During installation of Prodigy in the python virtualenv it might be installing a spacy version. Now in the same virtualenv if I dev install a spacy version with some changed code in it, will prodigy use that for its active learning? Or do I have to change config file of prodigy to use a custom built spacy version? I am about to embark on customizing some code in spacy as per Incorporating custom position feature into NER. So this question.