Python Libraries
Distinguishing between modules, packages and libraries.
Modules
Modules are simply any file that contains code. A helpful way to keep Jupyter Notebooks looking professional and easy to access is to create separate modules (modular programming) for definitions and other backend code in a separate file.
These modules can be created in an IDE or text editor. They must be .py files and they are a quick way to import functions and objects. When calling a module into your Jupyter Notebook, be sure to include the .py file in the same folder as your notebook, or refer to the path through sys.path.append('...')
Tips:
- Keep your module updated whenever changes are made. There are two ways to do this: include the magic command to actively reload all imported modules or use a reload function.
Update modules in your notebooks as they change in Python
- Jupyter Notebook will not accept magic commands or libraries from separate modules.
- Drop the py suffix in your import statement when importing.
- The Python Library Reference hosts a library of standard modules.
- Use dir to identify what is located in your module
Packages
Contain one or more modules. Packages are a special type of module that includes using dotted module names and are initiated through __init__.py.
Libraries
Libraries are essentially packages, but are typically referred to as a larger collection of packages. This is the top level of the hierarchy, serving as containers for modules/packages.
Managing Packages/Libraries
If you’re running into an error with your library, check on the version type that you are working with and then update as necessary. A quick query should reveal the latest version of a particular package available.
To view installed packages, run the following in your terminal or notebook:
conda list (Terminal)
pip list (Notebook)
To install a package, run:
# Terminal
$ conda install [package name]# Notebook
pip install [package name]
There an option to search for outdated packages and view the version type v. most updated version available:
# Notebook
pip list --outdated
Once these have been identified, here’s always the option to update all packages in your terminal:
# Terminal
$ conda update --all