant-default)"
+    required: true
+  dataset:
+    description: "dataset (i.e random-100)"
+    required: true
+  compose_file:
+    description: "path to docker compose"
+    required: true
+
+runs:
+  using: "composite"
+  steps:
+    - name: Install poetry
+      shell: bash
+      run: pip install poetry
+    - uses: actions/setup-python@v5
+      with:
+        python-version: "3.10"
+        cache: "poetry"
+    - name: Install deps
+      shell: bash
+      run: poetry install
+    - uses: hoverkraft-tech/[email protected]
+      with:
+        compose-file: "${{ inputs.compose_file }}"
+    - name: Execution
+      shell: bash
+      run: |
+        engine="${{ inputs.engine }}"
+        if [[ "$engine" == *"elasticsearch"* || "$engine" == *"opensearch"* ]]; then
+          ./tools/wait_for_green_status.sh
+        fi
+        source $(poetry env info -p)/bin/activate
+        poetry run python3 run.py --engines "${{ inputs.engine }}" --datasets "${{ inputs.dataset }}"

.github/workflows/manual-all-engines-benchmark.yaml

@@ -0,0 +1,138 @@
+name: Manual All Engines Default Benchmarks
+
+on:
+  push:
+    branches:
+     - "master"
+  pull_request:
+    types:
+      - opened
+      - reopened
+  workflow_dispatch:
+
+jobs:
+  elasticsearchBenchmark:
+    if: >
+      (
+      startsWith(github.event.head_commit.modified, 'engine/clients/elasticsearch') ||
+      startsWith(github.event.head_commit.modified, 'engine/servers/elasticsearch') ||
+      startsWith(github.event.head_commit.modified, 'engine/base_client/')
+      )
+    name: benchmark - elasticsearch-default - random-100 - against elasticsearch-single-node-ci
+    runs-on: ubuntu-latest
+    timeout-minutes: 30
+    steps:
+      - uses: actions/checkout@v3
+      - uses: ./.github/workflows/actions/run-engine-benchmark
+        with:
+          engine: "elasticsearch-default"
+          dataset: "random-100"
+          compose_file: "engine/servers/elasticsearch-single-node-ci/docker-compose.yaml"
+
+  milvusBenchmark:
+    if: >
+      (
+      startsWith(github.event.head_commit.modified, 'engine/clients/milvus') ||
+      startsWith(github.event.head_commit.modified, 'engine/servers/milvus') ||
+      startsWith(github.event.head_commit.modified, 'engine/base_client/')
+      )
+    name: benchmark - milvus-default - random-100 - against milvus-single-node
+    runs-on: ubuntu-latest
+    timeout-minutes: 30
+    steps:
+      - uses: actions/checkout@v3
+      - uses: ./.github/workflows/actions/run-engine-benchmark
+        with:
+          engine: "milvus-default"
+          dataset: "random-100"
+          compose_file: "engine/servers/milvus-single-node/docker-compose.yaml"
+
+  opensearchBenchmark:
+    if: >
+      (
+      startsWith(github.event.head_commit.modified, 'engine/clients/opensearch') ||
+      startsWith(github.event.head_commit.modified, 'engine/servers/opensearch') ||
+      startsWith(github.event.head_commit.modified, 'engine/base_client/')
+      )
+    name: benchmark - opensearch-default - glove-25-angular - against opensearch-single-node-ci
+    runs-on: ubuntu-latest
+    timeout-minutes: 30
+    steps:
+      - uses: actions/checkout@v3
+      - uses: ./.github/workflows/actions/run-engine-benchmark
+        with:
+          engine: "opensearch-default"
+          dataset: "glove-25-angular"
+          compose_file: "engine/servers/opensearch-single-node-ci/docker-compose.yaml"
+
+  pgvectorBenchmark:
+    if: >
+      (
+      startsWith(github.event.head_commit.modified, 'engine/clients/pgvector') ||
+      startsWith(github.event.head_commit.modified, 'engine/servers/pgvector') ||
+      startsWith(github.event.head_commit.modified, 'engine/base_client/')
+      )
+    name: benchmark - pgvector-default - random-100 - against pgvector-single-node
+    runs-on: ubuntu-latest
+    timeout-minutes: 30
+    steps:
+      - uses: actions/checkout@v3
+      - uses: ./.github/workflows/actions/run-engine-benchmark
+        with:
+          engine: "pgvector-default"
+          dataset: "random-100"
+          compose_file: "engine/servers/pgvector-single-node/docker-compose.yaml"
+
+  qdrantBenchmark:
+    if: >
+      (
+      startsWith(github.event.head_commit.modified, 'engine/clients/qdrant') ||
+      startsWith(github.event.head_commit.modified, 'engine/servers/qdrant') ||
+      startsWith(github.event.head_commit.modified, 'engine/base_client/')
+      )
+    name: benchmark - qdrant-default - random-100 - against qdrant-single-node
+    runs-on: ubuntu-latest
+    timeout-minutes: 30
+    steps:
+      - uses: actions/checkout@v3
+      - uses: ./.github/workflows/actions/run-engine-benchmark
+        with:
+          engine: "qdrant-default"
+          dataset: "random-100"
+          compose_file: "engine/servers/qdrant-single-node/docker-compose.yaml"
+
+  redisBenchmark:
+    if: >
+      (
+      startsWith(github.event.head_commit.modified, 'engine/clients/redis') ||
+      startsWith(github.event.head_commit.modified, 'engine/servers/redis') ||
+      startsWith(github.event.head_commit.modified, 'engine/base_client/')
+      )
+    name: benchmark - redis-default - random-100 - against redis-single-node
+    runs-on: ubuntu-latest
+    timeout-minutes: 30
+    steps:
+      - uses: actions/checkout@v3
+      - uses: ./.github/workflows/actions/run-engine-benchmark
+        with:
+          engine: "redis-default"
+          dataset: "random-100"
+          compose_file: "engine/servers/redis-single-node/docker-compose.yaml"
+
+  weaviateBenchmark:
+    if: >
+      (
+      startsWith(github.event.head_commit.modified, 'engine/clients/weaviate') ||
+      startsWith(github.event.head_commit.modified, 'engine/servers/weaviate') ||
+      startsWith(github.event.head_commit.modified, 'engine/base_client/')
+      )
+    name: benchmark - weaviate-default - random-100 - against weaviate-single-node
+    runs-on: ubuntu-latest
+    timeout-minutes: 30
+    steps:
+      - uses: actions/checkout@v3
+      - uses: ./.github/workflows/actions/run-engine-benchmark
+        with:
+          engine: "weaviate-default"
+          dataset: "random-100"
+          compose_file: "engine/servers/weaviate-single-node/docker-compose.yaml"

