how to "reimport" module to python then code be changed after import

I have a foo.py

def foo():
    print "test"

In IPython I use:

In [6]:  import foo
In [7]:  foo.foo()
test

Then I changed the foo() to:

def foo():
    print "test changed"

In IPython, the result for invoking is still test:

In [10]:  import foo
In [11]:  foo.foo()
test

Then I use:

In [15]: del foo
In [16]:  import foo
In [17]:  foo.foo()
test

I delete the foo.pyc in same folder foo.py exists, but still no luck.

May I know how to reimport the updated code in runtime?

For Python 2.x

reload(foo)

For Python 3.x

import importlib
import foo #import the module here, so that it can be reloaded.
importlib.reload(foo)

IPython3's autoreload feature works just right.

I am using the actual example from the webpage. First load the 'autoreload' feature.

In []: %load_ext autoreload
In []: %autoreload 2

Then import the module you want to test:

In []: import foo
In []: foo.some_function()
Out[]: 42

Open foo.py in an editor and change some_function to return 43

In []: foo.some_function()
Out[]: 43

It also works if you import the function directly.

In []: from foo import some_function
In []: some_function()
Out[]: 42

Make change in some_function to return 43.

In []: some_function()
Out[]: 43

In addition to gnibbler's answer:

This changed in Python 3 to:

>>> import imp
>>> imp.reload(foo)

As @onnodb points out, imp is deprecated in favor of importlib since Python 3.4:

>>> import importlib
>>> importlib.reload(foo)

If you want this to happen automatically, there is the autoreload module that comes with iPython.

In [15] instead of del foo, use

import sys
del sys.modules["foo"]

to delete foo from the module cache