build.gradle

import javax.inject.Inject
import org.apache.tools.ant.taskdefs.condition.Os
import org.gradle.process.ExecOperations

buildscript {
    repositories {
        mavenLocal()
        mavenCentral()
        maven {
            url "https://plugins.gradle.org/m2/"
        }
        maven {
            name = 'ajoberstar-backup'
            url = 'https://ajoberstar.org/bintray-backup/'
        }
    }
    dependencies {
        classpath group: 'org.yaml', name: 'snakeyaml', version: '1.19'
    }
}

plugins {
    id 'java'
    id 'application'
    id 'maven-publish'
    id 'jacoco'
    id 'com.gradleup.shadow' version '8.3.10'
    id "de.undercouch.download" version "4.1.1"
    id 'net.researchgate.release' version '3.1.0'
    id "org.jetbrains.kotlin.jvm" version "2.2.20"
}

group = "org.grobid"

description = """The goal of this GROBID module is to recognize in textual documents any expressions of 
measurements (e.g. pressure, temperature, etc.), to parse and normalization them, and finally to convert 
these measurements into SI units. We focus our work on technical and scientific articles (text, XML and PDF input) 
and patents (text and XML input)."""

// Deliberately *not* --first-parent: a release made on a branch and merged back with a merge
// commit leaves its tag on the second parent, so --first-parent walks straight past it and
// reports the previous release. After releasing v0.9.0 on a branch and merging, master would
// describe as v0.8.2-2-g<sha> instead of v0.9.0-4-g<sha>.
// --always keeps a bare commit sha as the fallback when no tag is reachable, e.g. in the
// shallow clones CI checks out.
def getGitRevision() {
    def gitRevision = "unknown"
    try {
        def result = providers.exec {
            workingDir = rootProject.rootDir
            commandLine 'git', 'describe', '--tags', '--always'
        }
        gitRevision = result.standardOutput.asText.get().trim()
    } catch (Exception e) {
        println "Could not get Git revision: ${e}"
    }
    return gitRevision
}

project.ext.gitRevision = getGitRevision()

tasks.register('collectGitRevision') {
    project.ext.gitRevision = getGitRevision()
    doLast {
        println "Git revision: ${project.ext.gitRevision}"
    }
}

tasks.withType(JavaCompile).configureEach {
    options.encoding = 'UTF-8'
    //    options.compilerArgs << "-Xlint:deprecation"
    //    options.compilerArgs << "-Xlint:unchecked"
}


kotlin {
    jvmToolchain(21)
}

java {
    toolchain {
        languageVersion.set(JavaLanguageVersion.of(21))
    }
}


tasks.run.workingDir = rootProject.rootDir

ext {
    // treating them separately, these jars will be flattened into grobid-core.jar on installing,
    // to avoid missing dependencies from the projects that include grobid-core (see 'jar' taskin grobid-core)
    localLibs = []
}

repositories {
    mavenLocal()
    mavenCentral()
    maven { url 'https://jitpack.io' }
    maven { url "https://oss.jfrog.org/artifactory/libs-snapshot" }
    maven { url "https://oss.jfrog.org/artifactory/libs-release" }
    flatDir {
        dirs 'localLibs'
    }
}


dependencies {
    //Tests
    testImplementation(platform('org.junit:junit-bom:5.14.1'))
    testRuntimeOnly("org.junit.platform:junit-platform-launcher") {
        because("Only needed to run tests in a version of IntelliJ IDEA that bundles older versions")
    }
    testRuntimeOnly("org.junit.jupiter:junit-jupiter-engine")
    testRuntimeOnly("org.junit.vintage:junit-vintage-engine")
    testImplementation('org.junit.jupiter:junit-jupiter')

    testImplementation 'org.easymock:easymock:5.6.0'
    testImplementation 'org.hamcrest:hamcrest-all:1.3'
    testImplementation 'org.mockito:mockito-core:5.14.2'
    // JUnit 4 API for compilation (was previously provided transitively by
    // powermock-module-junit4); JUnit 4 tests are executed by junit-vintage-engine.
    testImplementation 'junit:junit:4.13.2'
    testImplementation 'org.jetbrains.kotlin:kotlin-test'
    testImplementation "io.mockk:mockk:1.13.17"

