ftguess: fallback to magika (or others) when file type not identified

[decalage2]

This would be mostly useful for text-based formats, which cannot be easily identified.

https://opensource.googleblog.com/2024/02/magika-ai-powered-fast-and-efficient-file-type-identification.html
https://github.com/google/magika

Compared to other solutions, which are mostly magic-based, magika should provide better results for text-based formats such as JavaScript, VBScript, PowerShell, etc.

Other potential solutions:

• https://github.com/h2non/filetype.py (pure python, no dependencies)
• https://pypi.org/project/file-magic/
• https://github.com/cdgriffith/puremagic
• https://github.com/ahupp/python-magic
• https://filemagic.readthedocs.io/en/latest/
• TrID
• etc