Export Abstract

Wbarobinson

Hi all,

I am new to Zotero and I was wondering how I could export the abstracts that go with my citations. Every time I try to export I fail to bring the abstract. Also, if possible I would love to export to excel. Also, if possible I would love to export the location of the image file on my hard drive to excel.

Cheers,

Will

adamsmith

could you say more clearly what you mean by "export"?
Actual exports - e.g. to RIS - include abstracts.
Most citation styles don't, though, but see this thread for some options
http://forums.zotero.org/discussion/73/annotated-bibliography/#Item_16
Also reports have abstracts (right-click--> generate report).

There is no usable Excel export option - generally it wouldn't be impossible to create a citation style that exports to a semicolon-separated text file (which is like a .csv so Excel can import it). The main problem here is how you want it to look exactly - which fields go where etc.

Wbarobinson

Hmm, that is strange. I could not see anything that looked like my abstracts in RIS export which I opened with endnote. I will look. A little harder.

I am using zotero for non-citation based work. I am compiling webpages in an easily manageable format and hope to eventually use the new silverlight software, pivot.

My hope was to query my zotero data to produce .csv or .xls file which opened with excel to show each entry with the columns: path to image of website, blank, date that i entered (not any of the automatically generated dates), author name, title, abstract. The easiest way seemed like it would involve a zotero export function.

I don't really know if that is feasible or how to begin.

adamsmith

According to RIS specifications, Zotero puts the abstract in N2 - if I remember correctly, Endnote expects it in AB or so - which may sound more intutive, but is wrong.

For what you want, it would actually quite easy to write a csv export as a regular citation style if I understand you correctly:
All that would be for one item type (website)?
essentially all you would want is to have in the bibliography section of the style
<text variable="URL"/>
<text value="; ;"/>
<text variable="issued"/>
<text value=";"/>
<text variable="author"/>
<text value=";"/>
<text variable="title"/>
<text value=";"/>
<text variable="abstract"/>

you can see if this together with these instructions is enough,
http://www.zotero.org/support/csl_simple_edits
or I can quickly write that style for you, although I'd be helpful if you uploaded a google doc with three or four sample lines.

Wbarobinson

That is extremely kind of you, I will try my best with what you offered and only if I fail will I trouble you more. Thanks,

Will

Wbarobinson

I am trying, but I get the feeling that I should be modifying a particular citation style, given what you have written. Is that correct?

Will

adamsmith

oh well - that may have been a bit much (plust there were actually some mistakes in my post above), so take this as a start:
http://gist.github.com/500876
you can download the style using the "Raw" link on that page.
Install by dragging to an open Firefox Window.
There isn't much going on in that style, so you should be able to understand it easily.
Let me know how it goes.

Wbarobinson

So do I right click my collect, choose Create Bibliography from Collection...
Then pick the CSV-Export. I cannot seem to get the choice to make a .csv file. When I copy to clip board and paste into excel, or change .rtf to .csv, all the data appears on one line. How do I get in a spreadsheet form?

Will

adamsmith

oh I should have explained that:
select the items, right-click-->create bibliography --> clipboard --> copy this into a text editor (like notepad) and save as .csv
open that with Excel

Wbarobinson

Yeah, that ends up with the same thing where there are no columns, and every piece of information on a given entry is on the same line. :S

I would like it broken into columns: path to the saved image of the website (snapshot); blank; date; author; abstract.

Thanks for your prompt responses!

adamsmith

you need to tell Excel that the columns are separated by semicolons.
Maybe use import or something? Sorry, I don't have Excel - just tried this in open office calculator and that works.

adamsmith

oh and: I can't give you the path to the snapshot:
Zotero doesn't output that easily - it will just be the URL

Wbarobinson

Great! I got it to work, now how do I import the location of the saved snapshop?

Will

Wbarobinson

Sorry, failed to refresh the page before commenting... but is there anyone I can ask about that? It is mildly important.

noksagt

You do not have access to that information when using CSL styles to create bibliographies. You are correct that you would either need to build a translator from scratch to have that kind of information. If exporting attachments is acceptable, look at the BibTeX or RDF translators.

adamsmith

here would be the place but you're probably not going to have much luck.
Zotero doesn't output that path with a citation style, which means you have to write an export translator (similar to the RIS one) in java script - more work and less people able to do it.
But even there I'm not sure how easily you can get to the path for the snapshot, I've never seen that in Zotero output.

edit: generally what noksagt says, though I'm not sure how going through RIS would help - that doesn't export the path to the snapshot either, maybe he can clarify.

ajlyon

As for making RIS export work with Endnote, it seems to me that we could bow to the market and make importers and exporters that supported Endnote idiosyncrasies -- perhaps an RIS (Endnote-style) translator. There's an import issue as well, right?

noksagt

But EndNote's idiosyncrasies change from version-to-version of their importer. We endeavor to produce output that conforms to the RIS specification (which is hosted by Thomson Reuters, the makers of EndNote).

