Strings  ¶

• Replace accentuated characters by their ASCII equivalent in a unicode string:

  import unicodedata
  
  unicodedata.normalize("NFKD", "éèàçÇÉȲ³¼ÀÁÂÃÄÅËÍÑÒÖÜÝåïš™").encode("ascii", "ignore")
  

• Cleanest way I found to produce slugified / tokenized strings, based on boltons.strutils  :

  >>> from boltons import strutils
  >>> strutils.slugify(" aBc De F   1 23 4! -- ! 56--78 - -9- %$& +eée-", "-", ascii=True)
  b'abc-de-f-1-23-4-56-78-9-eee'
  

  Alternative: use awesome-slugify  package.

Sorting  ¶

• Sort a list of dicts by dict-key ( source  ):

  import operator
  
  [dict(a=1, b=2, c=3), dict(a=2, b=2, c=2), dict(a=3, b=2, c=1)].sort(
      key=operator.itemgetter("c")
  )
  

Date & Time  ¶

I recommend using Arrow . But if you can’t, here are some pure-python snippets.

• Add a month to the current date:

  import datetime
  import dateutil
  
  datetime.date.today() + dateutil.relativedelta(months=1)
  

Network  ¶

• Set urllib2 timeout ( source  ):

  import socket
  
  socket.setdefaulttimeout(10)
  

• Start a dumb HTTP server on port 8000 ( source  ):

  $ python -m SimpleHTTPServer 8000
  

Debug  ¶

• Add a Python’s debugger break point:

  import pdb
  
  pdb.set_trace()
  

• Delete all .pyc and .pyo files in the system:

  $ find / -name "*.py[co]" -print -delete
  

Version  ¶

• Print Python’s 3-elements version number:

  $ python -c "from __future__ import print_function; import sys; print('.'.join(map(str, sys.version_info[:3])))"
  2.7.6
  

• Compare Python version for use in shell scripts:

  $ python -c "import sys; exit(sys.version_info[:3] < (2, 7, 9))"
  $ if [[ $? != 0 ]]; then
  >     echo "Old Python detected.";
  > fi
  Old Python detected.
  

Style  ¶

• Use autopep8 to apply PEP8’s coding style on all Python files:

  $ find ./ -iname "*.py" -print -exec autopep8 --in-place "{}" \;
  

Configuration  ¶

I maintain a set of default configuration files in my dotfiles repository  :

• PDB: ~/.pdbrc

• Pip: ~/.pip/pip.conf

• PyPi: ~/.pypirc

• Pycodestyle: ~/.config/pycodestyle

• PyLint: ~/.pylintrc

Package Management  ¶

• Generate a binary distribution of the current package:

  $ python ./setup.py sdist
  

• Register, generate and upload to PyPi the current package as a source package, an egg and a dumb binary:

  $ python ./setup.py register sdist bdist_egg bdist_dumb upload
  

• Download Pygments’ source distribution from PyPi, without dependencies ( source  ):

  $ pip download --no-binary=:all: --no-deps pygments==2.14.0
  

• Hackish way to execute the CLI above with Pip’s internal (tested with pip==22.1 ), inspired by pip._internal.cli.base_command.Command._main()  :

  from pathlib import Path
  
  from pip._internal.cli.status_codes import SUCCESS
  from pip._internal.commands.download import DownloadCommand
  from pip._internal.utils.temp_dir import global_tempdir_manager, tempdir_registry
  
  
  tmp_path = Path("/tmp")
  
  # Emulate the following CLI call:
  #   $ pip download --no-binary=:all: --no-deps pygments==2.14.0
  cmd = DownloadCommand(name="dummy_name", summary="dummy_summary")
  
  with cmd.main_context():
      cmd.tempdir_registry = cmd.enter_context(tempdir_registry())
      cmd.enter_context(global_tempdir_manager())
      options, args = cmd.parse_args(
          [
              "--no-binary=:all:",
              "--no-deps",
              "--dest",
              f"{tmp_path}",
              f"pygments==2.14.0",
          ]
      )
      cmd.verbosity = options.verbose
      outcome = cmd.run(options, args)
      assert outcome == SUCCESS
  
  package_path = tmp_path.joinpath("Pygments-2.14.0.tar.gz")
  assert package_path.is_file()
  

Jinja  ¶

To generate curly braces:

>>> from jinja2 import Template
>>> Template(""" Yo! """).render()
u' Yo! '
>>> Template(""" {{'{{'}} """).render()
u' {{ '
>>> Template(""" {{'{'}} """).render()
u' { '
>>> Template(""" {{'{'}}machin{{'}'}} """).render()
u' {machin} '

Data  ¶

• Pandas snippets