Python - Modules

• Standard Modules - Many/all of these are built in modules for example sys

name variable with name of current module

Define a module like this:

crunchy.py


def crusty(x):
    x = x += 1
    for i in range(10):
        result += x * i
    return result

def crab(x):   # return Fibonacci series up to n
    x = x += 2
    for i in range(50):
        result += x * i
    return result

Use the module like this:

import crunchy

crunchy.crusty(1000)
crunchy.crab(100)
crunchy.__name__

crusty = crunchy.crusty
crusty(45)

• Modules contain code that is executed when they are imported for the first time ( used for initialization ).
• They can also be run as a script.

• Modules can import other modules

• They contain global variables:
  • Global vars within the module but won’t clash with global variables from importing module/script.
  • Accessible from scripts that import the module ( dot notation )

Import names directly into importing module’s name space:

from crunchy import crusty, crab
crab(500)

You can import anything that doesn’t start with _. This could result in name conflict and is frowned upon. Here is how you do it:

from crunchy import *
crab(500)

import module with different name:

import crunchy as cr
cr.crab(500)

Import fuction with different name:

from crunchy import crab as snipper
snipper(500)

If you want to reload a module that has been changed without restarting the interpreter:

import importlib;
importlib.reload(crunchy)

Run python module as a script:

python crunchy.py

Instructions to be run only when executed as a script ( great for testing ):

if __name__ == "__main__":
    x = sys.argv[1]
    crab(int(x))

Module Search Path

First search built-in modules:

sys.builtin_module_names

Then search in the directories listed here ( can be overwritten ):

sys.path

sys.path is initialized by:

• current script dir
• PYTHONPATH
• installation dependent default ( including site-package dir )

The current script dir is prioritized.

• reference doc

import sys
sys.path.append('/home/user1/my_libs')

Compiled Python files

• reference

Compiled version of each module is stored here:

__pycache__/module.version.pyc

• .pyc files load faster but doesn’t run faster

• module compileall - use to compile all modules in a dir

• module automatically recompiled if source modification time is more recent
• module recompiled and not stored if directly from the cli
• does not check cache if there is no source module

Compiled only distribution:

• compiled version needs to be in source dir
• source needs to not be there

For optimized compiled files:

-O	remove aserts
-OO	remove aserts and doc strings

dir() Function

Show names ( variables, functions, modules, etc. ) defined by a module as sorted list of strings:

import crunchy, sys
dir(crunchy)
dir(sys)
dir()      # names currently defined

The dir() function does not list built-in names, use the following to show these:

import builtins
dir(builtins)