Any Python file (ending with .py) can be imported by Python script. A single Python file is also called a module. This helps you to divide a bigger program into several smaller pieces.
For instance if you have a file names.py containing the following:
FIRST_NAMES = ['Alice', 'Bob', 'Charlie']
Then you can write (e.g. in a second Python file in the same directory):
import names
print(names.FIRST_NAMES)
For big programs, it is useful to divide up the code among several directories. A directory from which you can import Python modules is called a package. To create a package that Python will recognize you need to create a file __init__.py (it may be empty).
For instance, you could have the following files in a package namedata:
namedata/
__init__.py
names.py
To import from a module, a package or their contents, place its name (without .py) needs to be given in the import statement. Import statements can look like this:
import names
import names as n
from names import FIRST_NAMES
from namedata.names import FIRST_NAMES
It is strongly recommended to list the imported variables and functions explicitly and not write
from names import *
The latter makes debugging difficult.
When importing, Python generates intermediate files (bytecode) in the __pycache__ directory that help to execute programs more efficiently. It is managed automatically, and you can safely ignore it.
When importing modules or packages, Python needs to know where to find them. There is a certain sequence of directories in which Python looks for things to import:
- The current directory.
- The site-packages folder (where Python is installed).
- In directories in the
PYTHONPATHenvironment variable.
You can see all directories from within Python by checking the sys.path variable:
import sys
print sys.path