Python Programming/Modules< Python Programming
Modules are a simple way to structure a program. Mostly, there are modules in the standard library and there are other Python files, or directories containing Python files, in the current directory (each of which constitute a module). You can also instruct Python to search other directories for modules by placing their paths in the PYTHONPATH environment variable.
Importing a ModuleEdit
Modules in Python are used by importing them. For example,
This imports the math standard module. All of the functions in that module are namespaced by the module name, i.e.
import math print math.sqrt(10)
This is often a nuisance, so other syntaxes are available to simplify this,
from string import whitespace from math import * from math import sin as SIN from math import cos as COS from ftplib import FTP as ftp_connection print sqrt(10)
The first statement means whitespace is added to the current scope (but nothing else is). The second statement means that all the elements in the math namespace is added to the current scope.
Modules can be three different kinds of things:
- Python files
- Shared Objects (under Unix and Linux) with the .so suffix
- DLL's (under Windows) with the .pyd suffix
Modules are loaded in the order they're found, which is controlled by sys.path. The current directory is always on the path.
Directories should include a file in them called __init__.py, which should probably include the other files in the directory.
Creating a DLL that interfaces with Python is covered in another section.
Creating a ModuleEdit
From a FileEdit
The easiest way to create a module is by having a file called mymod.py either in a directory recognized by the PYTHONPATH variable or (even easier) in the same directory where you are working. If you have the following file mymod.py
class Object1: def __init__(self): self.name = 'object 1'
you can already import this "module" and create instances of the object Object1.
import mymod myobject = mymod.Object1() from mymod import * myobject = Object1()
From a DirectoryEdit
It is not feasible for larger projects to keep all classes in a single file. It is often easier to store all files in directories and load all files with one command. Each directory needs to have a
__init__.py file which contains python commands that are executed upon loading the directory.
Suppose we have two more objects called
Object3 and we want to load all three objects with one command. We then create a directory called mymod and we store three files called
Object3.py in it. These files would then contain one object per file but this not required (although it adds clarity). We would then write the following
from Object1 import * from Object2 import * from Object3 import * __all__ = ["Object1", "Object2", "Object3"]
The first three commands tell python what to do when somebody loads the module. The last statement defining __all__ tells python what to do when somebody executes from mymod import *. Usually we want to use parts of a module in other parts of a module, e.g. we want to use Object1 in Object2. We can do this easily with an from . import * command as the following file Object2.py shows:
from . import * class Object2: def __init__(self): self.name = 'object 2' self.otherObject = Object1()
We can now start python and import mymod as we have in the previous section.
Extending Module PathEdit
When import is requested, modules are searched in the directories (and zip files?) in the module path, accessible via sys.path, a Python list. The module path can be extended as follows:
import sys sys.path.append("/My/Path/To/Module/Directory") from ModuleFileName import my_function
Above, if ModuleFileName.py is located at /My/Path/To/Module/Directory and contains a definition of my_function, the 2nd line ensures the 3rd line actually works.
- The Module Search Path at Python doc
Module names seem to be limited to alphanumeric characters and underscore; dash cannot be used. While my-module.py can be created and run, importing my-module fails. The name of a module is the name of the module file minus the .py suffix.
Module names are case sensitive. If the module file is called MyModule.py, doing "import mymodule" fails while "import MyModule" is fine.
PEP 0008 recommends module names to be in all lowercase, with possible use of underscores.
Examples of module names from the standard library include math, sys, io, re, urllib, difflib, and unicodedata.