Namespaces

Let’s learn a bit about how Python keeps track of names

Namespaces and Modules

Namespaces are the way that Python organizes variable names

For example, suppose I write a script math2.py like this:

# Filename: math2.py
pi = 'foobar'

Now I start the Python interpreter and import it

>>> import math2

Next I import the math module from the standard library

>>> import math

Both of these modules have an attribute pi:

>>> math.pi
3.1415926535897931
>>> math2.pi
'foobar'

How is it that Python does not get confused with these two pi?

• Python keeps track of names by storing them in different namespaces
• Different namespaces can contain the same name, as in the example above

Whenever Python executes a module, it creates a namespace for that module

In Python, namespaces are implemented as dictionaries

We can look at the dictionary directly, using moduleName.__dict__:

>>> import math
>>> math.__dict__
{'pow': <built-in function pow>, ..., 'pi': 3.1415926535897931,...}

This namespace is created when we execute import math

When we access elements of the namespace using the dotted attribute notation

>>> math.pi
3.1415926535897931

this is in fact equivalent to math.__dict__['pi']

>>> math.__dict__['pi'] == math.pi
True

Viewing Namespaces

As we saw above, the math namespace can be printed by typing math.__dict__

Another way to see its contents is to type vars(math)

>>> vars(math)
{'pow': <built-in function pow>,...

Finally, if you just want to see the names, you can type

>>> dir(math)
['__doc__', '__file__', '__name__', 'acos', 'asin', 'atan',...

Notice the special names __doc__, __file__ and __name__

These are initialized in the namespace when any module is imported

• __doc__ is the doc string of the module
• __file__ is the name of the file where the module code was read from
• __name__ is the name of the module

>>> print math.__doc__
This module is always available.  It provides access to the
mathematical functions defined by the C standard.
>>> math.__file__
'/usr/lib/python2.5/lib-dynload/math.so'
>>> math.__name__
'math'

Interactive Sessions

In Python, commands typed at the prompt >>> are regarded as part of a module called __main__

An assignment such as

>>> x = 4

is stored in the namespace of __main__

To see the contents of __main__ use vars() rather than vars(__main__)

>>> vars()  # After starting Python, before making any assignments
{'__builtins__': <module '__builtin__' (built-in)>,
'__name__': '__main__',
'__doc__': None}

Now let’s make an assignment

>>> x = 4
>>> vars()
{'__builtins__': <module '__builtin__' (built-in)>,
'__name__': '__main__',
'__doc__': None,
'x': 4}

The variable x has been registered in the namespace

If we want to see just the names in __main__, we can use dir()

>>> dir()
['__builtins__', '__doc__', '__name__', 'x']

Running Scripts

Next, suppose we have a file called mod1.py with the following code:

x = 3
def f():
    pass  # 'pass' means do nothing

To use this code, one option is to import it

>>> import mod1

In this case, the code in the script is regarded as part of the module mod1.py

• Names x and f are registered in the namespace of mod1
• Accessed with dot notation: mod1.x, mod1.f

Another option is to run it interactively

• In IPython: run mod1.py
• In IDLE: Run Module
• etc., etc.

In this case, the code in the script is regarded as part of __main__

• Same as if we had typed the commands at the prompt
• x and f are registered in namespace of __main__
• Can be accessed directly (x rather than mod1.x, etc)

Another way to see the difference between running as __main__ and importing:

Let’s say we have a script mod2.py as follows

# Filename: mod2.py
print __name__

Now let’s look at two different ways of running it in IPython

In [1]: import mod2  # Standard import
mod2

In [2]: run mod2.py  # Run interactively
__main__

In the second case, the code is executed as part of __main__, so __name__ = '__main__'

Importing Modules, Again

Suppose we are working interactively, at the prompt >>>

When we import a module,

• the module name is registered in the namespace of __main__

>>> dir()
['__builtins__', '__doc__', '__name__']
>>> import math
>>> dir()
['__builtins__', '__doc__', '__name__', 'math']

• and the namespace of math is created

>>> dir(math)
['__doc__', '__file__', '__name__', 'acos', 'asin', 'atan',...

We can also import names directly into the current namespace using from

>>> from math import pi, e
>>> pi
3.1415926535897931
>>> e
2.7182818284590451
>>> dir()
['__builtins__', '__doc__', '__name__', 'e', 'math', 'pi']