pytest-watcher 0.3.0

Automatically rerun your tests on file modifications

A simple watcher for pytest

Overview

pytest-watcher is a tool to automatically rerun tests (using pytest by default) whenever your code changes.

Works on Unix (Linux, MacOS, BSD) and Windows.

Table of Contents

• Motivation
• File Events
• Installation
• Usage
• Using a different test runner
• Watching different patterns
• Delay
• Compatibility
• License

Motivation

Why not general tools (e.g. watchmedo, entr)?

• Easy to use and remember
• Works for most python projects out of the box
• Minimum dependencies (watchdog is the only one)
• Handles post-processing properly (see delay)

What about pytest-watch?

pytest-watch has been around for a long time and used to address exactly this problem. Unfortunately, pytest-watch is no longer maintained and does not work for many users. To provide a substitute, I developed this tool.

File events

By default pytest-watcher looks for the following events:

• New *.py file created
• Existing *.py file modified
• Existing *.py file deleted
• A *.py file moved either from or to the watched path

You can specify alternative file patterns to watch. See Watching different patterns

Installation

pip install pytest-watcher

Usage

Specify the path that you want to monitor:

ptw .

or

ptw /home/repos/project

Any arguments after <path> will be passed to the test runner (which is pytest by default). For example:

ptw . -x --lf --nf

will call pytest with the following arguments:

pytest -x --lf --nf

Using a different test runner

You can specify an alternative test runner using the --runner flag:

ptw . --runner tox

Watching different patterns

You can use the --patterns flag to specify file patterns that you want to monitor. It accepts a list of Unix-style patterns separated by a colon. The default value is "*.py".

Example:

ptw . --patterns '*.py,pyproject.toml'

You can also ignore certain patterns using the --ignore-patterns flag:

ptw . --ignore-patterns 'settings.py,db.py'

Delay

pytest-watcher uses a short delay (0.2 seconds by default) before triggering the actual test run. The main motivation for this is post-processors that can run after you save the file (e.g., black plugin in your IDE). This ensures that tests will be run with the latest version of your code.

You can control the actual delay value with the --delay flag:

ptw . --delay 0.2

To disable the delay altogether, you can provide zero as a value:

ptw . --delay 0

Compatibility

The code is tested for Python versions 3.7+

License

This project is licensed under the MIT License.