For the moment the dependency graph is determined in the tuttlefile. We need a way to create the graph according to the environement. For
example if one wants to process :
a list of images in a directory
some external stuf (eg a list of table) that are defined in a configuration file
create some specific resources, like process:// which will declare processes in the graph
when a process produces a process:// resources, it can only depend on primary resources, in order to be run first
when the workflow is executed, if first runs the process:// processes, which extend the dependency graph, then static check is run on all processes, then processes are run
should there declaring processes not produce resources used by other processes ?
For the moment the dependency graph is determined in the tuttlefile. We need a way to create the graph according to the environement. For example if one wants to process :
This might require that process retreive the list of inputs and outputs. For example, in Drake each process has access to environement variables $INPUT1, $INPUT2, $OUTPUT1, etc. Ex : https://github.com/Factual/drake/blob/develop/demos/McIlroy/Drakefile
A way to do this could be :