in src/classes/qadataset.py [0:0]
def read_original_dataset(self, file_path: str):
"""Reads the original/raw dataset into a List of QAExamples.
Args:
file_path: Local path to the dataset.
Returns:
List[QAExample]
"""
examples = []
with gzip.open(file_path, "rb") as file_handle:
header = json.loads(file_handle.readline())["header"]
for entry in file_handle:
entry = json.loads(entry)
for qa in entry["qas"]:
examples.append(
QAExample.new(
uid=qa["qid"],
query=qa["question"],
context=entry["context"],
answers=qa["answers"],
metadata={}, # NB: Put any metadata you wish saved here.
)
)
return examples