    //GROBID
    implementation 'org.grobid:grobid-core:0.9.1'
    implementation 'org.grobid:grobid-trainer:0.9.1'
    implementation 'org.grobid:grobid-service:0.9.1'
    implementation 'xerces:xercesImpl:2.12.2'
    implementation "net.arnx:jsonic:1.3.10"
    implementation "net.sf.saxon:Saxon-HE:9.6.0-9"
    implementation "xom:xom:1.3.2"
    implementation 'javax.xml.bind:jaxb-api:2.3.0'
    // Must match the opennlp-tools version Grobid 0.9.1 is built against (2.5.9):
    // grobid-core calls SentenceDetectorME.sentPosDetect(CharSequence), an OpenNLP 2.x
    // overload. A lower direct pin here overrides the transitive 2.5.9 and causes a
    // runtime NoSuchMethodError during sentence segmentation.
    implementation 'org.apache.opennlp:opennlp-tools:2.5.9'
    implementation 'black.ninia:jep:4.3.1'
    implementation "org.apache.httpcomponents:httpclient:4.5.14"
    implementation "org.apache.lucene:lucene-analyzers-common:4.5.1"
    implementation group: 'org.jruby', name: 'jruby-complete', version: '9.2.13.0'

    //Apache commons
    implementation 'org.apache.commons:commons-collections4:4.5.0'
    implementation 'org.apache.commons:commons-lang3:3.20.0'
    implementation 'commons-logging:commons-logging:1.2'
    implementation 'commons-io:commons-io:2.21.0'
    implementation 'commons-pool:commons-pool:1.6'
    implementation 'commons-codec:commons-codec:1.14'

    //Json
    implementation 'com.fasterxml.jackson.core:jackson-core:2.21.1'
    implementation 'com.fasterxml.jackson.core:jackson-databind:2.21.1'
    implementation 'com.fasterxml.jackson.module:jackson-module-afterburner:2.21.1'
    implementation 'com.fasterxml.jackson.dataformat:jackson-dataformat-yaml:2.21.1'

    // measurements

    // (1) UOM 2.x
    implementation 'tech.units:indriya:2.1.3'
    implementation group: 'si.uom', name: 'si-units', version: '2.1'
    implementation group: 'si.uom', name: 'si-quantity', version: '2.1'
    implementation group: 'systems.uom', name: 'systems-quantity', version: '2.1'
    implementation group: 'systems.uom', name: 'systems-common', version: '2.1'
    implementation group: 'systems.uom', name: 'systems-unicode', version: '2.1'
    implementation group: 'systems.uom', name: 'systems-ucum', version: '2.1'
    implementation group: 'systems.uom', name: 'systems-unicode', version: '2.1'

    //Dropwizard 5.x (Jakarta EE 10 / Jetty 12) to match Grobid 0.9.1. guicey 8.x is the
    // Dropwizard-5-compatible line. Jetty 12 removed the org.eclipse.jetty.servlets
    // CrossOriginFilter/QoSFilter; the service now uses CrossOriginHandler/QoSHandler.
    implementation 'ru.vyarus:dropwizard-guicey:8.0.2'

    // Dropwizard 5.0.1 ships Jetty 12.1.5; force the Jetty BOMs to 12.1.9 (CVE-2026-2332).
    implementation platform('org.eclipse.jetty:jetty-bom:12.1.9')
    implementation platform('org.eclipse.jetty.ee10:jetty-ee10-bom:12.1.9')

    implementation 'io.dropwizard:dropwizard-bom:5.0.1'
    implementation 'io.dropwizard:dropwizard-core:5.0.1'
    implementation 'io.dropwizard:dropwizard-assets:5.0.1'
    implementation 'io.dropwizard:dropwizard-testing:5.0.1'
    implementation 'io.dropwizard:dropwizard-forms:5.0.1'
    implementation 'io.dropwizard:dropwizard-client:5.0.1'
    implementation 'io.dropwizard:dropwizard-auth:5.0.1'
    implementation 'io.dropwizard.metrics:metrics-core:4.2.38'
    implementation 'io.dropwizard.metrics:metrics-servlets:4.2.38'

