airbyte/build.gradle

import com.bmuschko.gradle.docker.tasks.image.DockerBuildImage
import com.github.spotbugs.snom.SpotBugsTask

// The buildscript block defines dependencies in order for .gradle file evaluation.
// This is separate from application dependencies.
// See https://stackoverflow.com/questions/17773817/purpose-of-buildscript-block-in-gradle.
buildscript {
    repositories {
        maven {
            url "https://plugins.gradle.org/m2/"
        }
    }
    dependencies {
        classpath 'com.bmuschko:gradle-docker-plugin:8.0.0'
        // 6.x version of OpenApi generator is only compatible with jackson-core 2.13.x onwards.
        // This conflicts with the jackson depencneis the bmuschko plugin is pulling in.
        // Since api generation is only used in the airbyte-api module and the base gradle files
        // are loaded in first, Gradle is not able to intelligently resolve this before loading in
        // the bmuschko plugin and thus placing an older jackson version on the class path.
        // The alternative is to import the openapi plugin for all modules.
        // This might need to be updated when we change openapi plugin versions.
        classpath 'com.fasterxml.jackson.core:jackson-core:2.13.0'

        classpath 'org.codehaus.groovy:groovy-yaml:3.0.3'
    }
}

plugins {
    id 'base'
    id 'pmd'
    id 'com.diffplug.spotless' version '6.12.0'
    id 'com.github.hierynomus.license' version '0.16.1'
    id 'com.github.spotbugs' version '5.0.13'
    // The distribution plugin has been added to address the an issue with the copyGeneratedTar
    // task depending on "distTar". When that dependency has been refactored away, this plugin
    // can be removed.
    id 'distribution'
    id 'version-catalog'
    id 'maven-publish'
}

apply from: "$rootDir/publish-repositories.gradle"

repositories {
    mavenCentral()
    maven {
        url 'https://airbyte.mycloudrepo.io/public/repositories/airbyte-public-jars/'
    }
}

Properties env = new Properties()
rootProject.file('gradle.properties').withInputStream { env.load(it) }

if (!env.containsKey('VERSION')) {
    throw new Exception('Version not specified in .env file...')
}

// `version` is used as the application build version for artifacts like jars
// `image_tag` is used as the docker tag applied to built images.
// These values are the same for building an specific Airbyte release or branch via the 'VERSION' environment variable.
// For local development builds, the 'VERSION' environment variable is unset, and built images are tagged with 'dev'.
ext {
    version = System.getenv("VERSION") ?: env.VERSION
    image_tag = System.getenv("VERSION") ?: 'dev'
}

def createLicenseWith = { File license, String startComment, String endComment, String lineComment, boolean isPython ->
    /*
    In java, we don't have a second linter/styling tool other than spotless so it doesn't really
    matter if we write a newline or not for startComment/endComment.

    However, in python, we are using black that double-checks and reformats the code.
    Thus, writing an extra empty newline (not removed by trimTrailingWhitespace() is actually a
    big deal and would be reformatted (removed) because of black's specs.
    */
    def tmp = File.createTempFile('tmp', '.tmp')
    tmp.withWriter {
        def w = it
        if (startComment.length() > 0 || !isPython) {
            w.writeLine(startComment)
        }
        license.eachLine {
            w << lineComment
            w.writeLine(it)
        }
        if (endComment.length() > 0 || !isPython) {
            w.writeLine(endComment)
        }
        w.writeLine("")
        if (isPython) {
            w.writeLine("")
        }
    }
    return tmp
}

def createPythonLicenseWith = { license ->
    return createLicenseWith(license, '', '', '', true)
}

def createJavaLicenseWith = { license ->
    return createLicenseWith(license, '/*', ' */', ' * ', false)
}

// We are the spotless exclusions rules using file tree. It seems the excludeTarget option is super finicky in a
// monorepo setup and it doesn't actually exclude directories reliably. This code makes the behavior predictable.
def createSpotlessTarget = { pattern ->
    def excludes = [
            '.gradle',
            'node_modules',
            '.eggs',
            '.mypy_cache',
            '.venv',
            '*.egg-info',
            'build',
            'dbt-project-template',
            'dbt-project-template-mssql',
            'dbt-project-template-mysql',
            'dbt-project-template-oracle',
            'dbt-project-template-clickhouse',
            'dbt-project-template-snowflake',
            'dbt-project-template-tidb',
            'dbt-project-template-duckdb',
            'dbt_test_config',
            'normalization_test_output',
            'tools',
            'secrets',
            'charts', // Helm charts often have injected template strings that will fail general linting. Helm linting is done separately.
            'resources/seed/*_catalog.json', // Do not remove - this is also necessary to prevent diffs in our github workflows
            'resources/seed/*_registry.json', // Do not remove - this is also necessary to prevent diffs in our github workflows
            'airbyte-integrations/connectors/source-amplitude/unit_tests/api_data/zipped.json', // Zipped file presents as non-UTF-8 making spotless sad
            'airbyte-webapp', // The webapp module uses its own auto-formatter, so spotless is not necessary here
            'airbyte-webapp-e2e-tests', // This module also uses its own auto-formatter
            'airbyte-connector-builder-server/connector_builder/generated', // autogenerated code doesn't need to be formatted
    ]

