Restoring Execution Environments of Jupyter Notebooks

Wang, Jiawei and Li, Li and Zeller, Andreas
(2021) Restoring Execution Environments of Jupyter Notebooks.
In: ICSE 2021, May 2021, Virtual.
Conference: ICSE International Conference on Software Engineering


Download (3MB) | Preview


More than ninety percent of published Jupyter notebooks do not state dependencies on external packages. This makes them non-executable and thus hinders reproducibility of scientific results. We present SnifferDog, an approach that 1) collects the APIs of Python packages and versions, creat- ing a database of APIs; 2) analyzes notebooks to determine candidates for required packages and versions; and 3) checks which packages are required to make the notebook executable (and ideally, reproduce its stored results). In its evaluation, we show that SnifferDog precisely restores execution environments for the largest majority of notebooks, making them immediately executable for end users.


Actions (login required)

View Item View Item