    //Misc
    implementation 'com.googlecode.clearnlp:clearnlp:1.3.1'
    implementation 'com.google.guava:guava:33.5.0-jre'
    implementation "net.arnx:jsonic:1.3.10"

    // XML
    implementation 'org.codehaus.woodstox:stax2-api:4.2.1'
    implementation 'com.fasterxml.woodstox:woodstox-core:6.4.0'

    // Needed for compatibility with JDK > 8
    implementation 'javax.activation:activation:1.1.1'
}


configurations {
    implementation.exclude group: "org.slf4j", module: "slf4j-jdk14"
    implementation.exclude group: 'org.slf4j', module: "slf4j-log4j12"
    implementation.exclude group: 'log4j', module: "log4j"
}

configurations.configureEach {
    resolutionStrategy {
        force 'xml-apis:xml-apis:1.4.01'
    }
}

/** Other configuration **/

sourceSets.main.resources {
    srcDirs = ["src/main/resources", "resources/config"]
}
sourceSets.test.resources {
    srcDirs = ["src/test/resources"]
}

test {
    exclude '**/**IntegrationTest**'
    useJUnitPlatform()
    maxHeapSize = "1024m"

    def libraries = ""
    if (Os.isFamily(Os.FAMILY_MAC)) {
        if (Os.isArch("aarch64")) {
            libraries = "${file("./grobid-home/lib/mac_arm-64").absolutePath}"
        } else {
            libraries = "${file("./grobid-home/lib/mac-64").absolutePath}"
        }
    } else if (Os.isFamily(Os.FAMILY_UNIX)) {
        def jepDir = rootProject.rootDir.getAbsolutePath() + "/grobid-home/lib/lin-64/jep"
        libraries = jepDir
        jepDir = rootProject.rootDir.getAbsolutePath() + "/grobid-home/lib/lin-64"
        libraries += ":" + jepDir
    } else {
        throw new RuntimeException("Unsupported platform!")
    }

    if (JavaVersion.current().compareTo(JavaVersion.VERSION_1_8) > 0) {
        // Reflective access needed by the test mocking frameworks (EasyMock/Mockito) on Java 21
        jvmArgs "--add-opens", "java.base/java.util.stream=ALL-UNNAMED",
            "--add-opens", "java.base/java.io=ALL-UNNAMED",
            "--add-opens", "java.base/java.lang=ALL-UNNAMED",
            "--add-opens", "java.base/java.util=ALL-UNNAMED",
            "--add-opens", "java.base/java.util.regex=ALL-UNNAMED",
            "--add-opens", "java.base/java.math=ALL-UNNAMED",
            "--add-opens", "java.base/java.text=ALL-UNNAMED",
            "--add-opens", "java.xml/jdk.xml.internal=ALL-UNNAMED"
    }
    systemProperty "java.library.path", "${System.getProperty('java.library.path')}:" + libraries
}

tasks.run.workingDir = rootProject.rootDir

application {
    mainClass = 'org.grobid.service.main.GrobidQuantitiesApplication'
}

run {
    args = ['server', 'resources/config/config.yml']

    def libraries = ""
    if (Os.isFamily(Os.FAMILY_MAC)) {
        if (Os.isArch("aarch64")) {
            libraries = "${file("../grobid-home/lib/mac_arm-64").absolutePath}"
        } else {
            libraries = "${file("../grobid-home/lib/mac-64").absolutePath}"
        }
    } else if (Os.isFamily(Os.FAMILY_UNIX)) {
        libraries = "${file("../grobid-home/lib/lin-64/jep").absolutePath}:" +
            "${file("../grobid-home/lib/lin-64").absolutePath}:"
    } else {
        throw new RuntimeException("Unsupported platform!")
    }