    if (System.getenv().containsKey("SUB_BUILD")) {
        excludes.add("airbyte-integrations/connectors")
    }

    return fileTree(dir: rootDir, include: pattern, exclude: excludes.collect { "**/${it}" })
}

spotless {
    java {
        target createSpotlessTarget('**/*.java')

        importOrder()

        eclipse('4.21.0').configFile(rootProject.file('tools/gradle/codestyle/java-google-style.xml'))

        licenseHeaderFile createJavaLicenseWith(rootProject.file('LICENSE_SHORT'))
        removeUnusedImports()
        trimTrailingWhitespace()
    }
    groovyGradle {
        target createSpotlessTarget('**/*.gradle')
    }
    sql {
        target createSpotlessTarget('**/*.sql')

        dbeaver().configFile(rootProject.file('tools/gradle/codestyle/sql-dbeaver.properties'))
    }
    format 'styling', {
        target createSpotlessTarget(['**/*.yaml', '**/*.json'])

        prettier()
    }
}
check.dependsOn 'spotlessApply'

@SuppressWarnings('GroovyAssignabilityCheck')
def Task getPublishArtifactsTask(String buildVersion, Project subproject) {
    // generate a unique task name based on the directory name.
    return task("publishArtifact-$subproject.name" {
        apply plugin: 'maven-publish'
        publishing {
            repositories {
                publications {
                    "$subproject.name"(MavenPublication) {
                        from subproject.components.java

                        // use the subproject group and name with the assumption there are no identical subproject
                        // names, group names or subproject group/name combination.
                        groupId = "$subproject.group"
                        artifactId = "$subproject.name"
                        version = "$buildVersion"
                        repositories.add(rootProject.repositories.getByName('cloudrepo'))
                    }
                }
            }
        }
    })
}

allprojects {
    apply plugin: 'com.bmuschko.docker-remote-api'

    task copyDocker(type: Sync) {
        from "${project.projectDir}/Dockerfile"
        into "build/docker/"
    }
}

allprojects {
    apply plugin: 'base'

    // by default gradle uses directory as the project name. That works very well in a single project environment but
    // projects clobber each other in an environments with subprojects when projects are in directories named identically.
    def sub = rootDir.relativePath(projectDir.parentFile).replace('/', '.')
    group = "io.${rootProject.name}${sub.isEmpty() ? '' : ".$sub"}"
    project.archivesBaseName = "${project.group}-${project.name}"

    version = rootProject.ext.version
}

// Java projects common configurations
subprojects { subproj ->


    configurations {
        runtimeClasspath
    }


    // Common Docker Configuration:
    // If subprojects have the dockerImageName property configured in their gradle.properties file,
    // register:
    // 1) A copyGeneratedTar task to copy generated TARs. Subprojects that produce TARs can use this
    //    to copy the produced tar into the docker image.
    // 2) A buildDockerImage task to build the docker image.
    // 3) Make the docker image task depend on the default assemble task.
    if (subproj.hasProperty('dockerImageName')) {
        project.logger.lifecycle("configuring docker task for $subproj.name")

        // Although this task is defined for every subproject with the dockerImageName property,
        // It is not necessarily used for all subprojects. Non-TAR producing projects can ignore this.
        tasks.register("copyGeneratedTar", Copy) {
            dependsOn copyDocker
            dependsOn distTar
            from('build/distributions') {
                // Assume that tars are named <parent-project-name>-<child-project-name>-*.tar.
                // Because we only have a handle to the child project, and to keep things simple,
                // use a * regex to catch all prefixes.
                include "*$subproj.name-*.tar"
            }
            into 'build/docker/bin'
        }

        tasks.register("buildDockerImage", DockerBuildImage) {
            // This is currently only used for connectors.
            def jdkVersion = System.getenv('JDK_VERSION') ?: '17.0.4'

            def arch = System.getenv('BUILD_ARCH') ?: System.getProperty("os.arch").toLowerCase()
            def isArm64 = arch == "aarch64" || arch == "arm64"
            def buildArch = System.getenv('DOCKER_BUILD_ARCH') ?: isArm64 ? 'arm64' : 'amd64'
            def buildPlatform = System.getenv('DOCKER_BUILD_PLATFORM') ?: isArm64 ? 'linux/arm64' : 'linux/amd64'
            def alpineImage = System.getenv('ALPINE_IMAGE') ?: isArm64 ? 'arm64v8/alpine:3.14' : 'amd64/alpine:3.14'
            def nginxImage = System.getenv('NGINX_IMAGE') ?: isArm64 ? 'arm64v8/nginx:alpine' : 'amd64/nginx:alpine'

