Tensorshare

* Added `prepare_tensors_to_dict` for lazy formatting 39
* Added a Logger module to the package 40

* Added `TensorShareClient` for sending tensors over the network 23
* Updated `TensorProcessor` to encode and decode base64 bytes 23
* Added `TensorShareServer` schema and `from_dict` constructor for network exchange 28
* Added FastAPI integration code 28 29
* Created the `Usage` and `API Reference` sections of the documentation 29
* Add `numpy` as dependency for all backends 31

* Update the `TensorShare` schema with serialization and deserialization methods 15 20
* Adds the `TensorProcessor` interface for serializing any tensors to a TensorShare pydantic object and deserializing any TensorShare object to a dict of tensors in any backend. 7 15 16 20 22
* Improve package reliability by adding coverage and mypy 7 12
* Adds the `Backend` and `TensorType` Enums 7 22
* Rename/Add all the backends serialization and deserialization methods for convenience 15 16 22
* Add documentation `installation` and `usage` sections 17 22
* Fix the dependencies import to make backend dependencies optional 22

Switch LICENSE from Apache 2.0 to MIT (I don't know why I changed it initially 🤗 )

* Improved the TensorShare schema and its `model_config` parameters. 4
* Try to fix deployed documentation.

- Added import utils functions for backends 1
- Added backend-specific conversion functions 1
- Added a Roadmap to the docs 2