ProvBook: Provenance of the Notebook.
This extension displays the provenance of each notebook cell. Every cell is extended with a provenance area with a slider. The provenance area shows the history of the execution of each code cells. As you move the slider, you could see the start and end time of each execution, how long it took to execute, the number of runs, the source and output of each execution. It also shows the provenance of text cells where it shows the the modified time and the source. ProvBook provides two extensions:
- provbook : Displays the provenance of a Jupyter Notebook.
- notebook_rdf Convert Jupyter Notebooks to RDF and the converted RDF back to Jupyter Notebooks.
A video showing the installation and use of ProvBook with an example is available here
ProvBook: Provenance-based Semantic Enrichment of Interactive Notebooks for Reproducibility, Sheeba Samuel and Birgitta König-Ries, The 17th International Semantic Web Conference (ISWC) 2018 Demo Track
Prerequisite Python3 Jupyter Notebook
Installation steps of Jupyter Notebook can be found here
Installation of ProvBook
git clone https://github.com/Sheeba-Samuel/ProvBook
The provbook extension displays the extension of the Jupyter Notebook.
cd ProvBook sudo jupyter nbextension install provbook jupyter nbextension enable provbook/main
The notebook_rdf : Convert Jupyter Notebooks to RDF and the converted RDF back to Jupyter Notebooks
This utility converts Jupyter Notebooks to RDF. This is a command line utility which takes a notebook as input and generates RDF Turtle file. The RDF is generated using the REPRODUCE-ME ontology extended from W3C standard PROV-O and the P-Plan ontology. The RDF generated from the notebook can be converted back to a Jupyter Notebook. To download provenance information as RDF, Install notebook_rdf Extension. Change the directory to notebook_rdf. Perform the following commands:
cd notebook_rdf sudo python setup.py install
After installing the extension run the notebook
Example usage of notebook_rdf
Convert your notebook to RDF
python notebook_rdf your_notebook.ipynb
python notebook_rdf --from notebook your_notebook.ipynb --to RDF
Convert your RDF to notebook
python notebook_rdf notebook_rdf.ttl
python notebook_rdf --from RDF notebook_rdf.ttl --to notebook
The provenance of a code cell. The screenshot shows the toolbar button for the displaying the provenance of selected or all cells.
The Provenance Menu. A user can toggle the provenance display for a selected cell from Cell -> Provenance -> Toggle visibility (selected). A user can clear the provenance data from the metadata of the notebook from Cell -> Provenance -> Clear (all). The notebook can be downloaded as RDF from the File Menu -> Download as -> RDF (.ttl).
Jupyter version support
The extensions work with Jupyter 5.x.
The provenance is stored in the metadata of the notebook. Every time a code cell is executed, a new entry 'provenance' is added to the metadata of the code cell. The start and end time of the execution is added with the time it took to execute. The source and the output obtained from executing the cell is added to the metadata so that it can be shared with other collaborators to verify the output.