Add OCR support

[KMikeeU]

Fixes #28

Missing tests, but please let me know what you think!

[tghamm]

Hey there. Very cool! Would you be willing to add at least one test just to help me better understand flow of things?

[KMikeeU]

Added one test for uploading images and one for pdf documents.

The Document DTO currently expects the Type field to either be the string document_url or image_url. This is honestly a bit nonsensical, it would probably make more sense to use hierarchical types here (e.g. an abstract Document class that is implemented by an ImageDocument or DocumentDocument(? PdfDocument?)). Not sure what fits the project best.

[tghamm]

Hi @KMikeeU, so I think you're on the right track with an Abtract Document class, then using a custom converter to toggle between the document_url or image_url. If you look at my repo tghamm/Anthropic.SDK, I make pretty heavy use of this pattern there, and it seems to serve the sdk pretty well. Would you be willing to give that a go?

Thanks so much for contributing to the library!