I have written a nom parser that produces a syntax_tree which I can iterate over intermediate and leaf (token) nodes. For further processing I need to rewrite this into an AST with bottom-up evaluation. I've also been thinking about feeding into a parallel stream processor framework (e.g. timely) but I'm still unsure if this can be successful (e.g. timely workers are idiomatically more or less identical in their processing logic). Alternatively, some other reactive frame work to push reactors top-down to be evaluated when token leaf nodes are reached....
Some experience based hints are much appreciated!