Improve npm module parsing #22
Description
Papers does a fine job parsing the main project package.json and populating papers_manifest.yml with placeholders, but it doesn't currently go any deeper.
Installed node_modules are packaged with (at least some of) the metadata required to automatically fill out licence, license_url and project_url in the manifest.
Would be pretty helpful for papers to iterate through npm dependencies and extract that stuff to further reduce manual labor.