    if (JavaVersion.current().compareTo(JavaVersion.VERSION_1_8) > 0) {
        jvmArgs "--add-opens", "java.base/java.lang=ALL-UNNAMED"
    }
    workingDir = rootProject.rootDir
    def javaLibraryPath = "${System.getProperty('java.library.path')}:" + libraries

    systemProperty "java.library.path", javaLibraryPath
}

task integration(type: Test) {
    include '**'

    maxHeapSize = "1024m"

    def libraries = ""
    if (Os.isFamily(Os.FAMILY_MAC)) {
        if (Os.isArch("aarch64")) {
            libraries = "${file("./grobid-home/lib/mac_arm-64").absolutePath}"
        } else {
            libraries = "${file("./grobid-home/lib/mac-64").absolutePath}"
        }
    } else if (Os.isFamily(Os.FAMILY_UNIX)) {
        def jepDir = rootProject.rootDir.getAbsolutePath() + "/grobid-home/lib/lin-64/jep"
        libraries = jepDir
        jepDir = rootProject.rootDir.getAbsolutePath() + "/grobid-home/lib/lin-64"
        libraries += ":" + jepDir
    } else {
        throw new RuntimeException("Unsupported platform!")
    }

    if (JavaVersion.current().compareTo(JavaVersion.VERSION_1_8) > 0) {
        // Reflective access needed by the test mocking frameworks (EasyMock/Mockito) on Java 21
        jvmArgs "--add-opens", "java.base/java.util.stream=ALL-UNNAMED",
            "--add-opens", "java.base/java.io=ALL-UNNAMED",
            "--add-opens", "java.base/java.lang=ALL-UNNAMED",
            "--add-opens", "java.base/java.util=ALL-UNNAMED",
            "--add-opens", "java.base/java.util.regex=ALL-UNNAMED",
            "--add-opens", "java.base/java.math=ALL-UNNAMED",
            "--add-opens", "java.base/java.text=ALL-UNNAMED",
            "--add-opens", "java.xml/jdk.xml.internal=ALL-UNNAMED"
    }
    systemProperty "java.library.path", "${System.getProperty('java.library.path')}:" + libraries
}

// Training configuration

def trainerTasks = [
    //Training models
    "train_units"            : "org.grobid.trainer.UnitTrainer",
    "train_quantities"       : "org.grobid.trainer.QuantitiesTrainer",
    "train_values"           : "org.grobid.trainer.ValueTrainer",
    "train_quantifiedObjects": "org.grobid.trainer.QuantifiedObjectTrainer"
]


trainerTasks.each { taskName, trainerMainClass ->
    tasks.create(name: taskName, type: JavaExec, group: 'training') {
        mainClass = trainerMainClass
        classpath = sourceSets.main.runtimeClasspath
    }
}

// return the default value if the property has not been specified in command line
ext.getArg = { propName, defaultVal ->
    return project.hasProperty(propName) ? project.getProperty(propName) : defaultVal;
}

processResources {
    // The version and the revision are injected here rather than read from the source files, so
    // Gradle cannot see them change and keeps the task UP-TO-DATE: revision.txt would otherwise
    // hold whatever value it was first generated with. Building the release tag in a tree that
    // had already been built on master would then bake the master revision into the artifact.
    // Declaring them as inputs makes the task re-run whenever either changes.
    inputs.property('projectVersion', project.property('version').toString())
    inputs.property('gitRevision', rootProject.ext.gitRevision)

    filesMatching(["version.txt", "revision.txt"]) {
        expand(
            project_version: project.property('version') ?: "unknown",
            project_revision: rootProject.ext.gitRevision
        )
    }
}

