mlflow attempt task 2 prediction

Dockerfile

@@ -18,6 +18,7 @@ RUN pip3 install sklearn
 RUN pip3 install pymongo
 RUN pip3 install sacred
 RUN pip3 install mlflow
+RUN pip3 install tarfile
 
 CMD python3 data_expl.py
 CMD python3 nn_train.py

Jenkinsfile_pred

@@ -0,0 +1,41 @@
+pipeline {
+    agent { 
+        dockerfile  true
+    }
+    parameters {
+		buildSelector(
+            defaultSelector: lastSuccessful(),
+            description: 'Which build to use for copying artifacts',
+            name: 'BUILD_SELECTOR'
+        )
+        string(
+			defaultValue: 'input_example.json',
+			description: 'Input file name',
+			name: 'INPUT_FILE_NAME',
+			trim: false
+        )
+	}
+
+    stages {
+        stage('Stage 1') {
+            steps {
+                echo 'Hello world!' 
+            }
+        }
+
+        stage('Copy from different Pipeline') {
+			steps {
+                copyArtifacts fingerprintArtifacts: false, projectName: 's449288-training/master', selector: buildParameter('BUILD_SELECTOR')
+			}
+		}
+
+        stage('Prediction') {
+            steps {
+                    sh 'python3 predict.py $INPUT_FILE_NAME'
+                    sh 'rm -r ml'
+                
+            }
+        }
+    }
+}
+

predict.py

@@ -0,0 +1,18 @@
+import json
+import mlflow
+import numpy as np
+import sys
+import tarfile
+
+file = tarfile.open('mlruns.tar.gz')
+file.extractall('./ml')
+
+input = str((sys.argv[1:])[0])
+PATH = "ml/mlruns/1/f65f936936024133a2c03e1e486ba9cf/artifacts/model/"
+model = mlflow.pytorch.load_model(f"{PATH}/MLmodel")
+
+with open(f'[PATH]/{input}', 'r') as file:
+    json_data = json.load(file)
+    
+    print(f"Input: {json_data['inputs'][0]}")
+    print(f"Prediction: {model.predict(np.array([json_data['inputs'][0]], dtype=np.float32))}")