This folder contains some code examples for the DSL. Each example is in its own folder, containing two sub-folders: one for training and one for classification.
While the DSL does not directly support training, the current features allow k-NN feature vectors to be generated along with their ID in a format that is close to what a training set file would have. This allows us to specify a program in the DSL to generate the feature vectors to the standard output and then redirect the output to a file, replacing occurences of [ and ] in said file (these characters are present in the output since the DSL uses them to delimit vectors).