/*** Packaging and distribution ***/

shadowJar {
    archiveClassifier = 'onejar'
    mergeServiceFiles()
    zip64 true
    manifest {
        attributes 'Main-Class': 'org.grobid.service.main.GrobidQuantitiesApplication'
    }
    from sourceSets.main.output
    configurations = [project.configurations.runtimeClasspath]
}

jar {
    enabled true
}

artifacts {
    archives shadowJar
}

distZip.enabled = true
distTar.enabled = false
shadowDistZip.enabled = false
shadowDistTar.enabled = false

publishing {
    publications {
        mavenJava(MavenPublication) {
            from components.java
        }
    }
    repositories {
        mavenLocal()
    }
}

def conf = new org.yaml.snakeyaml.Yaml().load(new File("resources/config/config.yml").newInputStream())
def grobidHome = conf.grobidHome.replace("\$", "").replace('{', "").replace("GROBID_HOME:- ", "").replace("}", "")
if (grobidHome.startsWith("../")) {
    grobidHome = "${rootProject.rootDir}/${grobidHome}"
}

/** Model management **/

task copyModels(type: Copy) {
    from "${rootDir}/resources/models"
    include "**/*.wapiti"
    include "**/config.json"
    include "**/transformer-config.json"
    include "**/transformer-tokenizer/**"
    include "**/model_weights.hdf5"
    include "**/preprocessor.json"
    exclude "**/features-engineering/**"
    exclude "**/result-logs/**"
    into "${grobidHome}/models/"

    doLast {
        print "Copy models under grobid-home: ${grobidHome}"
    }
}

// Gradle 9 removed the dynamic `exec {}` shortcut inside task actions, so we
// inject ExecOperations via a helper interface and call it from doLast.
interface InjectedExecOps {
    @Inject ExecOperations getExecOps()
}

task downloadModelsGit(dependsOn: copyModels) {
    def targetDir = file("${grobidHome}/models/quantities_models")
    def injected = objects.newInstance(InjectedExecOps)
    doLast {
        targetDir.parentFile.mkdirs()
        // Requires git-xet to be installed beforehand (https://hf.co/docs/hub/git-xet),
        // since the HuggingFace repository stores large model files via Xet storage.
        if (targetDir.exists()) {
            injected.execOps.exec {
                workingDir = targetDir
                commandLine 'git', 'pull'
            }
        } else {
            injected.execOps.exec {
                commandLine 'git', 'clone',
                    'https://huggingface.co/sciencialab/grobid-quantities-models',
                    targetDir.absolutePath
            }
        }
    }
}

task installModels(dependsOn: [copyModels, downloadModelsGit]) {
    doLast {
        def sourceDir = file("${grobidHome}/models/quantities_models")
        def destDir = file("${grobidHome}/models/")

        Arrays.asList(
            "quantities",
            "quantities-BERT_CRF",
            "quantities-BidLSTM_CRF",
            "quantities-BidLSTM_CRF_FEATURES",
            "units-BERT_CRF",
            "units-BidLSTM_CRF",
            "units-BidLSTM_CRF_FEATURES",
            "units",
            "values-BERT_CRF",
            "values-BidLSTM_CRF",
            "values-BidLSTM_CRF_FEATURES",
            "values",
        ).each { String subDirName ->
            var subDir = new File(sourceDir, subDirName)
            if (subDir.isDirectory()) {
                copy {
                    from subDir
                    into new File(destDir, subDirName)
                }
            }
        }

        delete files("${sourceDir}")
    }
}

wrapper {
    gradleVersion = "9.0.0"
}

test {
    useJUnitPlatform()
}

jacocoTestReport {
    reports {
        xml.required = true // consumed by coverallsapp/github-action in CI
        html.required = true
    }
    dependsOn test // tests are required to run before generating the report
}

release {
    failOnUnversionedFiles = false
    failOnCommitNeeded = false
    tagTemplate = 'v${version}'
    git {
        // Releases must be cut from a branch whose name contains "release"
        // (e.g. `release/0.9.0`, `feature/release-0.9.0`, `prepare-release-0.9.0`).
        // Direct release from `master` is intentionally blocked: master is
        // protected and the gradle-release plugin cannot push there. Cut the
        // release on a release-named branch, then bring it back to master via
        // a PR merged with a merge commit (NOT squash) so the tagged commit
        // remains reachable from master history - which is also what lets
        // getGitRevision() above report the new release rather than the previous one.
        // The string is treated as a regex by net.researchgate.release.
        requireBranch.set('.*release.*')
    }
}