File/MIME Type Parsing & File Extension association

[douglasg14b]

Note: This is more difficult than expected....

Rough Plan/Idea

• Figure out how file works. Reading through simplemagic source will provide a lot of context
• Parse the magic file. Again simplemagic source will be helpful here, MimeMagicSharp also performs it's own parsing
  • Consider the format used by MimeMagicSharp. It's a JSON embodiment of the magic format

Why?

Why when there are other libraries, that you listed?

• Nuget Packages
  • Not all libs have a nuget package. This is a must.
• Cleaner code
  • Existing libraries are quite messy, ignoring clean code principles and language conventions
• Better API
  • The APIs are hit or miss. I want to take their best features and extend them
• DI/IoC Compatibility
  • Static classes are good an all, but IoC and dependency injection comparability should be baked in
• Performance
  • Some implementations read entire files to check for type, this doesn't work with huge files

Motivation

I enjoy improving upon others work, and I enjoy writing good software that users can enjoy.

Resources

Other Libraries & tools

• File(1) by a wide margin, the most robust source for file identification based on the magic(5) format
  • Contains the signatures for thousands of file types
• simplemagic A Java library that mimics File(1)
  • It utilizes the magic file(s) copied from CentOS
• magic(5)
  • The format specification, which is vital to understanding how the formats work.

Existing C# Libraries:

• filetypedetective [Deprecated] [Unlicenced]
  • Succeeded by Mime-Detective
  • Good source for reference material
• Mime-Detective [MIT]
  • Uses signatures and offsets from Gary Kesler
  • Performs best-effort in differentiating XLSX, PPTX and other document formats contained inside of ZIP files
  • MimeType contains the meat of how these checks are performed.
  • MimeDetective Notable for it's LearnMimeType method, the idea of which could be improved upon and put to good use.
• MimeMagicSharp
  • Uses signatures from the Magic format.
  • Neat, multi-level, format for them
  • Don't much like how it goes about checking the signatures.

C# Code Snippets

• Example effective file signature checking in Asp.Net Core Sample Code
• Large dictionary of extension -> MIME Type mappings

Signature Resources:

• Gary Kesler
• Filesignatures.net
• The magic file on pmuch any linux distro