logrus

Table of Contents

• Installation
• License
• Included Commands
  • human-time
• Included functions
  • thelogrus.cli
    • exec_subcommand(unfound)
    • find_subcommand(args)
    • is_program(name)
    • run(command)
  • thelogrus.logging
    • getCustomLogger(name, logLevel)
  • thelogrus.time
    • humanTime(seconds)
  • thelogrus.utils
    • mkdir_p(path)
    • obfuscateString(snippet, showLength=5, smear='*')
    • squashDicts(*dict_args)
  • thelogrus.yaml
    • readYamlFile(path: str)
    • writeYamlFile(path: str, data)

The logrus is a collection of random utility functions. Nothing in here is all that special, they're just yet another implementation of functions I've rewritten at every job to use in various utility scripts. By open sourcing them now, I'm hoping to not have to write them yet again.

And yes, the name is from Zelazny's Amber series.

Installation

pip install thelogrus

License

Apache 2.0 license.

Included Commands

human-time

Takes a value in seconds either from stdin or as arg 1 and converts it to a more meat-friendly format using the humanTime function.

human-time 1234 will print "20 minutes, 34 seconds"

Included functions

thelogrus.cli

exec_subcommand(unfound)

Creates a git-style driver command. If your script is named foo, and is run as foo bar baz and there is an executable in your $PATH named foo-bar, it will call foo-bar with baz as the command-line argument.

unfound is an optional argument that should be a function pointer and will be called if exec_subcommand can't find a suitable subcommand. Mainly useful for you to have a custom usage message.

Example usage:

#!/usr/bin/env python3
#
# Test script for thelogrus
#
# Confirms that thelogrus.cli.exec_subcommand works as expected
#
# Copyright 2019, Joe Block <jpb@unixorn.net>

import sys
from thelogrus.cli import exec_subcommand


def _usage(message):
  '''
  Custom usage printer
  '''
  print("%s" % sys.argv[0])
  print("Called as %s" % (' '.join(sys.argv)))
  print("Oh look, a custom usage message.")
  print("Attempted to find an executable using all the permutations of %s with no luck." % '-'.join(sys.argv))
  print("%s" % message)


if __name__ == '__main__':
  exec_subcommand(unfound=_usage)

find_subcommand(args)

Given a list ['foo','bar', 'baz'], attempts to create a command name in the format foo-bar-baz. If that command exists, we run it. If it doesn't, we check to see if foo-bar exists, in which case we run foo-bar baz. We keep slicing chunks off the end of the command name and adding them to the argument list until we find a valid command name we can run.

This allows us to easily make git-style command drivers where for example we have a driver script, foo, and subcommand scripts foo-bar and foo-baz, and when the user types foo bar foobar we find the foo-bar script and run it as foo-bar foobar

Example usage:

#!/usr/bin/env python3

import os
import subprocess
import sys
from thelogrus.cli import find_subcommand


def subcommander_driver():
  '''
  Process the command line arguments and run the appropriate subcommand.

  We want to be able to do git-style handoffs to subcommands where if we
  do `foo blah foo bar` and the executable foo-blah-foo exists, we'll call
  it with the argument bar.

  We deliberately don't do anything with the arguments other than hand
  them off to the foo subcommand. Subcommands are responsible for their
  own argument parsing.
  '''
  try:
    (command, args) = find_subcommand(sys.argv)

    # If we can't construct a subcommand from sys.argv, it'll still be able
    # to find this driver script, and re-running ourself isn't useful.
    if os.path.basename(command) == sys.argv[0]:
      print("Could not find a subcommand for %s" % ' '.join(sys.argv))
      sys.exit(1)
  except Exception as e:
    print(str(e))
    sys.exit(1)
  subprocess.check_call([command] + args)


if __name__ == '__main__':
  subcommander_driver()

is_program(name)

Search for a given program in $PATH, and return True if it exists and is executable.

run(command)

Runs a command (either a str or list) and returns its stdout.

thelogrus.logging

getCustomLogger(name, logLevel)

Returns a custom logger with nicely formatted output.

thelogrus.time

humanTime(seconds)

Takes a value in seconds, returns it in meat-friendly format. humanFriendlyTime(8675309) would return "100 days 9 hours 48 minutes 29 seconds".

thelogrus.utils

mkdir_p(path)

os module doesn't have a mkdir -p equivalent so added one.

obfuscateString(snippet, showLength=5, smear='*')

Takes a string, chops off showLength characters from the beginning and end and replaces everything else with the smear character. Really only useful for displaying just enough of a credential in a log to show it's using the correct credential, but not revealing the actual credential.

squashDicts(*dict_args)

Return a dict that is all the dict_args squashed together.

thelogrus.yaml

readYamlFile(path: str)

Return the data structure contained in a YAML file

Args: path (str): Path to read from

Returns: Data decoded from YAML file content

writeYamlFile(path: str, data)

Writes a data structure or object into a YAML file

Args: path (str): Path to data file data (any): Data to convert and write