caplog for log messages in PyTest - Framework Patterns

Framework Mode - caplog for log messages

Folder Structure

project-root/
├── tests/
│   ├── test_example.py
│   ├── conftest.py
│   └── utils.py
├── src/
│   └── app_code.py
├── pytest.ini
└── requirements.txt
    

This is a simple Pytest project structure. Test files go under tests/. The conftest.py holds fixtures and shared setup. Application code is in src/.

Test Framework Layers

• Test Layer: Test functions in tests/ using Pytest syntax.
• Logging Capture Layer: Use Pytest's caplog fixture to capture log messages during tests.
• Utilities Layer: Helper functions in tests/utils.py for common test tasks.
• Fixtures Layer: Shared setup in conftest.py, including configuring loggers if needed.
• Application Layer: Code under test in src/.

Configuration Patterns

Configure logging and test settings in pytest.ini or conftest.py. Example:

# pytest.ini
[pytest]
log_cli = true
log_cli_level = INFO
    

This enables live logging during tests. You can also configure loggers in conftest.py to set levels or formats.

Test Reporting and CI/CD Integration

Pytest outputs test results with pass/fail status. When using caplog, you can assert on log messages to verify behavior.

Integrate with CI/CD pipelines by running pytest commands. Use plugins like pytest-html for HTML reports or pytest-junitxml for XML reports.

Example CI command:

pytest --junitxml=reports/results.xml
    

Best Practices

• Use caplog fixture to capture logs instead of manual logger patching.
• Assert on specific log messages and levels to verify expected logging behavior.
• Keep log messages clear and consistent in the application code for easier testing.
• Configure logging levels in pytest.ini to control verbosity during tests.
• Use fixtures in conftest.py to set up logger configurations shared across tests.

Self Check

Where would you add a new test that verifies a warning log message is generated when invalid input is given?

Key Result

Use Pytest's caplog fixture to capture and assert log messages during tests for clear verification of logging behavior.

Practice

1. What is the main purpose of using caplog in pytest tests?

easy

A. To capture and check log messages generated during test execution

B. To mock external API calls in tests

C. To measure test execution time

D. To generate test reports automatically

Solution

1. Step 1: Understand caplog's role in pytest

   caplog is a pytest fixture used to capture log messages during tests.

2. Step 2: Identify the main use

   It allows assertions on log output to verify correct logging behavior.

3. Final Answer:

   To capture and check log messages generated during test execution -> Option A

4. Quick Check:

   caplog = capture logs [OK]

Hint: caplog captures logs so you can check messages easily [OK]

Common Mistakes:

• Confusing caplog with mocking tools
• Thinking caplog measures time
• Assuming caplog generates reports

2. Which of the following is the correct way to use caplog to capture only WARNING level logs in a test function?

easy

A. caplog.log_level(logging.WARNING)

B. caplog.set_level(logging.WARNING)

C. caplog.capture_level = logging.WARNING

D. with caplog.at_level(logging.WARNING):

Solution

1. Step 1: Recall caplog syntax for setting log level

   caplog uses a context manager method at_level() to set the log level temporarily.

2. Step 2: Match correct syntax

   The correct syntax is with caplog.at_level(logging.WARNING): to capture warnings.

3. Final Answer:

   with caplog.at_level(logging.WARNING): -> Option D

4. Quick Check:

   Use caplog.at_level() for log level [OK]

Hint: Use with caplog.at_level(level) to set capture level [OK]

Common Mistakes:

• Trying to set level by assignment
• Using non-existent caplog methods
• Forgetting to use 'with' for context

3. Given the following test code, what will be the output of the assertion?

import logging

def test_log_message(caplog):
    with caplog.at_level(logging.INFO):
        logging.info('Test started')
    assert 'Test started' in caplog.text

medium

A. The assertion passes because the log message is captured

B. The assertion fails because caplog.text is empty

C. The test raises a NameError due to missing import

D. The assertion fails because logging level is too low

Solution

1. Step 1: Analyze log capture level

   The code sets caplog to capture INFO level logs, which includes the info message logged.

