I thought it a bit odd, given its numerical/data analytical focus, that the article mentions mratsim's weave but not arraymancer which might in some situations allow a pure Nim approach.