bd99f6e648
So that they can be processed by `--parse-metadata` for example `webpage_url` is the same, but may be modified by the extractor