Zotero 7 Beta: DOI URLs causing wrong parent information to be generated for imported PDF
After a few days of wrestling with a problem, I've narrowed down what's causing it. I kept trying to drag and drop a PDF into a Zotero collection from an external hard drive. It would work properly for a second. Then it would create a parent item with a totally wrong name, then change the PDF name to match the incorrect parent. I have my system set up to automatically create the parent item from the PDF and to rename the PDF using the parent's information. But why this seemingly random data?
I finally saw that a footnote in the PDF was the source of the metadata. I tried everything I could think of, but it never behaved correctly until I removed the URL on the footnote. But the URLs on other footnotes in the PDF were not causing similar problems. Here is the URL that caused the problem:
https://doi.org/10.2307/41215028
So I tried to create a different footnote in the document that also had a URL from the doi.org site. It did the same thing--renaming my PDF to match that particular footnote within it, rather than using the true metadata of the PDF.
I can only guess it has something to do with an attempt to extract metadata by looking for a DOI number within the document's text, but perhaps the code needs to ignore the number if it's in a footnote.
I finally saw that a footnote in the PDF was the source of the metadata. I tried everything I could think of, but it never behaved correctly until I removed the URL on the footnote. But the URLs on other footnotes in the PDF were not causing similar problems. Here is the URL that caused the problem:
https://doi.org/10.2307/41215028
So I tried to create a different footnote in the document that also had a URL from the doi.org site. It did the same thing--renaming my PDF to match that particular footnote within it, rather than using the true metadata of the PDF.
I can only guess it has something to do with an attempt to extract metadata by looking for a DOI number within the document's text, but perhaps the code needs to ignore the number if it's in a footnote.
We'd want to see the actual PDF in question. Can you link to that, or send it to support@zotero.org with a link to this thread?