Fabian Zills

performance benchmarking

2

https://github.com/zincware/ZnFlow/blob/d1e6f5ddc75cf2849f546d9afd08ef271f843544/znflow/graph.py#L125 causes all the getters to fire and try to get the values. This would be okay if they were not trying to all read from file

Iteratively search gitlog for the given node names and provide git commit hash.

node_name being `123_123` might be cast to int

1

- check that a node name can be `f{123}_{123}` - ensure `self.name` is string - in the `_` warning, suggest using `-` instead

issue with attribute being a connection when accessed outside graph

3

also test, if changing self.params or self.x_path in run has an effect

if you rename a dependency from `a` to `b` and rebuild the graph you might get a keyerror. Why is the graph building reading anything from these files?!

e.g. `from_rev().frames` takes long

E.g. `zntrack list *EvaluateCalcResults`

Inspired by Typer you can use annotations instead of the Hydra like config object.