Modules

1. Module An object that serves as a unit of python code.
2. Package Analogous to directories in a file system, they are special kinds of modules which contain submodules or subpackages. Any module with a __path__ attribute is a package.
3. Namespace A place where a variable is stored. Namespaces are implemented as dictionaries. There are the local, global and built-in namespaces as well as nested namespaces in objects. Namespaces are used to implement scope.

We can use the import statement to use code from other modules in our module.

import numpy as np
import matplotlib.pyplot  # Note this also imports maplotlib

x = np.linspace(0, 10, 1000)
matplotlib.pyplot.plot(x, x**2)
matplotlib.pyplot.show()

When a module is first imported, Python searches for the module and if found, it creates a module object. We can now access the module's global namespace from within our module using mymodule.var.

Importing a module doesn't automatically import all of its submodules, but a package may choose to do this in its __init__.py. For more info see this post.

Any directory with an __init__.py file is recognised as a valid package.

When a regular package is imported, this __init__.py file is implicitly executed, and the objects it defines are bound to names in the package’s namespace. The __init__.py file can contain the same Python code that any other module can contain, and Python will add some additional attributes to the module when it is imported.

my_package/
    __init__.py
    my_module.py
    my_other_module.py

In __init__.py:

from .my_module import Foo
from .my_other_module import foo

In my_script.py:

import my_package

# We can now use Foo and foo
x = my_package.Foo()
my_package.foo()

It's common however for an __init__.py file to be left completely empty.