I recently switched from 3.3 to 3.6 and found that the new function “Get fulltext” next to the file-field is far too slow for my taste. What happened to the “Auto”-function? A more detailed version of my question follows.
In the older versions I had come up with a streamlined process to add a new entry to Jabref in minimal time and with minimal effort, but with high quality, consistent information: I imported the new entry through the medline web search which resulted in a unique key. Then I downloaded the pdf-file through the browser and stored it with the key as file name in my standard directory. Finally, I only needed to press “Auto” next to the file field in JabRef to add the pdf-file to the new entry.
I presume that the “Get fulltext” function replaced the “Auto” function, but this now takes over 30 seconds before anything happens. Eventually there will even be two file entries for the same pdf-file: one with absolute and one with relative path. This all slows down my previously highly efficient entry adding process.
So how can I return to a process which takes less than 10s on the JabRef side while maintaining the high quality, consistent information provided by Medline?
Info:
JabRef 3.6
linux 3.16.0-38-generic amd64
Java 1.8.0_101
Basically, “Auto” has just been renamed to “Get fulltext” as the functionality has been enhanced: Instead of not only looking locally for a PDF file it is also tried to automatically download the PDF from the web. So in quite a lot cases it is no longer necessary to leave JabRef to download the file but simply hit the “Get fulltext” button for an entry previously fetched (e.g., from MEDLINE).
However, your workflow still should be possible as well: Before trying to download the file it is still checked locally whether a matching file can be found. I just tested it locally and it still works for me.
Moreover, of course duplicates should not be created.
Can you please check the following settings:
Options -> Preferences -> File: Which options and values do you have in the “External File Links” section? (you can also paste a screenshot here)
File->Database properties: Do you have overridden the file settings here?
You’re right, in principal my process still works except that it takes super long (>30s) before the file is found and it is added twice, as I described. So how could I speed up the process to avoid unnecessary waiting times? It should take less than a second to check whether a file name starting with the key exists in the default directory.
Below is a screenshot of my global file settings. In the database properties the “general file directory” and “user-specific file directory” fields are blank.
