split_data.py

import dataset
import generate_test_splits as g
import pandas as pd

d = dataset.DataSet() 
g.generate_hold_out_split(d)

dataset_bodies = pd.read_csv("./dataset/train_bodies.csv")
dataset_stances = pd.read_csv("./dataset/train_stances.csv")
#train_bodies = pd.read_csv("./dataset/competition_test_bodies.csv")
#data_stances = pd.read_csv("./dataset/competition_test_stances_unlabeled.csv")
merged = pd.merge(dataset_bodies, dataset_stances, on="Body ID")
training_bodies = pd.read_csv("./splits/training_ids.csv")
training_bodies_final = merged.loc[merged['Body ID'].isin(training_bodies['Body ID'])]
train_bodies_csv = training_bodies_final[['Body ID','articleBody']].copy()
train_stances_csv = training_bodies_final[['Headline','Body ID','Stance']].copy()
train_bodies_csv.to_csv('./dataset/train_bodies1.csv', index=False)
train_stances_csv.to_csv('./dataset/train_stances1.csv', index=False)

dev_bodies = pd.read_csv("./splits/hold_out_ids.csv")
dev_bodies_final = merged.loc[merged['Body ID'].isin(dev_bodies['Body ID'])]
dev_bodies_csv = dev_bodies_final[['Body ID','articleBody']].copy()
dev_stances_csv = dev_bodies_final[['Headline','Body ID','Stance']].copy()
dev_bodies_csv.to_csv('./dataset/dev_bodies1.csv', index=False)
dev_stances_csv.to_csv('./dataset/dev_stances1.csv', index=False)