PLoS journals attaching snapshot even if I don't want it

kajeling

Any time i save a cite from a PLoS journal it attaches a snapshot, even though I unchecked the "automatically save snapshot" option. I looked in the translator and it's very short and I don't see any obvious place where it is attaching a snapshot so I'm not sure how to fix it.

eg: http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0005655
using the PLoS Journals.js translator in zotero 2.0b4

Thanks,
Keith

adamsmith

the "automatically attach snapshot" option refers to using the "Creat New item from the Page" option, rather than translators.
ATM you can only turn off associating any file including pdfs (by deselecting the pdf and other files) option, or get both the pdf and the snapshot from translators.
Afaik that is the case for all translators, not just PLOS.

kajeling

As far as I can determine, that's not really true. Turning the "automatically attach snapshot" option on and off definitely controls whether translators attach snapshots for other sites.

That would be a pretty annoying strategy anyway, as I always want it to grab PDFs, and never found snapshots particularly useful (especially when backing up/restoring all those little files for all my cites)

So at least on my computer, the PLOS translator is the only one that seems to ignore the setting.

Rintze

In my experience the option also holds for translators. The reason the PLoS translator still saves a snapshot seems to be that the RIS file, which the PLoS translator loads, contains a "UR" field with a link to the article webpage. The Zotero RIS digester seems to treat that as an instruction to save a snapshot.

This afternoon I rewrote part of the PLoS translator, and submitted it to the Zotero team. It should fix multiple limitations/issues, including this one.

dstillman

Yeah, generally speaking, the snapshot pref should currently affect translators as well, though there are issues with dependent translators, as Rintze notes.

The larger issue is that the snapshot pref probably shouldn't affect translators and should just be used for "Create New Item from Current Page". For translators, a snapshot should probably only be saved when it's full-text content such as a New York Times article and should be part of the download pref—and not the snapshot pref. (There could possibly be a "large files" sub-pref there to allow people to exclude, say, multi-megabyte YouTube videos while still saving NYT articles, but conceptually speaking it should be the same pref.)

Web link attachments, which is what, for example, the Amazon translator saves (i.e., non-canonical related links), should either be saved regardless or should be a separate new pref.

kajeling

If it ever ends up being implemented the way you describe, please be sure to provide an option we can use to disable the functionality. Automatically download/attaching a PDF is one thing, but automatically downloading a whole bunch of little files for a snapshot is something over which I would like veto power.

Don't get me wrong. I love snapshots for certain things, understand their usefulness, and appreciate all the work you guys have done to implement it, but it is the single-file versus many-file thing that I would like to at least have some control over.

dstillman

I could see there being a multi-part pref like this:

When saving items via translators, automatically download:

[ ] HTML snapshots

[ ] PDFs and other single files

[ ] Large media files such as videos

noksagt

The larger issue is that the snapshot pref probably shouldn't affect translators and should just be used for "Create New Item from Current Page". For translators, a snapshot should probably only be saved when it's full-text content such as a New York Times article

Having full-text HTML snapshots of journal articles (such as the article linked to in this thread) is still useful for annotation (and, sometimes, better indexing) even when a PDF can be downloaded.

dstillman

I hadn't noticed that the snapshot on that article is full text, but you're right. I was mainly referring to translator snapshots where it's just the database page—basically the citation and abstract and nothing more—and in those cases I suspect it's not necessary to save. For full content representations, though, there's no reason Zotero couldn't save all enabled and available media types based on a multi-part pref like the above.