Modules

In this part of the Python programming tutorial, we will talk about Python modules.

A module is a file in which we have Python code. The modules in Python have the .py extension.

There are several ways to manage Python code:

Python modules are used to organize Python code. For example, database related code is placed inside a database module, security code in a security module etc. Smaller Python scripts can have one module. But larger programs are split into several modules. Modules are grouped together to form packages.

Module names

A module name is the file name with the .py extension. When we have a file called empty.py, empty is the module name. The __name__ is a variable that holds the name of the module being referenced. The current module, the module being executed (called also the main module) has a special name: '__main__'. With this name it can be referenced from the Python code.

We have two files: empty.py and modulename.py. The second module is the main module, which is executed. It imports the first module. Modules are imported using the import keyword.

$ cat empty.py
"""
An empty module
"""

This is empty.py module.

#!/usr/bin/python

import empty
import sys

print __name__
print empty.__name__
print sys.__name__

In this code example we import two modules. One built-in module sys and one custom module (empty). We print the names of modules to the console.

$ ./modulename.py 
__main__
empty
sys

The name of the module, which is being executed is always '__main__'. Other modules are named after the file name. Modules can be imported into other modules using the import keyword.

Locating modules

When a module is imported the interpreter first searches for a built-in module with that name. If not found, it then searches in a list of directories given by the variable sys.path. The sys.path is a list of strings that specifies the search path for modules. It consists of the current working directory, directory names specified in the PYTHONPATH environment variable plus some additional installation dependent directories. If the module is not found, an ImportError is raised.

#!/usr/bin/python

import sys
import textwrap

sp = sorted(sys.path)
dnames = ', '.join(sp)

print textwrap.fill(dnames)

The script prints all directories from sys.path variable.

import textwrap

The textwrap module is used for easy formatting of paragraphs.

sp = sorted(sys.path)

We retrieve a list of directories from the sys.path variable and sort them.

dnames = ', '.join(sp)

We make a string out of the list.

$ ./syspath.py 
/home/janbodnar/programming/python/modules,
/usr/lib/pymodules/python2.7, /usr/lib/python2.7, /usr/lib/python2.7
/dist-packages, /usr/lib/python2.7/dist-packages/PIL,
...

Sample output.

The import keyword

The import keyword can be used in several ways.

from module import *

This construct will import all Python definitions into the namespace of another module. There is one exception. Objects beginning with underscore character _ are not imported. They are expected to be used only internally by the module being imported. This way of importing modules is not recommended.

#!/usr/bin/python

from math import *

print cos(3)
print pi

This import construct has imported all definitions from the built-in math module. We can call the math functions directly, without referencing the math module.

$ ./everything.py 
-0.9899924966
3.14159265359

The use of this import construct may result in namespace pollution. We may have several objects of the same name and their definitions can be overridden.

#!/usr/bin/python

from math import  *

pi = 3.14

print cos(3)
print pi

The example will print 3.14 to the console. Which may not be, what we wanted. The namespace pollution may become critical in larger projects.

The following example will show definitions that are not being imported using this import construct.

#!/usr/bin/python

"""
names is a test module
"""

_version = 1.0

names = ["Paul", "Frank", "Jessica", "Thomas", "Katherine"]

def show_names():
   for i in names:
      print i

def _show_version():
   print _version

This is the names.py module.

#!/usr/bin/python

from names import *

print locals()

show_names()

The _version variable and the _show_version() function are not imported into the private.py module. We do not see them in the namespace. The locals() function give us all the definitions available in the private module.

$ ./private.py 
{'__builtins__': <module '__builtin__' (built-in)>, '__file__': './private.py', 
'show_names': <function show_names at 0xb7dd233c>, 
'names': ['Paul', 'Frank', 'Jessica', 'Thomas', 'Katherine'],
 '__name__': '__main__', '__doc__': None}
Paul
Frank
Jessica
Thomas
Katherine
from module import fun, var

This import construct imports only specific objects from a module. This way we import only definitions that we need.

#!/usr/bin/python

from math import sin, pi

print sin(3)
print pi

We import two objects from the math module. There is no way, how we could reference other definitions like e.g. a cos function.

#!/usr/bin/python

from names import _version, _show_version

print _version
_show_version()

We could also import definitions beginning with an underscore. But this is a bad practice.