We could potentially improve support for EndNote by making a custom EndNote import filter and/or by pointing to a specific version of an RIS import filter that already exists. However, Thomson Reuters claims restrictions on all filters, including user-created ones. I would, therefore, direct anyone who had problems importing valid RIS into EndNote to refer to EndNote's support for a solution.

ajlyon

I knew there was a reason I was glad I never used Endnote. We are patient in explaining how to move back in forth in the forums, and we have extensive documentation. I suppose that's enough bowing to the market. Especially since we're probably not an insignificant market player ourselves.

jimgwc

I attempted to use the file posted by adamsmith to export to .csv, but ran into a problem. I was able to drag the file into firefox and install it, but when I tried to create a bibliography, an error message was displayed "An error occurred generating your bibliography. Please try again". The only suspicious thing that happened was a gray bar was displayed at the top of the window after installing the file, which displayed "This XML file does not appear to have any style information associated with it. The document tree is shown below."

Any help would be appreciated!

adamsmith

no, the grey bar is normal. Generate and send an error report (Gears Menu) after triggering the error and post the error ID here.

adamsmith

actually, before you do this, try the one here:
https://gist.github.com/843991

jimgwc

I tried the new link, but still got the same message. The Error ID is 906100320.

adamsmith

ok thanks - I'll need for one of the devs to tell me what's in that error report.

DWL-SDCA

Has there been any movement on the RIS N2 vs AB issue. I find that this is a great frustration when importing things from SpringerLink (especially) and other publishers' sites. Springer uses AB for the abstract and so does Wiley and Taylor and Frances. So when Zotero intercepts the RIS file the abstracts don't get imported.

Might I suggest that upon import of RIS if there is an N2 tag it should be imported into the Zotero abstract field. If there is no N2 tag but an AB tag that should be imported into the Zotero abstract field.

Also, SpringerLink (and if I remember correctly, others as well) use the tag DO for doi number instead of M3. If an RIS record contains no M3 tag but contains a DO tag, might there be a test of the DO field for a string that "looks like" a doi number and, if so import it.

Thanks.

ajlyon

Try this version-- I agree that we can be a bit more flexible in our input, since there are some unambiguous widespread abuses of the spec that we can anticipate and handle already on import.

Please go to http://github.com/ajlyon/zotero-bits/raw/master/RIS.js and save the file to the translators directory of your Zotero data directory (http://www.zotero.org/support/zotero_data).

If this works for you, please post here so that I can submit this change to be pushed to all users.

adamsmith

these sites have Zotero translators, right?
In those cases it's actually possible to re-assign the fields post RIS - the SpringerLink translator already does some cleaning up (though I thought it didn't work?).

Changing regular RIS import is trickier - the types of testing you suggest are possible, but also make a super-crucial translator more fragile, which I'm not sure is a good idea.

ajlyon

The way the RIS translator works, we can pretty safely map additional input fields. I just don't want out-of-spec hacks to multiply unmanageably in RIS.js. DO and AB are extremely widespread, so I think we can afford to handle them (import them, not export them) in the main translator. That also simplifies the site translators that use RIS.js, many of which have to hard-code this very common abuse of the spec.

DWL-SDCA

The new RIS import translator does the job with the abstract and the doi number.

However, there is a problem with the page numbers. The final page number is repeated.

"367-383" becomes "367-383-383"

see the article at:

10.1007/BF02287690

or at:

10.1023/B:JOFV.0000032635.42145.66

Responding to adamsmith's question about sites having translators --

With SpringerLink the toolbar icon brings full metadata for some articles, incomplete metadata for most articles, and prompts an error for many journal's articles. Even for the articles that cause a Zotero error, a download of a Reference Manager format of the RIS includes all metadata when viewed in a text editor. With some SpringerLink journals, the address bar icon doesn't appear (but I can download complete RIS metadata).

DWL-SDCA

The ProCite flavor of the RIS download works flawlessly. (I had been using the Reference Manager download.) There are no doubled final page numbers.

The ProCite is the default selection.

Thanks.

adamsmith

yeah - that means SL is still broken - Zotero is picking up the article's DOI.
Ajlyon will know this better, but remember there being some type of obstacle that made fixing it really hard to impossible.