Fabian Zills
There is a bug with using the Node after `zntrack.apply` as dependecy, e.g.: ```python mapped_md = zn.apply(ips.calculators.ASEMD, method='map')( data=md_selection.atoms, model=model1, thermostat=thermostat, checks=[], modifiers=[], steps=10, sampling_rate=1, pop_last=True, use_momenta=False, constraints=[], name="mapped_MD" )...
It would be handy to get some information once a stage has successfully finished. This could include - email - discord - slack - mattermost - ... The configuration should...
As the Nodes are part of groups so should the parameters be sorted into groups. There are two approaches to this: 1. sort in the `params.yaml` and have subsections with...
do not run the graph using `dvc repro` but some `zntrack repro ` which write the node only after/ shortly before it is being executed. Once you encounter an `if...
We can use https://github.com/microsoft/pyright/discussions/1910 To allow for: ```python class MyNode(zntrack.Node): params: zntrack.params[dict] = zntrack.params() ``` or even ```python class MyNode(zntrack.Node): params: zntrack.params[dict] ```
If you want to do something like: ```python class MyNode: self.inputs = zntrack.deps() def run(self): self.data = sum(self.inputs) class Linspace(zntrack.Node): args: tuple = zntrack.params() results = zntrack.outs() def run(self): self.results...
If you add an external node, the name is only a cryptic hash value. It would be beneficial to give a name, maybe even automatically generate one.
`zntrack.deps` should only accept nodes, therefore it should be easy to raise a good error message if e.g. a standard python class like a bumpy array is given to make...
This is not tested, but I'd assume that ```python class A(Node): paramter: int = zntrack.params() def do_stuff(self): ... class B(Node): paramter: int = zntrack.params() def do_stuff(self): ... a = A(5)...
- What happens if the `env.yaml` exists but no entry - what is the default value (seems to be None of the `env.yaml` exists, otherwise the default there)