engine/servers/elasticsearch-single-node-ci/docker-compose.yaml

@@ -0,0 +1,23 @@
+version: '3.5'
+
+services:
+  es:
+    image: docker.elastic.co/elasticsearch/elasticsearch:8.10.2
+    environment:
+      ELASTIC_PASSWORD: "passwd"
+      KIBANA_PASSWORD: "passwd"
+      SERVER_SSL_ENABLED: "false"
+      discovery.type: "single-node"
+      xpack.security.enabled: "false"
+    ports:
+      - "9200:9200"
+      - "9300:9300"
+    logging:
+      driver: "json-file"
+      options:
+        max-file: 1
+        max-size: 10m
+    deploy:
+      resources:
+        limits:
+          memory: 4Gb

engine/servers/opensearch-single-node-ci/docker-compose.yaml

@@ -0,0 +1,21 @@
+version: '3.5'
+
+services:
+  opensearch:
+    image: opensearchproject/opensearch:2.10.0
+    environment:
+      discovery.type: "single-node"
+      plugins.security.disabled: true
+      OPENSEARCH_JAVA_OPTS: "-Xms2g -Xmx2g"
+    ports:
+      - "9200:9200"
+      - "9300:9300"
+    logging:
+      driver: "json-file"
+      options:
+        max-file: 1
+        max-size: 10m
+    deploy:
+      resources:
+        limits:
+          memory: 4Gb

tools/wait_for_green_status.sh

@@ -0,0 +1,34 @@
+#!/bin/bash
+# This scripts helps to wait for Opensearch|Elasticsearch status to become Green
+
+set -e
+
+SEARCH_CLUSTER_HOST=${1:-"localhost:9200"}
+
+# Wait until the search cluster host is available
+until $(curl --output /dev/null --silent --head --fail "$SEARCH_CLUSTER_HOST"); do
+    printf '.'
+    sleep 1 # Wait for 1 second
+done
+
+# Wait for ES/OS to start
+response=$(curl --write-out %{http_code} --silent --output /dev/null "$SEARCH_CLUSTER_HOST")
+
+until [ "$response" = "200" ]; do
+    response=$(curl --write-out %{http_code} --silent --output /dev/null "$SEARCH_CLUSTER_HOST")
+    >&2 echo "Search cluster is unavailable - sleep 1s"
+    sleep 1
+done
+
+# Wait for ES/OS status to turn Green
+health="$(curl -fsSL "$SEARCH_CLUSTER_HOST/_cat/health?h=status")"
+health="$(echo "$health" | sed -r 's/^[[:space:]]+|[[:space:]]+$//g')"
+
+until [ "$health" = 'green' ]; do
+    health="$(curl -fsSL "$SEARCH_CLUSTER_HOST/_cat/health?h=status")"
+    health="$(echo "$health" | sed -r 's/^[[:space:]]+|[[:space:]]+$//g')"
+    >&2 echo "Search cluster status is not green yet - sleep 1s"
+    sleep 1
+done
+
+>&2 echo "Search cluster is up"