create a trino plugin to get slice info from vhost guid via registry

.circleci/config.yml

@@ -0,0 +1,85 @@
+version: 2.1
+
+orbs:
+  aws-ecr: circleci/aws-ecr@8.2.1
+
+jobs:
+  test:
+    docker:
+      - image: cimg/openjdk:24.0.2
+    steps:
+      - checkout
+      - run:
+          name: test with coverage
+          command: |
+            cd plugin/trino-af/
+            ../../mvnw clean test -Djacoco.skip=false -Djava.vendor="Eclipse Adoptium"
+      - store_test_results:
+          path: plugin/trino-af/target/surefire-reports
+      - store_artifacts:
+          path: plugin/trino-af/target/site/jacoco
+          destination: coverage-report
+      - run:
+          name: verify coverage summary
+          command: |
+            cd plugin/trino-af/
+            if [ -f target/site/jacoco/index.html ]; then
+              echo "Coverage report generated successfully"
+              echo "View coverage report in CircleCI artifacts"
+            else
+              echo "Coverage report generation failed"
+              exit 1
+            fi
+
+  build-and-push:
+    docker:
+      - image: cimg/openjdk:24.0.2
+    steps:
+      - checkout
+      - run:
+          name: build
+          command: |
+            cd plugin/trino-af/
+            ../../mvnw clean package -Djava.vendor="Eclipse Adoptium"
+      - run:
+          name: set environment variables
+          command: |
+            echo 'export REGISTRY_ACCOUNT_ID="307946643675"' >> "$BASH_ENV"
+            source "$BASH_ENV"
+
+            SHORT_COMMIT_SHA=${CIRCLE_SHA1:0:7}
+            echo "export SHORT_COMMIT_SHA='${SHORT_COMMIT_SHA}'" >> $BASH_ENV
+
+            TAG=${CIRCLE_BRANCH//\//-}-${SHORT_COMMIT_SHA}
+            echo "export TAG='${TAG}'" >> $BASH_ENV
+
+            DOCKER_HOST_BLOCK=$(curl -fsSL http://npm-yarn-blocklist.security.appf.io | tr ' ' '\n' | sed 's/^/--add-host=/' | sed 's/$/:127.0.0.1/' | tr '\n' ' ' | sed 's/ $//')
+            echo "export DOCKER_HOST_BLOCK='${DOCKER_HOST_BLOCK}'" >> $BASH_ENV
+
+            echo "ACCESS KEY ID: ${DATA_API_MACHINE_USER_AWS_ACCESS_KEY_ID}"
+      - setup_remote_docker:
+          docker_layer_caching: true
+      - aws-ecr/build-and-push-image:
+          registry-id: REGISTRY_ACCOUNT_ID
+          aws-access-key-id: DATA_API_MACHINE_USER_AWS_ACCESS_KEY_ID
+          aws-secret-access-key: DATA_API_MACHINE_USER_AWS_SECRET_ACCESS_KEY
+          region: us-west-2
+          create-repo: false
+          repo: data-api/trino-af
+          build-path: plugin/trino-af/
+          path: plugin/trino-af/
+          tag: ${TAG}
+          platform: linux/amd64,linux/arm64
+          skip-when-tags-exist: true
+          extra-build-args: ${DOCKER_HOST_BLOCK}
+
+
+workflows:
+  version: 2
+  test-build-push:
+    jobs:
+      - test
+      - build-and-push:
+          requires:
+            - test
+          context: data-api-pipeline-trigger-group

plugin/trino-af/Dockerfile

@@ -0,0 +1,4 @@
+FROM trinodb/trino:476
+
+# Copy the trino-af plugin files
+COPY --chown=trino:trino target/trino-af-476/ /usr/lib/trino/plugin/trino-af

