diff --git a/Jenkinsfile b/Jenkinsfile index 2b27ef9..a8825ec 100644 --- a/Jenkinsfile +++ b/Jenkinsfile @@ -27,7 +27,7 @@ node { "KAGGLE_KEY=${params.KAGGLE_KEY}" ]) { sh 'echo KAGGLE_USERNAME: $KAGGLE_USERNAME' sh 'chmod u+x ./download_dataset.sh' - archiveArtifacts artifacts: '*.csv', followSymlinks: false + archiveArtifacts artifacts: 'baltimore_train.csv, baltimore_test.csv, baltimore_dev.csv' } } } \ No newline at end of file diff --git a/download_dataset.sh b/download_dataset.sh index 105c18d..52ffd72 100644 --- a/download_dataset.sh +++ b/download_dataset.sh @@ -2,7 +2,10 @@ kaggle datasets download -d sohier/crime-in-baltimore unzip crime-in-baltimore.zip +wc -l BPD_Part_1_Victim_Based_Crime_Data.csv head -n 8000 BPD_Part_1_Victim_Based_Crime_Data.csv > baltimore_train.csv tail -n 2000 BPD_Part_1_Victim_Based_Crime_Data.csv > baltimore_test.csv -head -n 2000 baltimore_train.csv > baltimore_dev.csv \ No newline at end of file +head -n 2000 baltimore_train.csv > baltimore_dev.csv + +wc -l baltimore_train.csv \ No newline at end of file