This commit is contained in:
Adam Wojdyla 2022-04-02 22:21:12 +02:00
parent e09e36265e
commit 8320d74da3
5 changed files with 54 additions and 4 deletions

View File

@ -20,7 +20,7 @@ COPY . .
ARG KAGGLE_USERNAME
ARG KAGGLE_KEY
RUN chmod u+x ./script-download.py
RUN chmod u+x ./stats-docker.sh
RUN chmod u+x ./script-stats.py
# RUN ./download.sh 117928

View File

@ -20,13 +20,13 @@ pipeline {
}
agent {
dockerfile{
additionalBuildArgs '--build-arg KAGGLE_USERNAME="$KAGGLE_USERNAME" --build-arg KAGGLE_KEY="$KAGGLE_KEY" --no-cache=true'
additionalBuildArgs '--build-arg KAGGLE_USERNAME="$KAGGLE_USERNAME" --build-arg KAGGLE_KEY="$KAGGLE_KEY" -t s444507_create_dataset_image'
}
}
stages {
stage('Prepare dataset') {
steps {
sh 'python3 ./script-stats.py'
sh './stats-docker.sh'
}
}
}

38
Jenkinsfile-stats Normal file
View File

@ -0,0 +1,38 @@
pipeline {
parameters {
string(
defaultValue: 'heatedboss2',
description: 'Kaggle username',
name: 'KAGGLE_USERNAME',
trim: false
)
password(
defaultValue: '',
description: 'Kaggle token',
name: 'KAGGLE_KEY'
)
string(
defaultValue: '117928',
description: 'Cutoff',
name: 'CUTOFF',
trim: false
)
}
agent {
dockerfile{
additionalBuildArgs '--build-arg KAGGLE_USERNAME="$KAGGLE_USERNAME" --build-arg KAGGLE_KEY="$KAGGLE_KEY" -t s444507_create_dataset_image'
}
}
stages {
stage('Prepare dataset') {
steps {
sh 'python3 ./script-stats.py'
}
}
}
post {
success {
archiveArtifacts artifacts: 'Car_Prices_Poland_Kaggle_*', followSymlinks: false
}
}
}

View File

@ -9,7 +9,7 @@ import numpy as np
def unzip_package():
"""Unzip dataset"""
print('Unzipping dataset...')
os.system('unzip -o car-prices-poland.zip')
os.system('unzip -o ./car-prices-poland.zip')
print('Dataset unzipped')
print('Removing .zip file...')
os.system('rm ./car-prices-poland.zip')

12
stats-docker.sh Normal file
View File

@ -0,0 +1,12 @@
echo 'Total elements in Car Prices Poland dataset:'
wc -l ./Car_Prices_Poland_Kaggle.csv
echo 'Total elements in train dataset:'
wc -l ./Car_Prices_Poland_Kaggle_train.csv
echo 'Total elements in test dataset:'
wc -l ./Car_Prices_Poland_Kaggle_test.csv
echo 'Total elements in dev dataset:'
wc -l ./Car_Prices_Poland_Kaggle_dev.csv