Multitask and multimodal api improvements
Created by: Dreamlone
- Added new example with running multitask through API (it does not work and in this PR probably do not plan to fix it - for now this is an example of how I would like to apply data to the input)
- I added the ability to tune multimodal pipelines with multimodal data for multitask
- Redesigned the MultiModalData logic, so that this dataclass can be used for multitasking tasks with different target
- Prepared a prototype of how to run FEDOT on multimodal data via API (check tests)