Improve sort order of JSON returned; align with ScanCode toolkit

[pombredanne]

The files section of a JSON scan looks this way:

    {
      "for_packages": [
        "pkg:pypi/et-xmlfile@1.1.0?uuid=cbee3d4a-50c0-480f-8193-005b7b4328c8"
      ],
      "compliance_alert": "",
      "md5": "a02be95d0116629188cd4c2a3fa67660",
      "sha1": "71e4b08e9402a313dc3b1f682854e31f2ba4ee08",
      "extra_data": {},
      "copyrights": [],
      "holders": [],
      "authors": [],
      "licenses": [],
      "license_expressions": [],
      "emails": [],
      "urls": [],
      "path": "et_xmlfile-1.1.0.dist-info/AUTHORS.txt",
      "status": "application-package",
      "size": 64,
      "tag": "",
      "type": "file",
      "name": "AUTHORS.txt",
      "extension": ".txt",
      "programming_language": "",
      "mime_type": "text/plain",
      "file_type": "ASCII text",
      "is_binary": false,
      "is_text": true,
      "is_archive": false,
      "is_key_file": false,
      "is_media": false,
      "package_data": []
    },

The order of fields seems random.
We should instead adopt something like ScanCode toolkit and at least for the base file info something such as https://github.com/nexB/scancode-toolkit/blob/ded56e9120f5fdfb9a1a0309130bb4305a66aacb/src/scancode/api.py#L366