Gradescope Python Autograder Example

View project source on Github - - sample solution

Project Description

In this assignment, students will build an infix calculator REPL. The goal of this project is to teach the basics of parsing and evaluating a simple language.


  • Build an infix calculator read-eval-print loop
  • The calculator should handle the 4 basic operations, +, -, *, /, with operator precedence
  • In addition, it should handle parentheses and negative numbers
  • If the user types 'quit', exit the program
  • If there are syntax errors in the user input, raise CalculatorException

Dependencies (for tests)

  • Python 2.7+/3+
  • gradescope-utils provides decorators for setting point values for tests, and running tests with a JSON output. See the Github repository for more on what you can do with it, or you can look at the example tests in this project for some usage examples.
  • subprocess32 is a convenient backport of Python 3.2's subprocess module. Used in one of the tests to communicate with an instance of the REPL to verify that it responds to input correctly.

Python 3

Make sure to use pip3 and python3 when writing your code, because our autograder base image does include Python 2 as well, which are currently the defaults for pip and python. When installing Python 3, use the apt packages python3 and python3-pip. If you need a more recent version than what is packaged by Ubuntu 18.04, you can try using a PPA or installing from source.

Example Test

class TestSimpleArithmetic(unittest.TestCase):
    def setUp(self):
        self.calc = Calculator()

    def test_eval_add(self):
        """Evaluate 1 + 1"""
        val = self.calc.eval("1 + 1")
        self.assertEqual(val, 2)

The title of the test case is taken from the first line of the docstring. This is a unittest convention. The weight for each test is given by the @weight decorator.

See all tests here

Running Tests

suite ='tests')


This script installs Python and the pip package manager. Then it uses pip to install our two external dependencies.


This script copies the student's submission to the target directory, and then executes the test runner Python script.

This python script loads and runs the tests using the JSONTestRunner class from gradescope-utils. This produces the JSON formatted output to stdout, which is then captured and uploaded by the autograder harness.

This is a blank template file for the students to fill in. Note that their solution must be called for the autograder to work correctly.