Python – modules, packages, from, import basics.

A module is basically a single file, containing executable code, and class and function definitions.

When a module is imported all code in it is executed.

For example:

/modules
a.py
b.py

a.py:

import b

print "this is module a"

def function_a():
print "this is function a"

b.py:

print "this is module b"

def function_b():
    	print "this is function b"

run:

>python a.py
this is module b
this is module a

However, if you are declaring variables in the imported module, you need to reference the variable’s namespace. 

eg.

b.py:

var_b = 10

def function_b():
     print "this is function b"

class class_b:
     class_var_b = 10
     def class_function_b(self):
         print "this is class function b"

a.py:

import b

b.function_b()
print b.var_b
my_class = b.class_b()
my_class.class_function_b()

output:

>python a.py
this is function b
10
this is class function b

Using ‘from’. 

When you use from, you no longer need to specify the namespace, but all code within the imported module is still executed.

b.py:

print "this is module b"

var_b = 10

def function_b():
	print "this is function b"

a.py:

from b import function_b

print "this is module a" 

def function_a():
	print "this is function a"

function_a()
function_b()
print var_b
[/python]

output:

>python a.py
this is module b
this is module a
this is function a
this is function b
<snip> NameError: name 'var_b' is not defined

Here, we can access function_b because we explicitly imported it, but we can't access var_b because we haven't imported it.

Packages

We can use packages to organise our modules. Packages are a essentially a just a folder, and modules then include the folder name in the module's namespace.

Packages require an __init__.py in the folder in order to be detected.

/modules
    /folder_a
        __init__.py
        a.py
    main.py

__init__.py:

print "this is folder_a __init__.py"

a.py:

print "this is module a"

def function_a():
	print "this is function a"
	

main.py:

import folder_a.a
print "this is module main"

folder_a.a.function_a()

output:

>python main.py
this is folder_a __init.py
this is module a
this module main
this is function a

You can use 'as' to reduce the namespace length

eg.

import folder_a.a as a
print "this is module main"

a.function_a()

Putting imports into your __init__.pys

Whenever the a package is referenced, the code in the __init__.py is exectuted.

This means that you can include import statements in your __init__.py.

For example you can import the whole package, instead of individually importing the package contexts (or using import *).

eg.

/modules
    /folder_a
        __init__.py
        a.py
        a2.py
    main.py

__init__.py:

print "this is folder_a __init__.py"
import a, a2

a.py:

print "this is module a"

def function_a():
    print "this is function a"
    	

a2.py:

print "this is module a2"

def function_a_two():
	print "this is function a two"    	

main.py:

import folder_a 
print "this is module main"

folder_a.a.function_a()
folder_a.a2.function_a_two()

output:

>python main.py
this is folder_a __init.py
this is module a
this is module a2
this module main
this is function a
this is function a two

What if two modules contain a method with the same name?

If you are straight importing the modules, then this won't be a problem, as the method's namespace will distingush the two.

However, if you are using the 'from module import method' method, then the method that was important second will take precedent.

/modules
main.py
a.py
b.py

main.py:

from a import foo
from b import foo
foo()

a.py:

print "this is module a"

def foo():
	print "module a foo"

b.py:

print "this is module b"

def foo():
	print "module b foo"
	

output:

>python main.py
this is module a
this is module b
module b foo

eg.

 

Imports - from and import statements:

Here is a good stack overflow question that discusses 'from module import foo' vs 'import module'.

Resources

Namespaces http://bytebaker.com/2008/07/30/python-namespaces/