This is a collection of snippets that I tend to use frequently, or would need in the future. ############################################################################### To programmatically import a list of modules by name: _____ import importlib mods = ['os', 'sys', 'shutil', 'platform'] for m in mods: globals()[m] = importlib.import_module(m) _____ you can then use them as if you did: import os import sys import shutil import platform etc. this breaks pep-8, but sometimes you need to programmatically import modules. ref: https://www.python.org/dev/peps/pep-0008/#imports ############################################################################### To programmatically install modules via pip if they aren't installed: ____ import importlib import pip # I don't *think* pip/pypi is case-sensitive, but frequently module names are # not the same as their package names. ugh. # The key is the package name, the value is the module name. We use the above # trick here to try to import and install if it fails. mods = {'PyMySQL': 'pymysql', 'Jinja2': 'jinja2', 'psutil': None, # We show off a little bit here with this, see below. 'paramiko': None} # "" for m in mods.keys(): modname = mods[m] if not modname: modname = m try: globals()[modname] = importlib.import_module(modname) except ImportError: # We use --user to avoid conflicts with the host's python system. # pip.main() accepts all of pip (commandline)'s args! pip.main(['install', '--user', m]) try: globals()[modname] = importlib.import_module(modname) except ImportError: raise RuntimeError('Unable to install {0}!'.format(m)) ____ ############################################################################### To convert an argparse set of parsed arguments into a dict from a class, you simply do: ____ def GenArgs(): args = argparse.ArgumentParser() # args here return(args) def somefunc(): args = vars(GenArgs().parse_args()) ____ "args" in somefunc is a dict now. ############################################################################### To dynamically allocate class parameters into constants from a dict (such as from argparse - see above), do something like this: ____ class ClassName(object): def __init__(self, **kwargs): for i in kwargs.keys(): setattr(self, i, kwargs[i]) ---- ############################################################################### To store stdout and stderr to different files in a subprocess call: ---- with open('/tmp/test.o', 'w') as out, open('/tmp/test.e', 'w') as err: subprocess.run(['command'], stdout = out, stderr = err) ---- ############################################################################### To use optools logging lib (or other "shared" modules): ---- import os import re import importlib spec = importlib.util.spec_from_file_location( 'logger', '/opt/dev/optools/lib/python/logger.py') logger = importlib.util.module_from_spec(spec) spec.loader.exec_module(logger) log = logger.log(name = 'project.name') ---- ############################################################################### # TODO # https://stackoverflow.com/questions/10265193/python-can-a-class-act-like-a-module