Improved documentation - suggestions welcome

[jcmgray]

The tensor network documentation, in particular, has fallen a bit behind the functionality. This issue is to keep track of what extra documents will hopefully be written and any suggestions for content.

Docs

• [x] A better introduction to the tensor network functionality
  • [x] Describe the different levels of interface:
    • 'beginner', highest level interfaces (DMRG, Circuit etc) which manipulate tensor networks for you
    • specialized TensorNetwork classes like MPS, PEPS which encapsulate behaviour.
    • TensorNetwork for generic graphs
    • Tensor for manipulating labelled arrays
    • autoray for the actual backend agnostic array handling
  • [x] How to build TNs and address/select tensors from them including best practices
• [x] intro to TN drawing functionality
• [ ] intro to TNOptimizer, including maybe TensorNetwork.fit
• [ ] intro to contraction (exact)
  • [ ] explain sensitive tradeoff between pathfinding time and contraction time
  • [ ] explain usage of optimize={preset|path|optimizer}

Examples

• [ ] ?

[rezah]

@jcmgray That is super cool to extent documentation: I think we can tell people QUIMb can handle infinite/periodic TN such as iMPS and scale invariant MERA. I will complete the flowing list

1. transitional invariant TN like MPS with periodic boundary conditions
2. infinite TN: especially iMPS, scale-invariant MERA
3. learning CoTenGra in a more practical way? I am curious about that
4. 2D Tensor networks like MERA/TTN is missing here

[jcmgray]

A detailed guide to drawing has been added here - https://quimb.readthedocs.io/en/latest/tensor-drawing.html.

[jcmgray]

The basic introduction to tensor network functionality has been re-written - https://quimb.readthedocs.io/en/latest/tensor-basics.html.