Coverage in CI pipelines in PyTest

Introduction

Coverage in CI pipelines helps check how much of your code is tested automatically. It shows if important parts are missing tests.

When you want to make sure new code changes have enough tests before merging.

When you want to catch untested code early during development.

When you want to keep track of test quality over time.

When you want to prevent bugs by ensuring critical code is covered by tests.

Syntax

PyTest

pytest --cov=your_package tests/

This command runs tests and measures coverage for the specified package.

Use --cov-report=term to see coverage summary in the terminal.

Examples

Runs tests in the tests/ folder and measures coverage for the myapp package.

PyTest

pytest --cov=myapp tests/

Shows coverage and lists lines not covered by tests in the terminal.

PyTest

pytest --cov=myapp --cov-report=term-missing tests/

Generates an XML coverage report for CI tools to read.

PyTest

pytest --cov=myapp --cov-report=xml tests/

Sample Program

This simple test checks only the add function. The subtract function is not tested, so coverage will show it as missing.

PyTest

def add(a, b):
    return a + b

def subtract(a, b):
    return a - b


def test_add():
    assert add(2, 3) == 5

# Note: No test for subtract function

# Run with:
# pytest --cov=.

Important Notes

Always generate coverage reports in formats your CI tool supports, like XML or HTML.

Coverage shows what code runs during tests, but not if tests are correct.

High coverage is good, but focus also on meaningful tests.

Summary

Coverage in CI helps track how much code is tested automatically.

Use pytest --cov commands to measure and report coverage.

Integrate coverage reports in CI pipelines to catch missing tests early.

Practice

1. What is the main purpose of using coverage tools like pytest --cov in CI pipelines?

easy

A. To generate user documentation

B. To speed up the test execution time

C. To deploy the application automatically

D. To measure how much of the code is tested automatically

Solution

1. Step 1: Understand coverage tools in testing

   Coverage tools measure the percentage of code executed by tests.

2. Step 2: Role in CI pipelines

   In CI, coverage helps ensure tests cover enough code to catch bugs early.

3. Final Answer:

   To measure how much of the code is tested automatically -> Option D

4. Quick Check:

   Coverage measures tested code = D [OK]

Hint: Coverage shows tested code percentage in CI [OK]

Common Mistakes:

• Confusing coverage with test speed
• Thinking coverage deploys code
• Assuming coverage creates docs

2. Which of the following is the correct command to run pytest with coverage reporting?

easy

A. pytest --cov=.

B. pytest --cov-report=html --cov

C. pytest -coverage

D. pytest --coverage=report

Solution

1. Step 1: Identify correct pytest coverage syntax

   The correct syntax uses --cov to specify coverage on the current directory.

2. Step 2: Analyze options

   pytest --cov=. correctly runs coverage on the current directory with default terminal reporting.

3. Final Answer:

   pytest --cov=. -> Option A

4. Quick Check:

   Correct pytest coverage command = B [OK]

Hint: Use --cov and --cov-report together for coverage output [OK]

Common Mistakes:

• Using incorrect flags like --coverage=report
• Missing --cov option
• Wrong flag syntax like -coverage

3. Given this pytest command in a CI pipeline:

pytest --cov=myapp --cov-report=term

What will be the output shown in the CI logs?

medium

A. A summary of coverage percentages printed in the terminal

B. No coverage information will be shown

C. A detailed HTML coverage report saved to disk

D. An error because --cov-report=term is invalid

Solution

1. Step 1: Understand --cov-report=term option

   The term option prints coverage summary in the terminal output.

2. Step 2: Analyze expected CI log output

   The CI logs will show coverage percentages summary, not an HTML file or error.

3. Final Answer:

   A summary of coverage percentages printed in the terminal -> Option A

4. Quick Check:

   --cov-report=term shows summary in terminal = C [OK]

Hint: term report prints coverage summary in console [OK]

Common Mistakes:

• Expecting HTML report in terminal
• Assuming no coverage output
• Thinking --cov-report=term is invalid

4. You added coverage to your CI pipeline with this command:

pytest --cov=myapp --cov-report=html

But the coverage report is missing after the run. What is the most likely cause?

medium

A. The --cov option is misspelled

B. The coverage report is saved in a different directory, not checked by CI

C. pytest does not support coverage reporting

D. The tests did not run because of a syntax error

Solution

1. Step 1: Understand where coverage HTML reports are saved

   By default, HTML reports are saved in a folder named htmlcov in the current directory.

2. Step 2: Check CI pipeline file handling

   If the CI pipeline does not collect or upload this folder, the report will appear missing.

3. Final Answer:

   The coverage report is saved in a different directory, not checked by CI -> Option B

4. Quick Check:

   HTML report saved in htmlcov folder = A [OK]

Hint: Check htmlcov folder location in CI artifacts [OK]

Common Mistakes:

• Assuming --cov is misspelled without checking
• Believing pytest lacks coverage support
• Ignoring test run errors

5. In a CI pipeline, you want to fail the build if coverage falls below 80%. Which pytest coverage option helps enforce this?

hard

A. --fail-if-coverage-below=80

B. --coverage-threshold=80

C. --cov-fail-under=80

D. --cov-minimum=80

Solution

1. Step 1: Identify pytest coverage option for minimum coverage

   The correct option is --cov-fail-under which sets a minimum coverage percentage.

2. Step 2: Understand CI build failure behavior

   If coverage is below the set value, pytest returns a failure status causing CI to fail the build.

3. Final Answer:

   --cov-fail-under=80 -> Option C

4. Quick Check:

   Fail build if coverage under 80% = B [OK]

Hint: Use --cov-fail-under to enforce coverage minimum [OK]

Common Mistakes:

• Using incorrect or non-existent flags
• Confusing coverage reporting with build failure
• Assuming coverage threshold is set elsewhere