evaluate with jenkins

.gitignore

@@ -1,6 +1,5 @@
-df_atp.csv
-df_wta.csv
-atp-and-wta-tennis-data.zip
-data
-model.zip
-secret.txt
+*.csv
+*.zip
+*.png
+*.txt
+__pycache__

Dockerfile

@@ -9,9 +9,10 @@ RUN apt install -y figlet
 RUN export PATH=”$PATH:/usr/local/bin/python”
 RUN apt install python3-pip -y
 RUN apt install unzip -y
+RUN pip3 install numpy
 RUN pip3 install kaggle
 RUN pip3 install pandas
-RUN pip3 install pillow --global-option="build_ext" --global-option="--disable-zlib" --global-option="--disable-jpeg"
+RUN pip3 install pillow
 RUN pip3 install scikit-learn
 RUN pip3 install matplotlib
 RUN pip3 install torchvision

Jenkinsfile

@@ -1,7 +1,7 @@
 pipeline {   
     agent {
 		dockerfile {
-			additionalBuildArgs '-t ium'
+			additionalBuildArgs '-t s444498-init-datasets'
 			args '-e KAGGLE_USERNAME=${params.KAGGLE_USERNAME} -e KAGGLE_KEY=${params.KAGGLE_KEY}'
 		}
 	} 
@@ -29,7 +29,7 @@ pipeline {
         }
         stage('Archive datasets') {
             steps {
-                archiveArtifacts artifacts: 'atp_test.csv, atp_train.csv', onlyIfSuccessful: true
+                archiveArtifacts artifacts: 'atp_test.csv, atp_dev.csv, atp_train.csv', onlyIfSuccessful: true
             }
         }
         stage('Run training job') {

Jenkinsfile-evaluation

@@ -0,0 +1,49 @@
+pipeline {
+    agent {
+        dockerfile true
+    }
+
+    parameters {
+        gitParameter branchFilter: 'origin/(.*)', defaultValue: 'master', name: 'BRANCH', type: 'PT_BRANCH'
+        buildSelector(
+			defaultSelector: lastSuccessful(),
+			description: 'Which build to use for copying artifacts',
+			name: 'BUILD_SELECTOR'
+		)
+    }
+
+    stages {
+        stage('Copy artifacts') {
+            steps {
+                copyArtifacts fingerprintArtifacts: true, projectName: 's444498-create-dataset', selector: buildParameter('BUILD_SELECTOR')
+                copyArtifacts fingerprintArtifacts: true, projectName: 's444498-training/${BRANCH}', selector: buildParameter('BUILD_SELECTOR')
+                copyArtifacts filter: "eval_result.txt", projectName: 's444498-evaluation/${BRANCH}', optional: true
+            }
+        }
+        stage("Evaluation") {
+            steps {
+                sh "chmod u+x ./evaluation.py"
+                sh "python3 ./evaluation.py"
+                archiveArtifacts artifacts: "eval_result.txt", onlyIfSuccessful: true
+            }
+        }
+    }
+
+    post {
+        success {
+            emailext body: "SUCCESS", subject: "s444498-evaluation", to: "e19191c5.uam.onmicrosoft.com@emea.teams.ms"
+        }
+
+        failure {
+            emailext body: "FAILURE", subject: "s444498-evaluation", to: "e19191c5.uam.onmicrosoft.com@emea.teams.ms"
+        }
+
+        unstable {
+            emailext body: 'UNSTABLE', subject: "s444498-evaluation", to: "e19191c5.uam.onmicrosoft.com@emea.teams.ms"
+        }
+
+        changed {
+            emailext body: 'CHANGED', subject: "s444498-evaluation", to: "e19191c5.uam.onmicrosoft.com@emea.teams.ms"
+        }
+    }
+}

Jenkinsfile-training

@@ -45,6 +45,7 @@ pipeline {
     post {
         success {
             emailext body: "SUCCESS", subject: "s444498-training", to: "e19191c5.uam.onmicrosoft.com@emea.teams.ms"
+            build job: "s444498-evaluation/master"
         }
 
         failure {

evaluation.py

@@ -0,0 +1,36 @@
+import matplotlib.pyplot as plt
+import torch
+from torch.utils.data import DataLoader
+from neutral_network import MLP, AtpDataset, test
+
+def load_model():
+    model = MLP()
+    model.load_state_dict(torch.load('./model.zip'))
+    return model
+
+def load_dev_dataset(batch_size=64):
+    atp_dev = AtpDataset('atp_dev.csv')
+    return DataLoader(atp_dev, batch_size=batch_size)
+
+def make_plot(values):
+    build_nums = list(range(1, len(values) + 1))
+    plt.xlabel('Build number')
+    plt.ylabel('MSE Loss')
+    plt.plot(build_nums, values, label='Model MSE Loss over builds')
+    plt.legend()
+    plt.savefig('plot.png')
+
+model = load_model()
+dataloader = load_dev_dataset()
+
+loss_fn = torch.nn.MSELoss()
+
+loss = test(dataloader, model, loss_fn)
+with open('eval_result.txt', 'a+') as f:
+    f.write(f'{str(loss)}\n')
+with open('eval_result.txt', 'r') as f:
+    values = [float(line) for line in f.readlines() if line]
+    make_plot(values)
+
+
+

init.py

@@ -62,5 +62,6 @@ print("\nElements of dev set: " + str(len(atp_dev)))
 print("\nElements of train set: " + str(len(atp_train)))
 
 # Stworzenie plików z danymi trenującymi i testowymi
-atp_test.to_csv('atp_test.csv', encoding="utf-8", index=False)  
+atp_test.to_csv('atp_test.csv', encoding="utf-8", index=False)
+atp_dev.to_csv('atp_dev.csv', encoding="utf-8", index=False)  
 atp_train.to_csv('atp_train.csv', encoding="utf-8", index=False)