This commit is contained in:
PawelDopierala 2024-04-02 22:58:38 +02:00
parent 2044c0e46b
commit c62d532032
2 changed files with 1 additions and 2 deletions

1
Jenkinsfile vendored
View File

@ -35,6 +35,7 @@ pipeline {
steps { steps {
withEnv(["KAGGLE_USERNAME=${params.KAGGLE_USERNAME}", withEnv(["KAGGLE_USERNAME=${params.KAGGLE_USERNAME}",
"KAGGLE_KEY=${params.KAGGLE_KEY}" ]) { "KAGGLE_KEY=${params.KAGGLE_KEY}" ]) {
sh 'kaggle datasets download muhammadbinimran/housing-price-prediction-data --unzip'
sh 'chmod 777 ./data_processing.py' sh 'chmod 777 ./data_processing.py'
sh "python3 ./data_processing.py ${params.CUTOFF}" sh "python3 ./data_processing.py ${params.CUTOFF}"
} }

View File

@ -2,9 +2,7 @@ import sklearn
from sklearn.preprocessing import OneHotEncoder from sklearn.preprocessing import OneHotEncoder
from sklearn.model_selection import train_test_split from sklearn.model_selection import train_test_split
import pandas as pd import pandas as pd
import subprocess
subprocess.run(["kaggle", "datasets", "download", "muhammadbinimran/housing-price-prediction-data", "--unzip"])
housing_price_dataset = pd.read_csv('housing_price_dataset.csv') housing_price_dataset = pd.read_csv('housing_price_dataset.csv')
hp_train_test, hp_dev = sklearn.model_selection.train_test_split(housing_price_dataset, test_size=0.1) hp_train_test, hp_dev = sklearn.model_selection.train_test_split(housing_price_dataset, test_size=0.1)