Initialize a coverage measurement tool for tracking code execution.

Signal the intention to start measuring code coverage for testing or debugging.

When setting up test coverage collection, before running test suites, or when profiling code execution.

Used in test scripts, continuous integration pipelines, local debugging sessions, and coverage reporting scripts.

cov = coverage.Coverage()

<variable> = coverage.Coverage()

cov: variable name (any valid Python identifier); coverage: module name (must be 'coverage' if using the standard package); Coverage: class name from the coverage module

• .start()
• .stop()
• .report()
• .html_report(directory='cov_html')
• .xml_report(outfile='coverage.xml')
• .erase()

• cov = coverage.Coverage(source=['src'])
• cov = coverage.Coverage(omit=['*/tests/*'])
• cov = coverage.Coverage(branch=True)

Failing to call .start() before running tests, resulting in no coverage data. Calling .stop() but forgetting to generate a report, leaving data inaccessible. Using coverage without installing the coverage package, causing ImportError. Omitting .erase() between test runs leading to accumulated coverage data. Using .report() without show_missing, missing insight into missed lines.

Using pytest-cov plugin: integrates coverage directly with pytest, reducing boilerplate. Using unittest.TestCase: focuses on test assertions rather than coverage measurement. Using coverage.xml for CI integration: provides machine‑readable reports for trend tracking.

Coming from Java: may expect built‑in coverage like JaCoCo; in Python you must install and import the coverage package. Coming from JavaScript: may expect an Istanbul‑like CLI; in Python you typically invoke coverage via API or command line. Coming from Go: may assume a built‑in coverage tool; coverage.py requires explicit invocation.

• coverage
• testing
• testing-tools
• coverage-reporting

Not suitable for production performance profiling; adds measurable runtime overhead. Only measures executed lines by default; branch coverage requires branch=True argument. Data files (.coverage) can accumulate; regular cleanup via .erase() or parallel mode is needed for parallel test runs.

Enables teams to quantify test coverage, identify untested code, and improve test suite effectiveness, leading to higher software reliability.

Think of coverage.Coverage() as hiring an auditor who watches every line your code runs, ready to report what was missed.

Requires the coverage.py package installed via pip; version 6+ recommended for modern features.

Using pytest‑cov for integrated reporting or integrating coverage.xml into CI pipelines for trend analysis.