Pyle makes it easy to use Python as a replacement for command line tools such as sed or perl.
Pyle reads its standard input and evaluates each line with the expression specified, outputting the results on standard out.
To print the first 20 characters of each line of a file:
cat README.md | pyle -e "line[:20]"
or:
pyle -e "line[:20]" README.md
List all /tmp/ files with a filename with an even length:
ls /tmp/ | pyle -e "sh.ls('-l', line) if len(line) % 2 == 0 else None"
Perform an in-place string substitution, overwriting the original file with the updated file:
pyle -ie "re.sub(r'alien(s|)?', r'ghost\1', line)" TextAboutAliens.md
The special variableline is the current line (each line of input is evaluated through the given expression(s)). The variable words is the current line split by whitespace. To print just the URLs in an Apache access log (the seventh "word" in the line):
tail access_log | pyle -e "words[6]"
Print the SHA 256 sum of each *.py file in the current directory:
ls *.py | pyle -m hashlib -e "'%s %s' % (hashlib.sha256(line).hexdigest(), line)"
348e4a65e24bab4eed8e2bbe6f4c8176ddec60051d1918eea38b34b1103a8af6 pyle.py
b28c7f73e6df990a96cfb724be1d673c2d3c43f68d4b6c06d8e5a9b29e5d12cb pyle_test.py
If your expression returns a list or a tuple, the items will be printed joined by spaces. With that in mind we can simplify the above example:
ls *.py | pyle -m hashlib -e "(hashlib.sha256(line).hexdigest(), line)"
348e4a65e24bab4eed8e2bbe6f4c8176ddec60051d1918eea38b34b1103a8af6 pyle.py
b28c7f73e6df990a96cfb724be1d673c2d3c43f68d4b6c06d8e5a9b29e5d12cb pyle_test.py
Print the first five lines of each file with file names and line numbers:
pyle -e "'%-15s:%04d %s' % (filename, 1 + num, line) if num < 5 else None" *.py
You can also specify multiple expressions by repeating the -e option. Just
like in sed each expression acts on line (and words, etc.) as modified by
the previous expression. E.g. replace a letter in the first five words of a file:
pyle -e "words[:5]" -e "re.sub('A', 'B', line)" README.md
The idea for Pyle is based on Graham Fawcett's PyLine. Pyle is generally compatible with PyLine but requires a -e before the evaluation statement.
Pyle imports the sh module by default, which enables easy shell command execution.
Pyle is compatible with Python 2.7 and Python 3.4 and above.
pip install pyle
This file and pyle --help.
The following variables are available in the global scope:
* `line`: the current input line being processed
* `words`: line split by whitespace
* `num`: line number
* `filename`: the name of the current file
The following modules are imported by default:
* `re`: Python regular expressions
* `sh`: the [`sh` module](https://github.com/amoffat/sh)
The sh module makes it easy to run additional commands from within the expression.
Pyle can operate on a list of filenames in which case each file is read in order and evaluated line by line.
Some of us are just simply awful at remembering the sed, perl or even bash syntax but feel right at home with Python. Python code is often a little more verbose but what good is saving characters if you can't remember what they do?
Here's an example of sed vs pyle. This isn't a very good sed expression, admittedly, but the people who will find Pyle useful are not sed experts.
To change home directories from /var/X to /home/X with sed:
sed 's/^\(\([^:]*:\)\{5\}\)\/var\/\(.*\)/\1\/home\/\3/g' /etc/passwd
With Pyle:
pyle -e "re.sub(r'^(([^:]*:){5})/var/(.*)', r'\1/home/\3', line)" /etc/passwd
If you find the Python code more readable, Pyle is for you.
Tests need to be run both in Python 2 and Python 3. Best way to do that is to have one virtual environment for each. If you use virtualenv wrapper, something like this:
workon pyle2
python2 -3 -Werror -m unittest discover -p pyle_test.py
workon pyle3
python3 -m unittest discover -p pyle_test.py
Free to use and modify under the terms of the BSD open source license.