Fix missing fromTokenizer type definition

[mooyoul]

Currently fromTokenizer function only can be accessed by importing file-type/core due to missing type definition in type index.
This PR adds missing (re-exporting) fromTokenizer type definition.

---

BTW I don't want to leave off-topic message but I really LOVE that face-lifted API. Previously i manually peeked file in AWS S3 by sending ranged request. It added more complexity for inspecting file type. but now, we can just use handy tokenizers like @tokenizers/s3 package!

@sindresorhus and @Borewit, Thanks for maintaining this project and Kudos for your work ❤️

[sindresorhus]

BTW I don't want to leave off-topic message but I really LOVE that face-lifted API. Previously i manually peeked file in AWS S3 by sending ranged request. It added more complexity for inspecting file type. but now, we can just use handy tokenizers like @tokenizers/s3 package!

That would be a great tip for the readme ;)

[Borewit]

BTW I don't want to leave off-topic message but I really LOVE that face-lifted API. Previously i manually peeked file in AWS S3 by sending ranged request. It added more complexity for inspecting file type. but now, we can just use handy tokenizers like @tokenizers/s3 package!

@mooyoul very nice to hear you are appreciating the API and someone is actually using the fromTokenizer() method. @tokenizers/s3 was a bit of an idealistic project, you may be the first user directly using it. This micro component is detached from music-metadata-s3, to allow new tokenizer consumers, like file-type, to utilize the same efficiency.

Please star @tokenizers/s3, and the dependent @tokenizer/range, to give me some confirmation these niche components are actually in use, and give newcomers some confirmation it works.

That would be a great tip for the readme ;)

I will describe the S3 use case with an small example in the readme.