$ ./imnames.py 
1.0
1.0
import module

The last construct is most widely used. It prevents the namespace pollution and enables to access all definitions from a module.

#!/usr/bin/python

import math

pi = 3.14

print math.cos(3)
print math.pi
print math.sin(3)
print pi

In this case, we reference the definitions via the module name. As we can see, we are able to use both pi variables. Our definition and the one from the math module.

$ ./widely.py 
-0.9899924966
3.14159265359
0.14112000806
3.14
#!/usr/bin/python

# importas.py

import math as m

print m.pi
print m.cos(3)

We can change the name through which we can reference the module. To do this, we use the as keyword.

$ ./importas.py 
3.14159265359
-0.9899924966

An ImportError is raised if a module cannot be imported.

#!/usr/bin/python

try:
    import empty2
except ImportError, e:
    print 'Failed to import:', e

We have not created an empty2 module. Therefore an exception is raised.

$ ./importerror.py
Failed to import: No module named empty2

Example output.

Executing modules

Modules can be imported into other modules or they can be also executed. Module authors often create a testing suite to test the module. Only if the module is executed as a script, the __name__ attribute equals to __main__.

We will demonstrate this on a fibonacci module. Fibonacci numbers is a sequence of numbers, where each is the sum of its two immediate predecessors.

#!/usr/bin/python

"""
A module containing the fibonacci
function. 
"""

def fib(n):
    a, b = 0, 1
    while b < n:
        print b,
        (a, b) = (b, a + b)


# testing

if __name__ == '__main__':
   fib(500)

The module can be normally imported as usual. The module can be also executed.

$ ./fibonacci.py 
1 1 2 3 5 8 13 21 34 55 89 144 233 377

If we do import the fibonacci module, the test is not executed automatically.

>>> import fibonacci as fib
>>> fib.fib(500)
1 1 2 3 5 8 13 21 34 55 89 144 233 377

The fibonacci module is imported and the fib() function is executed.

The dir() function

The built-in dir() function gives a sorted list of strings containing the names defined by a module.

#!/usr/bin/python

"""
This is dirfun module 
"""

import math, sys

version = 1.0

names = ["Paul", "Frank", "Jessica", "Thomas", "Katherine"]

def show_names():
  
   for i in names:
      print i

print dir(sys.modules['__main__']) 

In this module, we import two system modules. We define a variable, a list and a function.

print dir(sys.modules['__main__']) 

The dir() function returns all the names available in the current namespace of the module. '__main__' is the name of the current module. The sys.modules is a dictionary that maps module names to modules which have already been loaded.

$ ./dirfun.py
['__builtins__', '__doc__', '__file__', '__name__', '__package__', 'math', 
'names', 'show_names', 'sys', 'version']

We can see some built-in names like '__file__' or '__name__' and all the others that we have defined.

The globals() function

The globals() function returns a dictionary that represents the current global namespace. It is a dictionary of global names and their values. It is the dictionary of the current module.

#!/usr/bin/python

import sys
import textwrap

version = 1.0

def myfun():
    pass

gl = globals()    
gnames = ', '.join(gl)

print textwrap.fill(gnames) 

We use the globals() function to print all the global names of the current module.

$ ./globalsfun.py
__builtins__, __file__, textwrap, __package__, sys, myfun, version,
__name__, gl, __doc__

These are the global names of the current module.

The __module__ attribute

The __module__ class attribute has the name of the module in which the class is defined.

"""
module animals
"""

class Cat:
  pass

  
class Dog:
  pass

This are the contents of the animals.py file. We have two classes.

#!/usr/bin/python

from animals import Cat

class Being:
    pass
  
  
b = Being()
print b.__module__

c = Cat()
print c.__module__

In this code we use the __module__ attribute.

from animals import Cat

From the animals module, we import the Cat class.

class Being:
    pass

In the current module, we define a class Being.

b = Being()
print b.__module__

An instance of the Being class is created. We print the name of its module.

c = Cat()
print c.__module__

We create an object from the Cat class. We also print the module, where it was defined.

$ ./mclass.py
__main__
animals

The current module's name is __main__. And the Cat's module name is animals.

This chapter was about modules in Python.