Test to Check

If you are looking at the SCTs of old DataCamp courses, you’ll notice they use test_x() functions instead of check_x() functions, and there is no usage of Ex(). The test_x() way of doing things has now been phased out in favor of the more transparent and composable check_x() functions that start with Ex() and are chained together with the . operator.

Common cases

Whenever you come across an SCT that uses test_x() functions, you’ll make everybody’s life easier by converting it to a check_x()-based SCT. Below are the most common cases you will encounter, together with instructions on how to translate from one to the other.

Something you came across that you didn’t find in this list? Just create an issue on GitHub. Content Engineering will explain how to translate the SCT and update this article.

test_student_typed

# Solution
y = 1 + 2 + 3

# old SCT
test_student_typed(r'1\s*\+2\s*\+3')

# new SCT
Ex().has_code(r'1\s*\+2\s*\+3')

test_object

# Solution
x = 4

# old SCT (checks equality by default)
test_object('x')

# new SCT
Ex().check_object('x').has_equal_value()

# Solution
x = 4

# old SCT
test_object('x', do_eval=False)

# new SCT
Ex().check_object('x')

test_function

# Solution
import numpy as np
arr = np.array([1, 2, 3, 4, 5])
np.mean(arr)

# old SCT (checks all args specified in solution)
test_function('numpy.array')

# new SCT
Ex().check_function('numpy.array').check_args('a').has_equal_value()

# Solution
import numpy as np
arr = np.array([1, 2, 3, 4, 5])
np.mean(arr)
np.mean(arr + arr)

# old SCT (1-based indexed)
test_function('numpy.array', index=1)
test_function('numpy.array', index=2)

# new SCT (0-based indexed)
Ex().check_function('numpy.array', index=0).check_args('a').has_equal_value()
Ex().check_function('numpy.array', index=1).check_args('a').has_equal_value()

test_function_v2

# Solution
import numpy as np
arr = np.array([1, 2, 3, 4, 5])
np.mean(arr)

# old SCT (explicitly specify args)
test_function_v2('numpy.array', params=['a'], index=1)

# new SCT
Ex().check_function('numpy.array', index=0).check_args('a').has_equal_value()

test_correct

# Solution
import numpy as np
arr = np.array([1, 2, 3, 4, 5])

# old SCT (use lambdas to defer execution)
test_correct(lambda: test_object('arr'),
             labmda: test_function('numpy.array'))

# new SCT (no need for lambdas)
Ex().check_correct(check_object('arr').has_equal_value(),
                   check_function('numpy.array').check_args('a').has_equal_value())

Enforcing check functions

For newer courses an updated version of the base Docker image is used that sets the PYTHONWHAT_V2_ONLY environment variable. When this variable is set, pythonwhat will no longer allow sct authors to use the old test_x() functions. If you are updating the Docker image for courses that have old skool SCTs, this would mean you have to rewrite all the SCTs to their check equivalents to make the build pass. If you want to work around this, thus being able to use test functions even with the latest base image, you can include the following line of code in your requirements.sh file:

echo "import os; os.environ['PYTHONWHAT_V2_ONLY'] = '0'" > /home/repl/.startup.py