pipeline {
    agent {
        dockerfile {
            filename 'Dockerfile'
            label 'docker'
        }
    }

    triggers {
        upstream(upstreamProjects: 's123456-create-dataset', threshold: hudson.model.Result.SUCCESS)
    }

    parameters {
        string(
            defaultValue: '--epochs 100 --batch_size 32 --learning_rate 0.01',
            description: 'Parametry trenowania',
            name: 'TRAINING_PARAMS'
        )
    }

    environment {
        DATASET_PROJECT = 's123456-create-dataset'
        DATA_DIR = 'data'
    }

    stages {
        stage('Clone repository') {
            steps {
                checkout scm
            }
        }

        stage('Copy Dataset') {
            steps {
                script {
                    copyArtifacts(
                        projectName: "${env.DATASET_PROJECT}",
                        selector: lastSuccessful(),
                        target: "${env.DATA_DIR}"
                    )
                }
            }
        }

        stage('Set execute permission') {
            steps {
                script {
                    sh 'chmod +x data_processing.sh'
                }
            }
        }

        stage('Run shell script') {
            steps {
                script {
                    sh './data_processing.sh'
                }
            }
            post {
                success {
                    archiveArtifacts artifacts: 'results.txt', allowEmptyArchive: true
                }
            }
        }

        stage('Train Model') {
            steps {
                sh "python3 model.py ${params.TRAINING_PARAMS}"
            }
        }

        stage('Archive Model') {
            steps {
                archiveArtifacts artifacts: 'orange_quality_model_tf.h5', allowEmptyArchive: true
                archiveArtifacts artifacts: 'predictions_tf.json', allowEmptyArchive: true
            }
        }
    }

    post {
        always {
            cleanWs()
        }
    }
}