Dataset from: https://conala-corpus.github.io/
Uploaded here for easy remote access.
To access in google colaboratory:
!pip install datasets
from datasets import load_dataset
import pandas as pd
train = load_dataset('json', data_files='https://huggingface.co/datasets/janrauhl/conala/resolve/main/conala-train.json')
test = load_dataset('json', data_files='https://huggingface.co/datasets/janrauhl/conala/resolve/main/conala-test.json')
train = train['train'].to_pandas()
test = test['train'].to_pandas()