Create custom Jupyter Notebook challenges with unit tests written in Python
You can add a custom challenge to any assessment or interview. To create a custom challenge on Coderbyte, navigate to your custom challenge library by clicking Customization and then Custom Challenges.
Create graded Jupyter Notebook challenges
Under "Unit Tests" click Create.
Then select the Jupyter option to get started.
You will then be brought to an editor page that loads a new Jupyter Notebook for you with a single file in it. This file, called coderbyte_grading.py is where you write your custom code to perform the grading. There are instructions in the file on how to compare candidate data outputs to your expected outputs.
You can then either create a new .ipynb file or upload en existing one you already have. When you are ready, provide a name for your challenge and click save.
Grading Jupyter Notebook solutions
Your expected output can either be some sort of internal JSON object or Python data structure, or it can be a larger file (e.g. a large CSV or Pickle file). Rather than adding the expected output file directly into this challenge, you upload the file to your account beforehand and then you can reference it via a URL in this grading file. This also prevents candidates from seeing the expected output data directly in their Notebook.