psyh

Common Linux commands implemented in Python.

My idea is to create simple Python methods that emulate the behavior of some common Linux command line utlities. I started with grep.

Sometimes I still find myself writing shell scripts. I guess it's because they seem, at that moment, more easy to write than to write Python code. Part of the reason for this is how amazingly well-designed these command-line utilities are. Another part is that Python code often has some setup needed before your code is actually ready.

This library aims to:

• minimize the amount of setup needed to invoke shell-script-like behavior.
• provide Pythonic interfaces equivalent to the command line utilities, to combine the clarity and portability of Python together with the succintness of shell scripting.
• provide a drop-in replacement command-line interface to those utilities, so that shell scripts may easily be ported to other operating systems, namely Windows.

Installation

Requirements:

• regex (for -E support)

Usage

From the command line:

python psyh.py grep [grep options]

Using the grep drop-in replacement function grep_sh:

import psyh, shlex
psyh.grep_sh(shlex.split("-P -o -i '(^|[^\w]|_)t\w*' " + __file__))

Using the grep() function:

import psyh
for match_tuple in psyh.grep(
		matcher=psyh.PcreMatcher(only_matching=True, ignore_case=True),
		inputs=[open(__file__)],
		patterns=[r'(^|[^\w]|_)t\w*']):
	print(match_tuple)

Known Issues

• It is significantly slower than grep (at least x4 and up to x7 slower).
• PcreMatcher is not wholly compatible with grep -P, and PosixMatcher is only loosely compatible with grep -E.
• The -o (--only-matching) might result in unclear behavior when used in conjunction with -v (--invert-match) or, if context is shown (i.e. -A (--after-context), -B (--before-context) or -C (--context)). This implementation prevents such usage.
• Asking for context lines with -NUM is not supported. Can't figure out how to do this with argparse.
• Errors are displayed as Python tracebacks, not like grep errors.

Todo

• grep_sh() prints, but there should probably be an intermediate implementation between it and grep(), that for instance handles all the context line overlap stuff.
• Implement the following options:
  • Matcher Selection
    • -E, --extended-regexp (properly implement ERE as PosixMatcher)
    • -G, --basic-regexp (implement BRE as BasicMatcher)
  • Matching Control
    • -w, --word-regexp
  • General Output Control
    • --color
  • Output Line Prefix Control
    • --byte-offset
    • -T, --initial-tab
    • -u, --unix-byte-offsets
    • -Z, --null
  • File and Directory Selection - traverse directories and do globbing, handle binary files (namely, -R)
  • Other Options
• Write nicer usage message (help)

History

• 2017-04-23: initial version

Credits

This implementation was written by Yuval Sedan. It is based on the interface to GNU grep 2.25.

License

See LICENSE.