pytest basics

Tests in their own file

pytestpytestA tool that finds test functions in files named test_*.py, runs each one, and reports what passed and failed. is that tool. It isn't part of Python: on a computer you install it, like the packages in lesson 8.8. Here it loads the first time you press Run tests.

grades.py is lesson 5.8's grade_for. Next to it, test_grades.py holds three test functionstest functionsA function whose name starts with test, like test_top, holding the assertions for one case. pytest finds it and runs it.. Each imports what it tests (lesson 8.5) and makes one assertion. Press Run tests:

def grade_for(score):
    if score >= 85:
        return "A"
    elif score >= 70:
        return "B"
    elif score >= 50:
        return "C"
    else:
        return "F"

from grades import grade_for


def test_top():
    assert grade_for(90) == "A"


def test_boundary():
    assert grade_for(85) == "A"


def test_fail():
    assert grade_for(49) == "F"

The command line shows what ran: pytest -q, "quietly". pytest looked in the folder for files named test_*.py or *_test.py, and in them for functions whose names start with test. Name yours test_ and the case, like test_boundary. It ran each one and printed a dot for each passingpassingA test whose assertions were all true when it ran. A failing test is one where an assertion was false, or an error happened. test, then 3 passed and how long it took (that time changes every run). All your test functions together are your test suitetest suiteAll the tests for a program, run together..

pytest only runs what fits its naming rules. The file below has one function with a true assertion, named check_top. Answer, then run it:

from grades import grade_for


def check_top():
    assert grade_for(90) == "A"

def grade_for(score):
    if score >= 85:
        return "A"
    elif score >= 70:
        return "B"
    elif score >= 50:
        return "C"
    else:
        return "F"

Reading a failure

Here's a bigger suite of six tests, and lesson 5.8's bug: > 85 instead of >= 85. Plain assert stopped at the first false assertion. Answer, then run it:

def grade_for(score):
    if score > 85:
        return "A"
    elif score >= 70:
        return "B"
    elif score >= 50:
        return "C"
    else:
        return "F"

from grades import grade_for


def test_top():
    assert grade_for(90) == "A"


def test_boundary():
    assert grade_for(85) == "A"


def test_middle():
    assert grade_for(75) == "B"


def test_pass():
    assert grade_for(50) == "C"


def test_fail():
    assert grade_for(49) == "F"


def test_zero():
    assert grade_for(0) == "F"

Read the report from the top. .F.... is one character per test, in order: the second one, test_boundary, is the failingfailingA test where an assertion was false, or an error happened, when it ran. one. Under FAILURES, pytest shows that test's code, with > at the line that failed. Then come the E lines, the explanation:

E       AssertionError: assert 'B' == 'A'
E         
E         - A
E         + B

This is the same assertion that gave a bare AssertionError in lesson 9.1, now with both values: grade_for(85) returned 'B', and the test said 'A'. The - line is the right side of == and the + line is the left side. Write the call on the left and the answer you expect on the right, as here, and - is always your answer, + the code's. Below them, test_grades.py:9 is the file and line. The last two lines sum it up: FAILED test_grades.py::test_boundary, then the count.

Each test starts fresh

A test file isn't a script that runs top to bottom. pytest calls each test function on its own, as the last two lines here do by hand. Predict:

def test_top():
    grade = "A"
    assert grade == "A"


def test_same_grade():
    assert grade == "A"


test_top()
test_same_grade()

Traceback (most recent call last):
  File "<your code>", line 11, in <module>
    test_same_grade()
    ~~~~~~~~~~~~~~~^^
  File "<your code>", line 7, in test_same_grade
    assert grade == "A"
           ^^^^^
NameError: name 'grade' is not defined

pytest does the same, and reports it as a failing test:

from grades import grade_for


def test_top():
    grade = grade_for(90)
    assert grade == "A"


def test_same_grade():
    assert grade_for(95) == grade

def grade_for(score):
    if score >= 85:
        return "A"
    elif score >= 70:
        return "B"
    elif score >= 50:
        return "C"
    else:
        return "F"

The report says E NameError: name 'grade' is not defined. Give each test everything it needs: here, assert grade_for(95) == "A".

Comparing floats

Lesson 2.1 showed that 0.1 + 0.2 == 0.3 is False: most floats are a tiny bit off from the decimal you wrote. pytest.approx compares with a small tolerance instead. It needs import pytest:

import pytest


def test_plain():
    assert 0.1 + 0.2 == 0.3


def test_approx():
    assert 0.1 + 0.2 == pytest.approx(0.3)

Use approx whenever the answer is a float that comes from a calculation.