Using Filename conventions to get PDF metadata

williamf · April 10, 2010

I have a large library of PDF articles that I've sorted into directories by journal name using the following filename convention: AuthorsName_ArticleTitle_JournalNameVolume.Issue_PageRange.pdf. I was wondering if there was anyway to:
(1) Import the whole directory structure straight into Zotero, having it create collections automatically by journal name AND/OR
(2) Use my filename convention to get the PDF metadata -- the problem is that many of the articles aren't OCRed and thus Zotero complains about not being able to download the metadata.

I'm very new to using this program, but in my brief search through the forums, I haven't found any discussion of these possibilities. All help is greatly appreciated -- thank you in advance!

adamsmith · April 10, 2010

neither of those are possible and I unfortunately don't see anything currently in Zotero that would help you much in that task.
The only thing I can tell you is that you can drag pdfs to Zotero which may make importing faster, but that's all - sorry to be the bearer of bad news.

williamf · April 11, 2010

Thanks for the quick response. That's quite annoying, and I noticed another problem with dragging files into Zotero -- it insists on copying the originals into its bizarre directory structure. The only way to create links it seems is to create new entries from scratch, unless there's some way to modify this default behavior when dragging and dropping. The search for the perfect iTunes-like utility for indexing and tagging PDFs continues . . .

dstillman · April 11, 2010

The ability to "link" (rather than "store") dragged files is planned, but linked files have a number of disadvantages compared to stored files—most importantly, they don't sync and they don't work cross-platform (at least until custom-root-directory-relative linked paths are implemented).

Friendlier directory names for stored files are a possibility, but they're not a priority, as there are trivial workarounds, such as using a saved search from your OS to show all PDFs under Zotero's directory structure or, when you're using Zotero, simply dragging files straight out of the Zotero pane. And, of course, if you know what you're looking for, a decent system-wide search feature should be able to find anything immediately regardless of where it is.

williamf · April 12, 2010

Thanks for your response, Dan. I would like to see linking/storing become two equally good options, as well as a smarter directory structure feature, particularly as someone who has already set up his directory structure as usefully as possible. In terms of the workarounds you suggest, while they are all possible, I sometimes would like to navigate directly to a PDF whose location I know, without having to fire up Firefox and use Zotero (or any other application, for that matter). Perhaps this could be done by having Zotero set up shortcuts or symbolic links to a central bank of files, or by having a primary tag/collection where the file would be located and have other tags/collections point to it?

adamsmith · April 12, 2010

william - I think Dan's point is that you don't need to fire up FF ( :-) ) fo that - you could used an intelligent saved-search/virtual folder or a good desktop search (google search etc.) - especially when you have good naming conventions for files (also, there will likely be tools like Gnotero for other operating systems than linux in the future). The point is that you can achieve effects close to the smarter folder structure through other, readily available means, so development on such a structure is not a major priority.
On the other hand treating links and stored files as "equally good" options is not going to happen for some time if ever - syncing linked files just seems like a potential nightmare, and relative paths - while generally a good idea, should be a while off, too.

If none of that satisfies you, you could also take a look at Papers (if you're on Mac) and Mendeley (on any OS), who, while imho as overall bib tools not on par with Zotero, have some elegant solutions for pdf management.

williamf · April 12, 2010

Thanks adamsmith. I realize there are other avenues to use, but it would be nice to have it all under one application. I'm not on a Mac, so Papers and other Mac programs are out. I'm about to try Mendeley, which I just noticed extracts PDF XMP data, which many of my PDFs have. Thanks to you and Dan for your help!