Sergio Sánchez Ramírez

Replace OMEinsum with TensorOperations

5

After some benchmarking, I've found out that OMEinsum is terribly slow and would like to see how TensorOperations performs compared to it. This PR replaces OMEinsum with TensorOperations. If it...

# To do - [ ] Refactor into a `draw` function (just like `quimb`) - It should be customizable - [ ] Take defaults from GraphMakie extension https://github.com/user-attachments/assets/255b4f0e-1301-4cf1-96e4-325a6a85d54e

KrylovKit.jl uses [VectorInterface.jl](https://github.com/Jutho/VectorInterface.jl) to use any object "that can be treated as a vector" (e.g. an MPS) to run Krylov-subspace methods on top of it. By integration with it, using...

`active_reg_inner` crashes on `Tuple{X} where {X}`

2

If you run this code: ```julia Enzyme.Compiler.active_reg_inner(Vector{Tuple{Code}} where {Code

"Tags with side effect (<script> and <style>) are ignored in client component templates" error on WGLMakie + DocumenterVitepress

2

I'm trying to plot some graphs with GraphMakie and WGLMakie on my docs. I'm trying to make it work with DocumenterVitepress.jl but I'm running into the following error: ```js [vitepress]...

## notes - MPI types are not ABI stable, but fortunately there is [MPItrampoline](https://github.com/eschnett/MPItrampoline) which provides some wrappers over them - `MPIABI_Comm` - `MPIABI_Request` - `MPIABI_Op` - docs on how...

fix fft rev diff rule

1

CC @avik-pal should fix #170

Diagrams not correctly included with DocumenterVitepress

3

This package is not working for me. Instead I get these text messages in the HTML:  I'm using Julia 1.11.2 with Documenter 1.8.0, DocumenterVitepress 0.1.6.

Move to a interface system

1

The main change is to stop relying on the `AbstractTensorNetwork`, `AbstractQuantum` and `AbstractAnsatz` abstract inheritance system and move to a [duck typing](https://wikipedia.org/wiki/Duck_typing) system. Since duck typing cannot be enforced, the...

This PR implements the ideas of #314 but with the minimal breaking changes; i.e. `Quantum` and `Ansatz` still exist. Breaking changes should come in another PR.