pytest-acereports

Zero-setup HTML reporting for pytest: one self-contained file per run with logs, API traffic, screenshots, retries and flaky-test detection. The same report as Ace Reports for .NET.

Website: https://harshadlambate.github.io/AceReportsWebsite/ · Live demo report: https://harshadlambate.github.io/AceReportsWebsite/demo.html

Setup

Install the package. That's all:

pip install pytest-acereports

The plugin registers itself with pytest. Run your tests as usual and the report is written to test-results/AceReport_<timestamp>.html in your project's root folder; the path is printed at the end of the run.

To turn reporting off: pytest --no-acereports, or acereports = false in pytest.ini.

Configuration

Add this hook to a conftest.py. Ace Reports calls it before the first test runs:

from acereports import ace

def pytest_acereports_configure(options):
    ace.suite_name("Checkout Regression")   # default: the name of the project folder
    ace.environment_name("UAT")             # see "Environment name" below
    ace.dark_theme()                        # or light_theme() / system_theme() (default)
    # ace.output_path("reports/report.html")
    # ace.capture_http_traffic(False)       # API capture is on by default

Or set the same things in pytest.ini (or pyproject.toml / tox.ini):

[pytest]
acereports_suite_name = Checkout Regression
acereports_environment = UAT
acereports_theme = dark
acereports_output = reports/report.html
acereports_capture_http = true
acereports_capture_logging = true
acereports_readable_names = true

pytest --acereports-output=PATH overrides the output path for one run.

Environment name

Resolved in this order, first match wins:

1. ace.environment_name("QA")
2. the ACEREPORTS_ENVIRONMENT environment variable
3. acereports_environment in pytest.ini
4. Local

Logging from a test

from acereports import ace

ace.info("Submitting payment")
ace.pass_("Payment authorized")            # 'pass' is a Python keyword, hence the underscore
ace.fail("Refund was accepted but should have been rejected")   # logs only; use assert to fail
ace.debug("Skipping optional 3-D Secure step")

Each call attaches to whichever test is running. Records from Python's logging module are added to the test's log as well. pytest's own log_level option decides what gets through, so set log_level = INFO in pytest.ini to see INFO messages.

Screenshots

A screenshot is taken automatically when a test fails, before fixtures are torn down, so the browser is still open. Take your own at any point; each becomes its own tab in the report:

ace.take_screenshot()                    # "Screenshot 1", "Screenshot 2", ...
ace.take_screenshot("After login")       # named tab
ace.take_screenshot("Cart page", image=png_bytes)   # an image you already have

Ace Reports finds the page by itself when the test uses a fixture whose name contains page, driver or browser, such as pytest-playwright's page, or has such an attribute on its test class. A browser created inside the test function has to be handed over once:

page = browser.new_page()
ace.use_screenshot_source(page)          # Playwright Page, Selenium WebDriver, ...

Works with Playwright (sync API), Selenium, and anything with a screenshot() or get_screenshot_as_png() method. With Playwright's async API, take the screenshot yourself and pass it in with image=await page.screenshot().

API traffic

Every HTTP call made with requests or httpx (sync and async) is captured: method, URL, request and response bodies, status, duration, and connection errors. It is shown per attempt in the API tab. Authorization, Proxy-Authorization, Cookie and Set-Cookie headers are redacted. Turn it off with ace.capture_http_traffic(False).

Categories

Markers are categories, from the test, its class and its module:

pytestmark = pytest.mark.checkout          # every test in the file

@pytest.mark.smoke
@pytest.mark.category("UAT, Payments")     # names that are not valid identifiers
def test_pay():
    ...

Markers that steer pytest or a plugin (parametrize, skip, xfail, flaky, usefixtures, ...) are not categories. A test with no category is filed under "Uncategorized".

Retries and flaky tests

Retries come from pytest-rerunfailures:

@pytest.mark.flaky(reruns=3)
def test_checkout():
    ...

or pytest --reruns 3 for every test. Each attempt keeps its own logs, API calls, error and screenshots, shown as separate attempt tabs. A test is retried when it took more than one attempt, and flaky when it ultimately passed after an earlier attempt failed.

What is recorded

pytest	Report
test function / method	one row; test_create_post is shown as "Create post" (set acereports_readable_names = false to keep the function name)
@pytest.mark.parametrize	one row per case; parameters listed as name=value
class, or module when there is no class	the "Test Group" column
markers	category tags (a test can have several)
setup or teardown errors	failed, with "Error in setup: ..." / "Error in teardown: ..."
skip, skipif, xfail	left out of the table; skipped tests are counted
@pytest.mark.ace_name("Valid User Login")	the name shown in the report (also works on pytest.param(..., marks=...))

pytest-bdd

Gherkin scenarios run with pytest-bdd are recorded with their scenario title as the name, their feature as the group, their tags as categories and their example row as arguments. Each step is a GIVEN / WHEN / THEN / AND / BUT line in the log; a failing step carries its error, and the steps it prevented are listed as skipped. Nothing to configure.

Parallel runs

Works with pytest-xdist (pytest -n auto): every worker's results end up in one report.

Requirements

Python 3.9+ and pytest 7.0+.

License

Ace Reports is free to use under the PolyForm Perimeter License 1.0.1.

Commercial terms (OEM, white-label, or a full buy-out) are available: see the website.

Found it useful? Buy me a coffee ☕