runez

Friendly misc/utils/convenience library

.. image:: https://img.shields.io/pypi/v/runez.svg :target: https://pypi.org/project/runez/ :alt: Version on pypi

.. image:: https://github.com/codrsquad/runez/workflows/Tests/badge.svg :target: https://github.com/codrsquad/runez/actions :alt: Tested with Github Actions

.. image:: https://codecov.io/gh/codrsquad/runez/branch/main/graph/badge.svg :target: https://codecov.io/gh/codrsquad/runez :alt: Test code codecov

.. image:: https://img.shields.io/pypi/pyversions/runez.svg :target: https://github.com/codrsquad/runez :alt: Python versions tested (link to github project)

Overview

runez is a convenience "utils" library for common operations I found myself rewriting multiple times.

The name was initially meant as "run ez" ("run easy"), the fact that it sounds like "runes" gives it a bit of a mystery/magic side that's also relatively appropriate (it does indeed concentrate a bit of invocation magic, as you can save quite a few lines of repetitive code by using it)

Features

• Usable with any python version

• Pure python standalone library, does not bring in any additional dependency

• Takes care of most edge cases, with nice errors

  • Functions can be called without checking for return code etc (abort by default, with nice error)

  • They can also be called with fatal=False, in which case the return value will indicate whether call succeeded or not

• Support for dryrun mode (show what would be done, but don't do it)

• Perform most typical logging setups in one call to runez.log.setup()

• Log operations systematically (at debug level mostly), examples::

  Running: foo ... Copy foo -> bar Would move foo -> bar (for dryrun)

• CaptureOutput context manager -> grab output/logging from any code section

• 100% test coverage

Example

Run a program::

import runez

# Aborts if "foo" doesn't exist
output = runez.run("ls", "foo")

# Output can also be ignored
runez.run("ls", "foo")

# Don't capture output, just run the command and let output "pass through"
runez.run("ls", "foo", stdout=None, stderr=None)

# Don't abort, return False on failure (or actual output when successful)
output = runez.run("ls", "foo", fatal=False)

File operations::

import runez

runez.touch("foo")
runez.copy("foo", "bar")
runez.move("foo", "baz")
runez.delete("foo")

runez.write("foo", "bar\nbaz\n")
content = "\n".join(runez.readlines("foo", first=10))

full_path = runez.resolved_path("foo/bar")
folder = runez.parent_folder(full_path)
runez.ensure_folder(folder)
with runez.Anchored(folder):
    assert runez.short(full_path) == "bar"

Installation

As usual, available on pypi_: pip install runez

Philosophy

runez tries to provide a consistent interface across functions. Here are the main tenets for functions involving I/O (such as writing, reading, copy-ing files etc):

All IO-related functions NOT returning content (run(), delete(), ...) have this common signature: fatal=True, logger=UNSET, dryrun=UNSET

• fatal: decides whether operation should raise an exception on failure or not

  • fatal=True (default): raise an exception on failure, log a meaningful error

  • fatal=False: don't raise on failure, log a meaningful error

  • fatal=None: don't raise on failure, don't log anything

  • In non-fatal mode, calls try to return a usable value appropriate for the call (see docstring of each function)

• logger: decides how chatty the operation should be

  • LOG.error() is used for failures, except when fatal is not True AND provided logger is a callable

  • logger=UNSET (default):

    • LOG.debug("Running: ...") to trace activity

    • print("Would run: ...") in dryrun mode

  • logger=False: Log errors only (used internally, to avoid unnecessary log chatter when one operation calls another)

  • logger=mylogger: call provided mylogger() to trace activity (example: logger=MY_LOGGER.info)

    • mylogger("Running: ...") to trace activity

    • mylogger("Would run: ...") in dryrun mode

  • logger=None: Don't log anything (even errors)

• dryrun allows to override current runez.DRYRUN setting just for that call

All IO-related functions returning content (read_json(), readlines(), ...) use a simpler convention based on: default=UNSET, which decides whether operation should raise an exception on failure or not:

• When default is NOT provided, the function call will abort on failure with an exception, logging a meaningful error via LOG.error()

• When default is provided (even if None), the function call will NOT abort, but return the specified default instead, it is up to the caller to log anything in that case (no log chatter comes from runez in that case, at all)

.. _pypi: https://pypi.org/