Allow to look up extractors by JSON-LD provider identifier

That is, IATA or UIC codes. This allows us to find custom extractors for
content we so far only processed with a generic extractor. With that we
can achieve the same result on a stand-alone PDF as we would get with an
email context.