GitHub | PyPi | Warehouse | ReadTheDocs | Travis-CI

Pyline: a grep-like, sed-like command-line tool (Python package)



Somewhat unsurprisingly, I found the original pyline recipe while searching for “python grep sed” (see AUTHORS.rst and LICENSE.psf).

I added an option for setting p = Path(line) in the eval/compile command context and added it to my dotfiles ; where it grew tests and an optparse.OptionParser; and is now promoted to a GitHub project with ReadTheDocs documentation, tests with tox and Travis-CI, and a for PyPi.


Pyline is an ordered MapReduce tool:

Input Readers:
  • stdin (default)
  • file (, 'r', encoding='utf-8'))
Map Functions:
  • Python module imports (-m os)

  • Python regex pattern (-r '\(.*\)')

  • path library (p from --pathpy OR --pathlib)

  • Python codeobj eval output transform:

    ls | pyline -m os 'line and os.path.abspath(line.strip())'
    ls | pyline -r '\(.*\)' 'rgx and (, or line'
    ls | pyline -p 'p and p.abspath() or ("# ".format(line))'
    # With an extra outer loop to bind variables in
    # (because (_p = p.abspath(); <codeobj>) does not work)
    find $PWD | pyline --pathpy -m os -m collections --input-delim='/' \
        'p and [collections.OrderedDict((
                ("p", p),
                ("_p", _p),
                ("_p.split()", str(_p).split(os.path.sep)),
                ("line.rstrip().split()", line.rstrip().split(os.path.sep)),
                ("l.split()", l.split(os.path.sep)),
                ("words", words),
                ("w", w)))
            for _p in [p.abspath()]][0]' \
           -O json
Partition Function:
Compare Function:
Reduce Functions:
bool(), sorted()
Output Writers:

ResultWriter classes

pyline -O csv
pyline -O tsv
pyline -O json


Install from PyPi:

pip install pyline

Install from GitHub as editable (add a pyline.pth in site-packages):

pip install -e git+


Print help:

pyline --help


# Print every line (null transform)
cat ~/.bashrc | pyline line
cat ~/.bashrc | pyline l

# Number every line
cat ~/.bashrc | pyline -n l

# Print every word (str.split(input-delim=None))
cat ~/.bashrc | pyline words
cat ~/.bashrc | pyline w

# Split into words and print (default: tab separated)
cat ~/.bashrc | pyline 'len(w) >= 2 and w[1] or "?"'

# Select the last word, dropping lines with no words
pyline -f ~/.bashrc 'w[-1:]'

# Regex matching with groups
cat ~/.bashrc | pyline -n -r '^#(.*)' 'rgx and'
cat ~/.bashrc | pyline -n -r '^#(.*)'

## Original Examples
# Print out the first 20 characters of every line
tail access_log | pyline "line[:20]"

# Print just the URLs in the access log (seventh "word" in the line)
tail access_log | pyline "words[6]"

Work with paths and files:

# List current directory files larger than 1 Kb
ls | pyline -m os \
  "os.path.isfile(line) and os.stat(line).st_size > 1024 and line"

# List current directory files larger than 1 Kb
#pip install
ls | pyline -p 'p and p.size > 1024 and line'