autograde.py 0.0.2

Run tests written with unittest against a specified module.

Installation

pip install autograde.py

Module Interface

Run tests written with unittest against a specified module. Additional options:

• location of tests (if not in the current directory),
• fallback module to run the tests against,
• where to save a CSV of the test results, and
• config file to specify how results are processed

usage: grader [-h] [--fallback FALLBACK] [--submission SUBMISSION] [--tests TESTS]
              [--test-pattern TEST_PATTERN] [--output OUTPUT] [--log LOG] [--config CONFIG]
              path

positional arguments:
  path                  Module to grade

optional arguments:
  -h, --help            show this help message and exit
  --fallback FALLBACK   Fallback module to grade
  --submission SUBMISSION
                        Submission name to grade.
  --tests TESTS         Path of tests to run. Defaults to ./
  --test-pattern TEST_PATTERN
                        Test name pattern to match. Defaults to "test*.py"
  --output OUTPUT       Output file for report. Defaults to stdout.
  --log LOG             Log file to use. Defaults to stdout.
  --config CONFIG       Config file to use.

Tests Discovery

Tests to run can be located anywhere using a combination of the --tests and --test-pattern args. By default, they are searched for under the current directory and match the test*.py pattern. Tests are discovered using the unittest module.

Running Tests

While tests are running, the script will change to the directory of path. This means that file locations in tests should be relative to the provided path. --fallback helps the module determine all tests that should be run. This is helpful when the code located at path is untrusted and may totally fail the test discovery step.

Config

The grader config can be specified with the --config arg. It is a json file that can specify what happens when a certain test runs or when a certain submission is graded. Individual test configs should be under the "tests" map and the key should be of the form: testFunction (test_filename.TestCaseName). Inividual submission configs should be under the "submissions" map and match the name of the submission. An example format is below:

{
    "tests": {
        "testSimple1 (test_simple.TestSimpleTestCase)": {
            "name": "Simple Test 1",
            "weight": 2
        }
    },
    "names": {
        "submission_name": {
            "name": "custom label here"
        }
    }
}