[Feature] Multimodality (Vision and Text Inputs)

[DhruvSrikanth]

[Features]

• Adds multimodality to the client. This means that users can now query with text, images or even text and images as input to their AI functions. Images can be attached through local paths, public URLs and base64 encodings
• Add HTTP2 protocol support (enabled by default for more efficient communication)
• Improve compatibility for earlier versions of Python

[Tests]

• Adds tests for text, image, and multimodal inputs. Includes tests for synchronous, asynchronous and batched queries

[Docs]

• Update docs for contributing to the package