micropytest 0.19.0

A micro test runner

microPyTest

microPyTest is a minimal, pure python-based test runner that you can use directly in code.

Key Points

• Code-first approach: Import and run tests from your own scripts.
• Artifact tracking: Each test can record artifacts (files or data) via a built-in test context.
• Command execution: Run and interact with external processes with real-time output processing.
• Test filtering: Run only the tests you need by specifying patterns.
• Test arguments: Pass command-line arguments directly to your tests.
• Lightweight: Just Python. No special config or advanced fixtures.
• Optional CLI: You can also run tests via the micropytest command, but embedding in your own code is the primary focus.

Installation

Install directly from PyPI:

pip install micropytest

Usage in Code

Suppose you have some test files under my_tests/:

# my_tests/test_example.py
def test_basic():
    assert 1 + 1 == 2

def test_with_context(ctx):
    ctx.debug("Starting test_with_context")
    assert 2 + 2 == 4
    ctx.add_artifact("numbers", {"lhs": 2, "rhs": 2})

You can run them from a Python script:

import micropytest.core

results = micropytest.core.run_tests(tests_path="my_tests")
passed = sum(r.status == "pass" for r in results)
total = len(results)
print(f"Test run complete: {passed}/{total} passed")

• Each test that accepts a ctx parameter gets a TestContext object with .debug(), .warn(), .add_artifact(), etc.
• Results include logs, artifacts, pass/fail/skip status, and duration.

Command Execution

microPyTest includes a Command class for running and interacting with external processes:

from micropytest.command import Command
import sys

def test_interactive_command(ctx):
    # Run a Python interpreter interactively
    with Command([sys.executable, "-i"]) as cmd:
        # Send a command
        cmd.write("print('Hello, world!')\n")
        
        # Check the output
        stdout = cmd.get_stdout()
        
        # Exit the interpreter
        cmd.write("exit()\n")
    
    # Verify the output
    assert any("Hello, world!" in line for line in cmd.get_stdout())

Key features:

• Run commands with callbacks for real-time output processing
• Interact with processes via stdin
• Access stdout/stderr at any point during execution
• Set custom environment variables and working directories
• Allow ignoring test files by using .micropytestignore files.

Test Filtering

You can run a subset of tests by specifying a filter pattern:

# Run only tests with "artifact" in their name
results = micropytest.core.run_tests(tests_path="my_tests", test_filter="artifact")

This is especially useful when you're focusing on a specific area of your codebase.

Passing Arguments to Tests

Tests can accept and parse command-line arguments using standard Python's argparse:

def test_with_args(ctx):
    import argparse
    
    # Create parser
    parser = argparse.ArgumentParser(description="Test with arguments")
    parser.add_argument("--string", "-s", default="default string", help="Input string")
    parser.add_argument("--number", "-n", type=int, default=0, help="Input number")
    
    # Parse arguments (ignoring unknown args)
    args, _ = parser.parse_known_args()
    
    # Log the parsed arguments
    ctx.debug(f"Parsed arguments:")
    for key, value in vars(args).items():
        ctx.debug(f"  {key}: {value}")
    
    # Use the arguments in your test
    assert args.string != "", "String should not be empty"
    assert args.number >= 0, "Number should be non-negative"

When running from the command line, you can pass these arguments directly:

micropytest --test test_with_args --string="Hello World" --number=42

The arguments after your test filter will be passed to your test functions, allowing for flexible test parameterization.

Differences from pyTest

• Code-first: You typically call run_tests(...) from Python scripts. The CLI is optional if you prefer it.

• Artifact handling is built-in: ctx.add_artifact("some_key", value) can store data for later review. No extra plugin required.

• Command execution built-in: No need for external plugins to run and interact with processes.

• No fixtures or plugins: microPyTest is intentionally minimal. Tests can still share state by passing a custom context class if needed.

• No configuration: There's no pytest.ini or conftest.py. Just put your test functions in test_*.py or *_test.py.

• Time estimates for each test

Quickstart

See the examples subfolder

Optional CLI

If you prefer a command-line flow:

micropytest -p tests/

• --verbose: Show all debug logs & artifacts.
• --quiet: Only prints a final summary.
• --test: Run only tests matching the specified pattern.
• --dry-run: Show what tests would be run without actually running them (will assume them to pass).

Examples:

# Run all tests in my_tests directory
micropytest -v my_tests

# Run only tests with "artifact" in their name
micropytest -t artifact my_tests

# Run a specific test and pass arguments to it
micropytest -t test_cmdline_parser --string="Hello" --number=42

Development

To develop with microPyTest, install the required dependencies and run the tests in the project:

pip install rich
python -m micropytest .

Uploading to PyPI

To build and upload a new version to PyPI:

# Install build tools
pip install build twine

# Build the distribution packages
python -m build

# Upload to PyPI
python -m twine upload dist/micropytest-xxx.tar.gz

Make sure to update the version number in __init__.py and pyproject.toml before building a new release.

Changelog

• v0.19.0 – Add support for ignoring test files by using .micropytestignore files
• v0.18.5 – Fix bugs in TestStore (artifacts and logs not being transmitted)
• v0.18.4 – Fix bug in TestStore
• v0.18.3 – Fix bugs related to KeepAlive
• v0.18.2 – Fix error handling, fix bug in SVN helper
• v0.18.1 – Transmit logs and artifacts asynchronously for improved performance
• v0.18.0 – Allow passing Command callbacks and timeout when using context manager
• v0.17.8 – Add TestAttributes.with_args()
• v0.17.7 – Fix problem with KeyboardInterrupt not being raised on headless Windows
• v0.17.6 – Fix exception cases in Command
• v0.17.5 – Improve error handling in test store
• v0.17.4 – Terminate also child processes in Command.terminate()
• v0.17.3 – Bugfix in Command.wait()
• v0.17.2 – Refactor logging
• v0.17.1 – Bugfix in SVN helper
• v0.17 – Improve SVN helper
• v0.16 – Improve test store terminology
• v0.15 – Add support for environment variables for SVN credentials
• v0.14 – Add support for storing test results on a remote server using REST API, make core functions sync
• v0.13 – Make ctrl+c responsively abort tests, fix bugs, add VCS helper functions
• v0.12 – Improved type safety, allow detecting changes using VCS
• v0.11.1 – Add dry-run mode
• v0.11 – Add support for parameterized tests
• v0.10 – Refactoring, make VSCHelper extensible, breaking: you need to create an instance of VCSHelper to use its functions
• v0.9.2 – Fix dependency issues and require Python 3.9 or newer
• v0.9.1 – Make pyproject.toml compatible with Python older than 3.9 (attempt did not work)
• v0.9 – Several fixes
• v0.8 – Several fixes and minor improvements
• v0.7 – Added asyncio support
• v0.6 – Added rich display support, tag filtering, improved warnings display, VCS helper, and improved command execution
• v0.5 – Added test filtering and argument passing capabilities
• v0.4 – Added Command class for process execution and interaction
• v0.3.1 – Fixed screenshot in pypi
• v0.3 – Added ability to skip tests
• v0.2 – Added support for custom context classes
• v0.1 – Initial release

Enjoy your micro yet mighty test runner!