2. Step 2: Check assertion on caplog.text

   The message 'Test started' is logged and appears in caplog.text, so the assertion passes.

3. Final Answer:

   The assertion passes because the log message is captured -> Option A

4. Quick Check:

   Logged message in caplog.text = assertion pass [OK]

Hint: caplog.text holds captured logs as a string [OK]

Common Mistakes:

• Assuming caplog.text is empty without at_level
• Confusing log levels and capture
• Forgetting to import logging

4. Identify the error in the following pytest test using caplog:

import logging

def test_error_log(caplog):
    caplog.at_level(logging.ERROR)
    logging.error('An error occurred')
    assert 'An error occurred' in caplog.text

medium

A. logging.error() is called before caplog is imported

B. caplog.at_level() must be used as a context manager with 'with'

C. assertion should check caplog.records instead of caplog.text

D. logging level ERROR is too high to capture

Solution

1. Step 1: Check usage of caplog.at_level()

   caplog.at_level() is a context manager and must be used with 'with' keyword.

2. Step 2: Identify missing 'with' statement

   The code calls caplog.at_level() without 'with', so it does not set the level properly.

3. Final Answer:

   caplog.at_level() must be used as a context manager with 'with' -> Option B

4. Quick Check:

   Use 'with caplog.at_level()' not just call [OK]

Hint: Always use 'with' for caplog.at_level() context [OK]

Common Mistakes:

• Calling caplog.at_level() without 'with'
• Checking wrong caplog attribute
• Misunderstanding log level capture

5. You want to test that your function logs a WARNING message only when a certain condition is true. Which pytest test code correctly uses caplog to check this behavior?

hard

A. def test_warn_log(caplog):
with caplog.at_level(logging.WARNING):
my_function(True)
assert 'Warning triggered' in caplog.text
with caplog.at_level(logging.WARNING):
my_function(False)
assert 'Warning triggered' not in caplog.text

B. def test_warn_log(caplog):
with caplog.at_level(logging.INFO):
my_function(True)
assert 'Warning triggered' in caplog.text

C. def test_warn_log(caplog):
with caplog.at_level(logging.WARNING):
my_function(True)
my_function(False)
assert caplog.text.count('Warning triggered') == 1

D. def test_warn_log(caplog):
caplog.set_level(logging.WARNING)
my_function(True)
assert 'Warning triggered' in caplog.text
my_function(False)
assert 'Warning triggered' not in caplog.text

Solution

1. Step 1: Understand the requirement

   We want to check that the warning message logs only once when condition is true, not when false.

2. Step 2: Analyze options for correct caplog usage

   def test_warn_log(caplog): with caplog.at_level(logging.WARNING): my_function(True) my_function(False) assert caplog.text.count('Warning triggered') == 1 uses one context capturing both calls and asserts the warning appears exactly once, matching the requirement.

3. Step 3: Why other options fail

   def test_warn_log(caplog): with caplog.at_level(logging.WARNING): my_function(True) assert 'Warning triggered' in caplog.text with caplog.at_level(logging.WARNING): my_function(False) assert 'Warning triggered' not in caplog.text: separate contexts but caplog.text accumulates logs, so second assert 'Warning triggered' not in caplog.text fails (message from first call persists); def test_warn_log(caplog): caplog.set_level(logging.WARNING) my_function(True) assert 'Warning triggered' in caplog.text my_function(False) assert 'Warning triggered' not in caplog.text: caplog.set_level does not exist (AttributeError), and accumulation would fail second assert; def test_warn_log(caplog): with caplog.at_level(logging.INFO): my_function(True) assert 'Warning triggered' in caplog.text: tests only True case with INFO level (captures WARNING but no False verification).

4. Final Answer:

   def test_warn_log(caplog): with caplog.at_level(logging.WARNING): my_function(True) my_function(False) assert caplog.text.count('Warning triggered') == 1 -> Option C

5. Quick Check:

   Use one caplog context and count messages [OK]

Hint: Use one with caplog.at_level() block and count messages [OK]

Common Mistakes:

• Using multiple caplog contexts losing logs
• Using wrong log level in at_level
• Assuming caplog.set_level exists