plugin/trino-af/README.md

@@ -0,0 +1,112 @@
+# Guide for trino-af Plugin
+The `trino-af` plugin implements the `af_find_datasource_by_guid` function which receives a vhost guid and then calls [the registry app](https://github.com/appfolio/registry_app) to get the slice the vhost is active on.
+
+Vhost databases with the same guid can exist on multiple slices. The function can be used in Trino's row filter to make sure Trino only return rows belong to the slice a vhost is active on.
+
+# Development
+
+## Install dependencies
+
+```commandline
+brew install trino jenv
+```
+
+## Install Java
+
+Trino works with Temurin JDK 24.
+```commandline
+brew install temurin
+jenv add /Library/Java/JavaVirtualMachines/temurin-24.jdk/Contents/Home
+```
+
+## Create a branch
+
+### Pull all tags
+
+May need to run the following if new trino verions are released.
+
+```commandline
+git remote add upstream https://github.com/trinodb/trino.git
+git fetch upstream --tags
+git push origin --tags
+```
+
+### Check out the Branch
+
+```commandline
+git checkout 476-af
+```
+
+All development work happen in this branch.
+
+## Test
+
+Run automated tests:
+
+```commandline
+cd plugin/trino-af/
+../../mvnw test
+```
+
+Start a test server for manual testing:
+
+```commandline
+cd plugin/trino-af
+../../mvnw compile test-compile exec:java
+```
+
+Access the test server:
+
+```commandline
+trino --server http://localhost:8080 --execute "select af.af.af_find_datasource_by_guid('test-guid-1')"
+```
+
+### Code Coverage
+
+JaCoCo code coverage report is generated on CircleCI.
+
+To manually generate coverage reports:
+
+```commandline
+cd plugin/trino-af/
+mvn clean test -Djacoco.skip=false -Djava.vendor="Eclipse Adoptium"
+```
+
+After running, view the coverage report:
+
+```commandline
+open target/site/jacoco/index.html
+```
+
+The coverage report shows:
+- **Instruction Coverage**: Percentage of bytecode instructions executed
+- **Branch Coverage**: Percentage of conditional branches tested
+- **Line Coverage**: Percentage of source code lines executed
+- **Method Coverage**: Percentage of methods called
+- **Class Coverage**: Percentage of classes instantiated
+
+## Build
+
+The plugin is built on CircleCI.
+
+To manually build the plugin:
+
+```commandline
+cd plugin/trino-af/
+../../mvnw package
+```
+
+`target/trino-af-476.zip` is the plugin file
+
+## Push
+
+CircleCI pushes a Trino container image with the plugin installed to ECR in the data-api-prod account.
+
+To manually push an image:
+
+```commandline
+awsume data-api-prod
+./build-and-push.sh
+```
+
+The output should have the image tag which can be used in Kubernetes Trino setup.

plugin/trino-af/build-and-push.sh

@@ -0,0 +1,36 @@
+#!/bin/bash
+set -e
+
+TRINO_BASE_VERSION=476
+TRINO_AF_REPO=data-api/trino-af
+
+echo "Step 1: Building Maven package..."
+../../mvnw clean package
+
+echo "Step 2: Finding ECR repository for trino-af..."
+ECR_REPO_URI=$(aws ecr describe-repositories --repository-names $TRINO_AF_REPO --query 'repositories[0].repositoryUri' --output text)
+if [ "$ECR_REPO_URI" == "None" ] || [ -z "$ECR_REPO_URI" ]; then
+    echo "Error: ECR repository 'trino-af' not found"
+    exit 1
+fi
+echo "Found ECR repository: $ECR_REPO_URI"
+
+echo "Step 3: Finding latest version tag..."
+LATEST_VERSION=$(aws ecr list-images --repository-name $TRINO_AF_REPO --query 'imageIds[?imageTag!=`null`].imageTag' --output text | tr '\t' '\n' | grep -E '\-[0-9]+$' | sort -V | tail -1)
+if [ -z "$LATEST_VERSION" ]; then
+    NEW_VERSION="$TRINO_BASE_VERSION-1"
+else
+    # Extract the number after the dash and increment it
+    VERSION_NUMBER=$(echo "$LATEST_VERSION" | grep -oE '[0-9]+$')
+    NEW_VERSION="$TRINO_BASE_VERSION-$((VERSION_NUMBER + 1))"
+fi
+echo "Latest version: ${LATEST_VERSION:-none}, New version: $NEW_VERSION"
+
+echo "Step 4: Authenticating with ECR..."
+aws ecr get-login-password | docker login --username AWS --password-stdin $ECR_REPO_URI
+
+echo "Step 5: Building multi-architecture Docker image..."
+docker buildx create --use --name multiarch-builder 2>/dev/null || docker buildx use multiarch-builder
+docker buildx build --platform linux/arm64,linux/amd64 -t $ECR_REPO_URI:$NEW_VERSION --push .
+
+echo "Successfully built and pushed $ECR_REPO_URI:$NEW_VERSION"