Fix ScienceDirect fetcher

[koppor]

For the test PDF, the json structure does not contain the full text any more. One has to use other meta data. This PR does that

Excerpt of new meta data:

    "pdfDownload": {
      "linkType": "DOWNLOAD",
      "isPdfFullText": false,
      "urlMetadata": {
        "queryParams": {
          "md5": "2b19b19a387cffbae237ca6a987279df",
          "pid": "1-s2.0-S1674775515001079-main.pdf"
        },
        "pii": "S1674775515001079",
        "pdfExtension": "/pdfft",
        "path": "science/article/pii"
      }
    },

• Change in CHANGELOG.md described in a way that is understandable for the average user (if applicable)
• Tests created for changes (if applicable)
• Manually tested changed features in running JabRef (always required)
• Screenshots added in PR description (for UI changes)
• Checked documentation: Is the information available and up to date? If not created an issue at https://github.com/JabRef/user-documentation/issues or, even better, submitted a pull request to the documentation repository.