Speedy Python library to determine MIME type of file.
Defity (Detect file type) is a library for Python application to guess file type in a reliable way, not based on filename extension ( *.png, *.pdf), but on actual file content. It is like what file_ command and libmagic_ library do, but with different strategy.
📕 Documentation: defity.readthedocs.io
$ pip install defity>>> import defity
>>> defity.from_file('path/to/landscape.png')
'image/png'
>>> with open('path/to/landscape.png', 'rb') as f:
... defity.from_file(f)
...
'image/png'
>>> defity.from_bytes(b'some-binary-content')
'image/png'There are many Python libraries also do the same thing, most of them are based on wellknown libmagic. Defity is based on Rust tree_magic_mini_ library, which in turn is a fork of tree_magic_ , another Rust library. Quote from tree_magic to see how it differs from libmagic:
Unlike the typical approach that libmagic and file(1) uses, this loads all the file types in a tree based on subclasses. (EX: application/vnd.openxmlformats-officedocument.wordprocessingml.document (MS Office 2007) subclasses application/zip which subclasses application/octet-stream) Then, instead of checking the file against every file type, it can traverse down the tree and only check the file types that make sense to check. (After all, the fastest check is the check that never gets run.)
This library also provides the ability to check if a file is a certain type without going through the process of checking it against every file type.
And what tree_magic_mini has improved over tree_magic:
Reduced copying and memory allocation, for a slight increase in speed and decrease in memory use.
So, Defity should have better performance than other libraries for the same purpose.
Another advantage is that, Defity is static linked to the underlying Rust library, not depend on discrete libmagic.so. It will be easier to deploy to cloud function platforms, where you don't have control over what system libraries is present there.
In general, Defity is licensed under Apache-2.0 if it is built without tree_magic_mini_ embedded MIME database, and is licensed under GPL-3.0 otherwise. Concretely:
- On Linux, it is licensed under Apache-2.0.
- On Windows and MacOS, it is licensed under GPL-3.0.
It is because, Linux boxes already come with FreeDesktop's MIME database, Defity just uses it. Windows and MacOS don't have this database and Defity has to embed with it.
- Author: Nguyễn Hồng Quân.
- Free icon is made by Vitaly Gorbachev from flaticon.com.