            // Used by the platform -- Must be an Amazon Corretto-based image for build to work without modification to Dockerfile instructions
            def openjdkImage = System.getenv('JDK_IMAGE') ?: "airbyte/airbyte-base-java-image:1.0"

            platform = buildPlatform
            images.add("airbyte/$subproj.dockerImageName:$rootProject.ext.image_tag")
            buildArgs.put('JDK_VERSION', jdkVersion)
            buildArgs.put('DOCKER_BUILD_ARCH', buildArch)
            buildArgs.put('ALPINE_IMAGE', alpineImage)
            buildArgs.put('NGINX_IMAGE', nginxImage)
            buildArgs.put('JDK_IMAGE', openjdkImage)
            buildArgs.put('VERSION', rootProject.ext.version)

        }

        tasks.named("assemble") {
            dependsOn buildDockerImage
        }
    }

    if (subproj.name == 'airbyte-webapp' || subproj.name == 'airbyte-webapp-e2e-tests') {
        return
    }

    apply plugin: 'java'
    apply plugin: 'jacoco'
    apply plugin: 'com.github.spotbugs'
    apply plugin: 'pmd'

    sourceCompatibility = JavaVersion.VERSION_17
    targetCompatibility = JavaVersion.VERSION_17

    repositories {
        mavenCentral()
        maven {
            url 'https://jitpack.io'
        }
        maven {
            url 'https://airbyte.mycloudrepo.io/public/repositories/airbyte-public-jars/'
        }
    }

    pmd {
        consoleOutput = true
        ignoreFailures = false
        rulesMinimumPriority = 5
        ruleSets = []
        ruleSetFiles = files(rootProject.file('tools/gradle/pmd/rules.xml'))
        toolVersion = '6.51.0'
    }

    jacoco {
        toolVersion = "0.8.8"
    }

    jacocoTestReport {
        dependsOn test
        reports {
            html.required = true
            xml.required = true
            csv.required = false
        }
    }

    jacocoTestCoverageVerification {
        violationRules {
            failOnViolation = false
            rule {
                element = 'CLASS'
                excludes = ['**/*Test*', '**/generated*']
                limit {
                    counter = 'BRANCH'
                    minimum = 0.8
                }
                limit {
                    counter = 'INSTRUCTION'
                    minimum = 0.8
                }
            }
        }
    }

    // make tag accessible in submodules.
    ext {
        cloudStorageTestTagName = 'cloud-storage'
        numberThreads = project.hasProperty('numberThreads') ? project.getProperty('numberThreads') as int : Runtime.runtime.availableProcessors() ?: 1
    }

    spotbugs {
        ignoreFailures = false
        effort = 'max'
        excludeFilter = rootProject.file('spotbugs-exclude-filter-file.xml')
        reportLevel = 'high'
        showProgress = false
        toolVersion = '4.7.3'
    }

    test {
        //This allows to set up a `gradle.properties` file inside the connector folder to reduce number of threads and reduce parallelization.
        //Specially useful for connectors that shares resources (like Redshift or Snowflake).
        maxParallelForks = numberThreads
        jacoco {
            enabled = true
            excludes = ['**/*Test*', '**/generated*']
        }
        useJUnitPlatform {
            excludeTags(cloudStorageTestTagName)
        }
        testLogging() {
            events "passed", "skipped", "failed"
            exceptionFormat 'full'
            // uncomment to get the full log output
            // showStandardStreams = true
        }
        finalizedBy jacocoTestReport
    }

    task allTests(type: Test) {
        useJUnitPlatform()
        testLogging() {
            events "passed", "failed", "started"
            exceptionFormat 'full'
            // uncomment to get the full log output
            // showStandardStreams = true
        }
        finalizedBy jacocoTestReport
    }

    dependencies {
        if (subproj.name != 'airbyte-commons') {
            implementation project(':airbyte-commons')
        }

        implementation(platform("com.fasterxml.jackson:jackson-bom:2.13.0"))
        implementation(platform("org.glassfish.jersey:jersey-bom:2.31"))


