Upgrade StorageAdvanced Metadata Retrieval
Hello,
Is it possible to further improve Zotero's metadata retrieval capabilities in the future (perhaps even with AI support)?
https://s3.amazonaws.com/zotero.org/images/forums/u8284066/3arrbyaeqm0r9r28oyes.png
While it's true that Zotero mostly works quite well—especially with open access articles that have DOIs—there are still cases where it fails, as shown in the screenshot. For example, the PDF file in SS is a book review, has no DOI, but it's fully open access, unprotected, and doesn’t require OCR. Even though the relevant settings are enabled in Zotero, it still left the file as a plain PDF, and I’m not sure why.
Additionally, when Zotero fails to retrieve metadata, usually it tries to autofill the metadata itself. However, this can sometimes lead to incorrect or corrupted entries. For example, it might mistakenly insert the author's name as the title, use the book title in place of the publisher, or include unrelated phrases like “All rights reserved.” etc. Can this be improved so Zotero better understands and extracts accurate metadata? And along with this, could an option like 'Retry metadata retrieval' be offered?
Regards,
Is it possible to further improve Zotero's metadata retrieval capabilities in the future (perhaps even with AI support)?
https://s3.amazonaws.com/zotero.org/images/forums/u8284066/3arrbyaeqm0r9r28oyes.png
While it's true that Zotero mostly works quite well—especially with open access articles that have DOIs—there are still cases where it fails, as shown in the screenshot. For example, the PDF file in SS is a book review, has no DOI, but it's fully open access, unprotected, and doesn’t require OCR. Even though the relevant settings are enabled in Zotero, it still left the file as a plain PDF, and I’m not sure why.
Additionally, when Zotero fails to retrieve metadata, usually it tries to autofill the metadata itself. However, this can sometimes lead to incorrect or corrupted entries. For example, it might mistakenly insert the author's name as the title, use the book title in place of the publisher, or include unrelated phrases like “All rights reserved.” etc. Can this be improved so Zotero better understands and extracts accurate metadata? And along with this, could an option like 'Retry metadata retrieval' be offered?
Regards,