When using the following example with bdfiddle, there is an error parsing the metadata.jsonld:
There is metadata that is not part of the extraction that is about the user doing the extraction that breaks the parser. If no extractor_id is found, just ignore the metadata. Additionally, the extractor_id string split expects all extractor id's to us a URI like "https://clowder.ncsa.illinois.edu/clowder/api/extractors/ncsa.cv.eyes" however, this is not true for some (e.g. ncsa.cv.caltech101), the extractor_id is simply "ncsa.cv.caltech101" so the datawolf tool finds no metadata.