Python Imports and Module Caching

This is another post based on Dr. Fred Baptiste Python 3: Deep Dive series. For more, head over to Udemy.

What happens behind the scenes in a python import statement? Glad I asked. Heres how it works:

1. Check sys.modules : Python first checks sys.modules to see if your module has already been loaded.

2. Create a module object: If python does not find the module in sys.modules, it creates a new ModuleType object and loads the modules source code into it

3. Cache the module: The new module objects is added to the sys.modules cache

4. Compile and execute: Python compiles and executes the source code from your module

5. Add to globals() : Python and adds your module to globals(), making it available for use going forward.

Key Point : It is often said that importing a python module will execute all of the code in the module. That is only partly true. Python only executes a module the first time it is imported, then caches it in sys.modules. Subsequent import statements throughout your code, even in different files, will refer back to the cached version and will not execute the module again.

Below is a bit of code to bring all of this home. We will:

1. Create a simple module (module1) with a print statement so we can see when the module gets executed.

2. Import module1 into main.py, which will execute module1.

3. Delete module1 from globals(), then try accessing it to demonstrate that it is no longer available.

4. Re-import module1. This will add module1 back to globals(), but will use the cached version from sys.modules and will not re-execute module1.

5. Access module1 again to demonstrate that is has in fact been added back to globals()
   

# module1.pyprint("Running module1")# main.pyprint("======== Running main.py ======== ")print("Importing module1")import module1print("Deleting module1")del globals()["module1"] # Deletes from globals(), but not from sys.modulestry: print("Accessing module 1 after deleting from globals()") # This will cause an exception. module1 is still cached in `sys.modules`, # but the reference to the cache has been removed from globals() module1except NameError as e: print("NameError", e)print("Re-importing module1")# Adds the cached version of module1 back to globals(), without# re-executing the moduleimport module1print("Accessing module 1 after re-importing from globals()")module1print("Done!")print("=================================")

Output from main.py:

======== Running main.py ========Importing module1Running module1Deleting module1Accessing module 1 after deleting from globals()NameError name 'module1' is not definedRe-importing module1Accessing module 1 after re-importing from globals()Done!=================================

]]>