by Juan Altmayer Pizzorno and Emery Berger at UMass Amherst's PLASMA lab.
SlipCover is a fast code coverage tool. It tracks a Python program as it runs and reports on the parts that executed and those that didn't. That can help guide your testing (showing code that isn't being tested), debugging, fuzzing or to find "dead" code.
Past code coverage tools can make programs significantly slower; it is not uncommon for them to take twice as long to execute. SlipCover aims to provide the same information with near-zero overhead, often almost as fast as running the original Python program.
Previous coverage tools like Coverage.py rely on
Python's tracing facilities,
which add significant overhead.
Instead, SlipCover uses just-in-time instrumentation and de-instrumentation.
When SlipCover gathers coverage information, it modifies the program's Python byte codes,
inserting instructions that let it keep track the lines executed by the program.
As the program executes, SlipCover gradually removes instrumentation that
is no longer needed, allowing those parts to run at full speed.
Care is taken throughout SlipCover to keep things as efficient as possible.
On Python 3.12, rather than rewrite bytecode, SlipCover uses the new
sys.monitoring API
to collect coverage information.
The first image on the right shows SlipCover's speedup, ranging from 1.1x to 3.4x, in relation to Coverage.py, running on CPython 3.10.5.
The first two benchmarks are the test suites for scikit-learn and Flask; "sudoku" runs Peter Norvig's Sudoku solver while the others were derived from the Python Benchmark Suite.
More "Python-intensive" programs such as sudoku and those from the benchmark suite (with a larger proportion of execution time spent in Python, rather than in native code) generate more tracing events, causing more overhead in Coverage.py. While each program's structure can affect SlipCover's ability to de-instrument, its running time stays relatively close to the original.
On PyPy 3.9, the speedup ranges from 2.1x to 104.9x. Since it is so high for some of the benchmarks, we plot it on a logarithmic scale (see the second image on the right).
In a proof-of-concept integration with a property-based testing package, SlipCover sped up coverage-based testing 22x.
We verified SlipCover's accuracy against Coverage.py and against a simple script of our own that collects coverage using Python tracing. We found SlipCover's results to be accurate, in fact, in certain cases more accurate.
SlipCover is available from PyPI. You can install it like any other Python module with
pip3 install slipcover
You could then run your Python script with:
python3 -m slipcover myscript.py
SlipCover can also execute a Python module, as in:
python3 -m slipcover -m pytest -x -v
which starts pytest, passing it any options (-x -v in this example)
after the module name.
No plug-in is required for pytest.
$ python3 -m slipcover -m pytest
================================================================ test session starts ================================================================
platform darwin -- Python 3.9.12, pytest-7.1.2, pluggy-1.0.0
rootdir: /Users/juan/project/wally/d2k-5, configfile: pytest.ini
plugins: hypothesis-6.39.3, mock-3.7.0, repeat-0.9.1, doctestplus-0.12.0, arraydiff-0.5.0
collected 439 items
tests/box_test.py ......................... [ 5%]
tests/image_test.py ............... [ 9%]
tests/network_equivalence_test.py .........................................s................................................................. [ 33%]
.............................................................................. [ 51%]
tests/network_test.py ....................................................................................................................... [ 78%]
............................................................................................... [100%]
=================================================== 438 passed, 1 skipped, 62 warnings in 48.43s ====================================================
File #lines #miss Cover% Lines missing
--------------------------------- -------- ------- -------- ------------------------
d2k/__init__.py 3 0 100
d2k/box.py 105 27 74 73, 142-181
d2k/image.py 38 4 89 70-73
d2k/network.py 359 1 99 236
tests/box_test.py 178 0 100
tests/darknet.py 132 11 91 146, 179-191
tests/image_test.py 45 0 100
tests/network_equivalence_test.py 304 30 90 63, 68, 191-215, 455-465
tests/network_test.py 453 0 100
$
As can be seen in the coverage report, d2k lacks some coverage, especially in
its box.py and image.py components.
Our GitHub workflows run the automated test suite on Linux, MacOS and Windows, but really it should work anywhere where CPython/PyPy does.
SlipCover is under active development; contributions are welcome! Please also feel free to create a new issue with any suggestions or issues you may encounter.
For more details about how SlipCover works please see the following paper, published at ISSTA'23: SlipCover: Near Zero-Overhead Code Coverage for Python.
Logo design by Sophia Berger.
This material is based upon work supported by the National Science Foundation under Grant No. 1955610. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.
FAQs
Near Zero-Overhead Python Code Coverage
We found that slipcover demonstrated a healthy version release cadence and project activity because the last version was released less than a year ago. It has 1 open source maintainer collaborating on the project.
Did you know?
Socket for GitHub automatically highlights issues in each pull request and monitors the health of all your open source dependencies. Discover the contents of your packages and block harmful activity before you install or update your dependencies.
Security News
Research
The Socket Research Team breaks down a malicious wrapper package that uses obfuscation to harvest credentials and exfiltrate sensitive data.
Research
Security News
Attackers used a malicious npm package typosquatting a popular ESLint plugin to steal sensitive data, execute commands, and exploit developer systems.
Security News
The Ultralytics' PyPI Package was compromised four times in one weekend through GitHub Actions cache poisoning and failure to rotate previously compromised API tokens.