data preparation

get-data.sh

@@ -0,0 +1,5 @@
+#!/bin/bash
+kaggle datasets download zynicide/wine-reviews
+mkdir -p data
+unzip -o wine-reviews.zip -d data
+rm wine-reviews.zip

prepare_dataset.py

@@ -0,0 +1,22 @@
+import pandas as pd
+import os
+from sklearn.model_selection import train_test_split
+
+CUTOFF = int(os.environ['CUTOFF'])
+
+wines = pd.read_csv('data/winemag-data_first150k.csv', engine='python', encoding='ISO-8859-1', sep=',')
+
+wines = wines.dropna()
+
+wines = wines.sample(100)
+
+X, Y = wines, wines
+
+# SPLIT BETWEEN DEV, TRAINS, AND TEST
+wines_train, wines_temp, wines_train, wines_temp = train_test_split(X, Y, test_size=0.2, random_state=1)
+wines_dev, wines_test, wines_dev, wines_test = train_test_split(wines_temp, wines_temp, test_size=0.2)
+
+wines_train.to_csv('wines_train.csv', index=False)
+wines_dev.to_csv('wines_dev.csv', index=False)
+wines_test.to_csv('wines_test.csv', index=False)
+