Skip to main content

Jupyter notebook assignment formatting and distribution for Otter-Grader

Project description

otter-assign: Jupyter Notebook Assignments with Otter-Grader

Format and tools for authoring and distributing Jupyter notebook assignments with autograding using Otter-Grader.

Requires: Python 3 (even if it's installed, check that it's your working version python --version)

Getting started

Course instructors author assignments as Jupyter notebooks by creating a notebook that contains setup code, questions, solutions, and tests to validate those solutions. This project prepares an assignment to be distributed to students and later scored automatically.

The notebook format is not specific to a programming language or autograding framework, but was designed to be used with Otter, which is Python based. Contributions to support other testing frameworks, such as nbgrader and other programming languages, are welcome.

An example notebook appears in test/example.ipynb, which uses the notebook format. To convert it, run:

oassign test/example.ipynb test/output
  • test/example.ipynb: an example notebook path that you'll need to replace with the path to the master solution notebook, which was augmented with the metadata and commands from the notebook format.
  • test/output: the path to where the output will be stored
    • the output contains two directories autograder and student
    • the autograder directory contains the full set of tests and a solution notebook (a solution notebook is different from the master notebook, because it is not formatted accordidng to the notebook format but instead looks like the student notebook with solutions)
    • the student directory contains an automatically created redacted version.
  • at the end of the command, specify paths to any support files (e.g. data files) needed by the notebooks to be copied into the autograder and student directories

Before you run the oassign command, make sure that you run the entire notebook top to bottom (Cell -> Run All) to make sure that every cell has the correct output -- the output of the cells will be turned into the appropriate tests stored in the provided output directory (second argument of the oassign command). If you change the tests, you need to re-generate the files by re-running the notebook and the oassign command.


Test outside of a question

File "/opt/conda/lib/python3.6/site-packages/oassign/", line 141, in gen_ok_cells
    assert not is_test_cell(cell), 'Test outside of a question: ' + str(cell)
AssertionError: Test outside of a question:

If you get this error, this means that you have more than one cell between the markdown cell that declared the question (i.e., the one that contains #BEGIN QUESTION) and the cell that contains the # TEST.

SOLUTION: remove the extra code/markdown cell(s) either between the solution cell and the markdown cell with the #BEGIN QUESTION or between the solution cell and the # TEST cell.

Test cell with a blank on the last line

If your test contains a blank/newline after the test, otter-assign seems to automatically add a semicolon at the end of the test, thus, supressing the output of the command.


movies.head(1)['plots'][0]=='Two imprisoned men bond over a number of years, finding solace and eventual redemption through acts of common decency.'

Turns into the following failed test in the students' notebook:

>>> movies.head(1)['plots'][0]=='Two imprisoned men bond over a number of years, finding solace and eventual redemption through acts of common decency.';

# Error: expected
#     True
# but got

SOLUTION: remove the blank line at the end of the # TEST cell.



  • Added custom prompts for manually graded questions

v0.0.1: Initial release

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

otter-assign-0.0.2.tar.gz (9.3 kB view hashes)

Uploaded source

Built Distribution

otter_assign-0.0.2-py3-none-any.whl (11.0 kB view hashes)

Uploaded py3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page