From archive to annotation: the pipeline of scanning, crowdsourcing, transcribing and modelling