        // version is handled by "com.fasterxml.jackson:jackson-bom:2.10.4", so we do not explicitly set it here.
        implementation libs.bundles.jackson

        implementation libs.guava

        implementation libs.commons.io

        implementation libs.bundles.apache

        implementation libs.slf4j.api


        // SLF4J as a facade over Log4j2 required dependencies
        implementation libs.bundles.log4j

        // Bridges from other logging implementations to SLF4J
        implementation libs.bundles.slf4j

        // Dependencies for logging to cloud storage, as well as the various clients used to do so.
        implementation libs.appender.log4j2
        implementation libs.aws.java.sdk.s3
        implementation libs.google.cloud.storage

        implementation libs.s3

        // Lombok dependencies
        compileOnly libs.lombok
        annotationProcessor libs.lombok

        testCompileOnly libs.lombok
        testAnnotationProcessor libs.lombok

        testRuntimeOnly libs.junit.jupiter.engine
        testImplementation libs.bundles.junit
        testImplementation libs.assertj.core

        testImplementation libs.junit.pioneer

        // adds owasp plugin
        spotbugsPlugins libs.findsecbugs.plugin
        implementation libs.spotbugs.annotations
    }

    tasks.withType(Tar) {
        duplicatesStrategy DuplicatesStrategy.INCLUDE
    }

    tasks.withType(Zip) {
        duplicatesStrategy DuplicatesStrategy.INCLUDE
    }

    tasks.withType(SpotBugsTask) {
        // Reports can be found under each subproject in build/spotbugs/
        reports {
            xml.required = false
            html.required = true
        }
    }

    tasks.withType(Pmd) {
        exclude '**/generated/**'
        exclude '**/jooq/**'
    }

    javadoc.options.addStringOption('Xdoclint:none', '-quiet')
    check.dependsOn 'jacocoTestCoverageVerification'
}

task('generate') {
    dependsOn subprojects.collect { it.getTasksByName('generateComponentManifestClassFiles', true) }
    dependsOn subprojects.collect { it.getTasksByName('generateJsonSchema2Pojo', true) }
}

task('format') {
    dependsOn generate
    dependsOn spotlessApply
    dependsOn subprojects.collect { it.getTasksByName('airbytePythonFormat', true) }
}

// add licenses for python projects.
subprojects {
    def pythonFormatTask = project.tasks.findByName('blackFormat')

    if (pythonFormatTask != null) {
        apply plugin: "com.github.hierynomus.license"
        license {
            header rootProject.file("LICENSE_SHORT")
        }
        task licenseFormatPython(type: com.hierynomus.gradle.license.tasks.LicenseFormat) {
            header = createPythonLicenseWith(rootProject.file('LICENSE_SHORT'))
            source = fileTree(dir: projectDir)
                    .include("**/*.py")
                    .exclude(".venv/**/*.py")
                    .exclude("**/airbyte_api_client/**/*.py")
                    .exclude("**/__init__.py")
            strictCheck = true
        }
        def licenseTask = project.tasks.findByName('licenseFormatPython')
        def isortFormatTask = project.tasks.findByName('isortFormat')
        if (isortFormatTask != null) {
            isortFormat.dependsOn licenseTask
        }
        def flakeCheckTask = project.tasks.findByName('flakeCheck')
        if (flakeCheckTask != null) {
            flakeCheck.dependsOn licenseTask
        }
        blackFormat.dependsOn licenseTask

        def generateManifestFilesTask = project.tasks.findByName('generateComponentManifestClassFiles')
        if (generateManifestFilesTask != null) {
            licenseTask.dependsOn generateManifestFilesTask
        }
    }
    task listAllDependencies(type: DependencyReportTask) {}
}

// produce reproducible archives
// (see https://docs.gradle.org/current/userguide/working_with_files.html#sec:reproducible_archives)
tasks.withType(AbstractArchiveTask) {
    preserveFileTimestamps = false
    reproducibleFileOrder = true
}

// definition for publishing
catalog {
    versionCatalog {
        from(files("deps.toml"))
    }
}

publishing {
    publications {
        // This block is present for dependency catalog publishing.
        maven(MavenPublication) {
            groupId = 'io.airbyte'
            artifactId = 'oss-catalog'

            from components.versionCatalog
            // Gradle will by default use the subproject path as the group id and the subproject name as the artifact id.
            // e.g. the subproject :airbyte-example:example-models is imported at io.airbyte.airbyte-config-oss:config-persistence:<version-number>.
        }
    }
    repositories.add(rootProject.repositories.getByName('cloudrepo'))
}