INNER CODE UNIT · Python
load_dir_reviews
NirantK/NLP_Quickbook · Part-08 Web Deployments/utils.py:38
def load_dir_reviews(reviews_path):
files_list = list(reviews_path.iterdir())
reviews = []
for filename in files_list:
f = open(filename, "r", encoding="utf-8")
reviews.append(f.read())
return pd.DataFrame({"text": reviews})
pos_path = dir_path / "pos"
neg_path = dir_path / "neg"
pos_reviews, neg_reviews = load_dir_reviews(pos_path), load_dir_reviews(neg_path)
pos_reviews["label"] = 1
neg_reviews["label"] = 0
merged = pd.concat([pos_reviews, neg_reviews])
df = merged.sample(frac=1.0) # shuffle the rows