pypi package 'zxpy'

Activity

What's new in version 1.6.3

Delta between version 1.6.2 and version 1.6.3

Source: Github

Commits:

• 0d999adf9448415e179f47edf57c67f4d6f26b65, February 20, 2023 1:11 PM:

  Add requirements.txt (#51)

• 5b0db6b7739b683ade721fa3f1ec555b1769eb03, February 21, 2023 7:53 AM:

  Make CLI arguments accessible in zxpy programs (#52)
  
  * Fix sys.argv passed to tests
  
  * Improve test assertion
  
  * Add $ arg support
  
  * Normalize /bin/sh as first argument of script_args
  
  * Add type def

• ad22c8b748eb885527828387b0abf99a2d676fa9, February 21, 2023 10:55 AM:

  Add docs for shell args

• 13adcc00f3ec15df16c4eb8540d033f21a90b974, February 21, 2023 10:55 AM:

  Bump version

Files changed:

Readme

zxpy

Shell scripts made simple 🐚

zxpy lets you seamlessly write shell commands inside Python code, to create readable and maintainable shell scripts.

Inspired by Google's zx, but made much simpler and more accessible using Python.

Rationale

Bash is cool, and it's extremely powerful when paired with linux coreutils and pipes. But apart from that, it's a whole another language to learn, and has a (comparatively) unintuitive syntax for things like conditionals and loops.

zxpy aims to supercharge bash by allowing you to write scripts in Python, but with native support for bash commands and pipes.

Let's use it to find all TODOs in one of my other projects, and format them into a table:

#! /usr/bin/env zxpy
todo_comments = ~"git grep -n TODO"
for todo in todo_comments.splitlines():
    filename, lineno, code = todo.split(':', 2)
    *_, comment = code.partition('TODO')
    print(f"{filename:40} on line {lineno:4}: {comment.lstrip(': ')}")

Running this, we get:

$ ./todo_check.py
README.md                                on line 154 : move this content somewhere more sensible.
instachat/lib/models/message.dart        on line 7   : rename to uuid
instachat/lib/models/update.dart         on line 13  : make int
instachat/lib/services/chat_service.dart on line 211 : error handling
server/api/api.go                        on line 94  : move these to /chat/@:address
server/api/user.go                       on line 80  : check for errors instead of relying on zero value

Writing something like this purely in bash or in Python would be much harder than this. Being able to use linux utilities seamlessly with a readable, general purpose language is what makes this a really powerful tool.

A larger, practical example

You can find a comparison between a practical-ish script written in bash and zxpy in EXAMPLE.md

Installation <a href="https://pypi.org/project/zxpy"><img src="https://img.shields.io/badge/pypi-zxpy-blue?style=flat"></a>

pip install zxpy

pipx

If you have pipx installed, you can try out zxpy without installing it, by running:

pipx run zxpy

Basic Examples

Make a file script.py (The name and extension can be anything):

#! /usr/bin/env zxpy
~'echo Hello world!'

file_count = ~'ls -1 | wc -l'
print("file count is:", file_count)

And then run it:

$ chmod +x ./script.py

$ ./script.py
Hello world!
file count is: 3

Run >>> help('zx') in Python REPL to find out more ways to use zxpy.

A slightly more involved example: run_all_tests.py

#! /usr/bin/env zxpy
test_files = (~"find -name '*_test\.py'").splitlines()

for filename in test_files:
    try:
        print(f'Running {filename:.<50}', end='')
        output = ~f'python {filename}'  # variables in your shell commands :D
        assert output == ''
        print('Test passed!')
    except:
        print(f'Test failed.')

Output:

$ ./run_all_tests.py
Running ./tests/python_version_test.py....................Test failed.
Running ./tests/platform_test.py..........................Test passed!
Running ./tests/imports_test.py...........................Test passed!

More examples are in EXAMPLE.md, and in the examples folder.

stderr and return codes

To get stderr and return code information out of the shell command, there is an alternative way of invoking the shell.

To use it, just use 3 variables on the left side of your ~'...' shell string:

stdout, stderr, return_code = ~'echo hi'
print(stdout)       # hi
print(return_code)  # 0

More examples are in the examples folder.

CLI Arguments

When writing a shell script, you often want to pass CLI arguments to it.

Like so:

$ cat ./foo.sh
echo arg is: $1

$ ./foo.sh 123
arg is: 123

To do the same in zxpy, pass the script arguments after a -- in the zxpy CLI command.

#!/usr/bin/env zxpy

import sys
print("Argv is:", sys.argv)

~"echo output: $1 $2 $3"

$ ./test.py
Argv is: ['/bin/sh']
output:

$ ./test.py -- abc def
Argv is: ['/bin/sh', 'abc', 'def']
output: abc def

Both $1 and sys.argv[1] will do the same thing.

Quoting

Take this shell command:

$ uname -a
Linux pop-os 5.11.0 [...] x86_64 GNU/Linux

Now take this piece of code:

>>> cmd = 'uname -a'
>>> ~f'{cmd}'
/bin/sh: 1: uname -a: not found

Why does this not work?

This is because uname -a was quoted into 'uname -a'. All values passed inside f-strings are automatically quoted to avoid shell injection.

To prevent quoting, the :raw format_spec can be used:

>>> cmd = 'uname -a'
>>> ~f'{cmd:raw}'
Linux pop-os 5.11.0 [...] x86_64 GNU/Linux

This disables quoting, and the command is run as-is as provided in the string.

Note that this shouldn't be used with external data, or this will expose you to shell injection.

Interactive mode

$ zxpy
zxpy shell
Python 3.8.5 (default, Jan 27 2021, 15:41:15)
[GCC 9.3.0]

>>> ~"ls | grep '\.py'"
__main__.py
setup.py
zx.py
>>>

Also works with path/to/python -m zx

It can also be used to start a zxpy session in an already running REPL. Simply do:

>>> import zx; zx.install()

and zxpy should be enabled in the existing session.

Development/Testing

To install from source, clone the repo, and do the following:

$ source ./venv/bin/activate  # Always use a virtualenv!
$ pip install -r requirements-dev.txt
Processing ./zxpy
[...]
Successfully installed zxpy-1.X.X
$ pytest  # runs tests