Python's __loader__, what is it?

Question

I\'ve seen the term __loader__ floating around some Python files and I can\'t find any documentation on it aside from a few brief descriptions about it\'s purpose,

Answer 1

What is __loader__?

__loader__ is an attribute that is set on an imported module by its loader. Accessing it should return the loader object itself.

In Python versions before 3.3, __loader__ was not set by the built-in import machinery. Instead, this attribute was only available on modules that were imported using a custom loader.

However, this functionality changed in Python 3.3 because of PEP 0302. Now, __loader__ is available on every module that is imported:

>>> # Python 3.3 interpreter
>>> import os
>>> os.__loader__
<_frozen_importlib.SourceFileLoader object at 0x01F86370>
>>>

---

What is a loader?

A loader is an object that is returned by a finder. It uses its load_module() method to load a module into memory. importlib.abc.Loader is an example of an abstract base class for a loader.

---

What is a finder?

A finder is an object that uses its find_module() method to try and find the loader for a module. importlib.abc.Finder is an example of an abstract base class for a finder. Note however that it is deprecated in favor of importlib.abc.MetaPathFinder and importlib.abc.PathEntryFinder.

---

How can I use it, if at all?

The main use of __loader__ is for introspection. However, there are two other common uses:

1. __loader__ may be used to gather data on a specific module's loader.

2. In Python versions before 3.3, __loader__ could be used with hasattr to check whether or not a module was imported with the built-in import machinery:

   >>> # Python 3.2 interpreter
   >>> import os
   >>> hasattr(os, '__loader__')
   False
   >>>
   

   If hasattr(os, '__loader__') had returned True, it would mean that the os module was imported using a custom loader. Since it did not, it means that the module was imported using the built-in import machinery.

   Note: The above test will not work in Python 3.3+ because of the changes made by PEP 0302.