Introduction
============

One of the reasons for the success and repeated use of tensor networks (TN) in Physics and other fields is the possibility to visualise complicated calculations diagramatically. The central idea of this package is to introduce a layer of abstraction that *one-to-one* represents such diagramms and the manipulations on it.

Pro-List
--------

* One-to-one translation from a TN-scetch to the python code.
* Independent development of low level (contraction, svd,..) and high level (contraction order, svd sweep directions,...) algorithms.
* Encapsulating the lower level allows for a better (fairer) comparison of higher level algoithms.
* Potential to encode extra information that are relevant to the TN. For example: specifying that two indices of a tensor are in fact the same (connected by a kronecker delta).
* Potential to use extra encoded information to optimise the execution of the lower level (contraction, svd,...).

Con-List
--------

* More overhead then good low-level implementations.
* *hmmm...*