Right now you can choose between two output formats: python/repr and tikz.
However, for many users it would be more convenient to have something in between, like abstract index notation.
E.g. Tr(ABC) can be represented as A_{i,j} B_{j,k} C_{k,i}.
The notation is convenient (over e.g. outputting in graph/DOT format) because we can also output brackets. Like x_i (A_{i,j} + B_{i,j}).
Typically this kind of index tensor notation distinguishes between upper/lower indices.
It's not clear that this would be useful for tensorgrad.
Right now you can choose between two output formats: python/repr and tikz.
However, for many users it would be more convenient to have something in between, like abstract index notation.
E.g.
Tr(ABC)can be represented asA_{i,j} B_{j,k} C_{k,i}.The notation is convenient (over e.g. outputting in graph/DOT format) because we can also output brackets. Like
x_i (A_{i,j} + B_{i,j}).Typically this kind of index tensor notation distinguishes between upper/lower indices.
It's not clear that this would be useful for tensorgrad.