In .yaml this is the proper format of an edge:
- [some_node.output.some_file, other_node.input.some_input]
Defining the edge this way:
- [some_node.outputs.some_file, other_node.inputs.some_input]
(notice the S in outputS/inputS) will result in an error from vh lint and can't be used for an --adhoc run.
If yaml containing such edge is to be committed and then used for a pipeline (via UI) both versions would be acceptable.
Allowing the use of outpuS/inputS in the edge definition (via yaml) sounds reasonable and is kinda more intuitive.
(The same goes for passing parameterS between the nodes btw)
I would imagine allowing the use of both versions would not be an issue - and would be an ideal solution.
In .yaml this is the proper format of an edge:
- [some_node.output.some_file, other_node.input.some_input]Defining the edge this way:
- [some_node.outputs.some_file, other_node.inputs.some_input](notice the S in outputS/inputS) will result in an error from
vh lintand can't be used for an--adhocrun.If yaml containing such edge is to be committed and then used for a pipeline (via UI) both versions would be acceptable.
Allowing the use of outpuS/inputS in the edge definition (via yaml) sounds reasonable and is kinda more intuitive.
(The same goes for passing parameterS between the nodes btw)
I would imagine allowing the use of both versions would not be an issue - and would be an ideal solution.