Import : Use unAPI, mods, json or else ?

gduh

Hi,

I'm new with Zotero, I read a lot of documentation (not always easy to find).

First, my goal is to create a web site including data that can be imported with Zotero/Firefox. I found the best solution is to use unAPI, even if it seems that the developers's team doesn't exist anymore.

So I try to write a small web site just to try this solution.

After severals tests using different formats like mods, json and rdf_zotero, it appears that Zotero using "only" the rdf_zotero's format. In fact, the import function show the option, "Enregistrer dans Zotero en utilisant unAPI" only if rdf_zotero format exists. Otherwise, the import function show only basic options "Enregistrer dans Zotero comme page web".

So my question is: Am I oblige to use the format : rdf_zotero ?
For differents reasons, it will be easier for me to use json format, is it possible ?

Finally, if I'm oblige to use rdf_format, is there a framework or something else to help build the flow (web application) since the rdf_format is quite complexe !!!!

Thanks for your answers.
Gregory

noksagt

So my question is: Am I oblige to use the format : rdf_zotero ?

No; other formats should work (I use MODS XML with unAPI). UnAPI allows you to support multiple formats, so there's no reason not to also support rdf_zotero if you have already implemented it well (though it sounds like that may not yet be the case). The currently recognizable formats for the unAPI translator are:RECOGNIZABLE_FORMATS = ["rdf_zotero", "rdf_bibliontology", "mods", "marc", "unimarc", "ris",
"refer", "bibtex", "rdf_dc"];
You'll note that JSON does not appear in that list. I have limited experience with it and don't know of the strategic thinking behind whether it should/could be a real inter-change format...But you do have plenty of other options if you'd like to support something other than rdf_zotero.

adamsmith

no, unAPI should work with most (all?) formats Zotero supports (edit: see noksagt's answer for a more precise version of this). My guess would be that something about the MODS you serve was invalid. Zotero would check that and doesn't import if it's invalid.

I know for sure that MODS works, but don't have a great example at hand.
E.g. the U Toronto library catalog serves RIS over unAPI:
http://search.library.utoronto.ca/UTL/index?N=0&Ntk=Anywhere&Ntt=nimni+challenge+of+post-zionism&Ntx=mode%2Bmatchallpartial&Nu=p_work_normalized&Np=1&formName=search_form_simple

DWL-SDCA

My SafetyLit (short for safety literature) site does much of what you desire. [SafetyLit is a free service of the SafetyLit Foundation in cooperation with San Diego State University and the World Health Organization]

Export to personal bibliography management software: each record is available in multiple formats for manual download and is recognized by Zotero for one-click download. Several thousand records are downloaded each day in multiple metadata formats.

Import of metadata into SafetyLit: Although some journal article metadata is brought into the SafetyLit database via ftp directly from publishers, we find that to be useful only for journals we index cover-to-cover. Other journal articles' metadata is downloaded to Zotero after we identify relevant articles on publisher sites, edited within Zotero, then exported in MODS, and imported into SafetyLit. SafetyLit draws content from about 15,000 scholarly journals.

The SafetyLit database houses about 485,000 journal article records. Also included are about 30,000 book, technical report, conference proceedings, and doctoral theses.

We find that Zotero is a very useful, probably an indispensable, tool for getting accurate metadata into the SafetyLit system. Publisher metadata is messy. It is inconsistent and often omits key information. Zotero translators automatically fix most of the errors and inconsistencies. The right Zotero panel allows us to quickly identify and correct problems. Thus, very few records require further editing after they are brought into SafetyLit.

adamsmith

That's http://www.safetylit.org -- yes, that's a nice example of unAPI serving MODS (and working flawlessly with Zotero).

gduh

no, unAPI should work with most (all?) formats Zotero supports (edit: see noksagt's answer for a more precise version of this). My guess would be that something about the MODS you serve was invalid. Zotero would check that and doesn't import if it's invalid.

To check the MODS, I get a MODS from an existing website, so I think it must be well formed. But if I use it in my own website to test it, Zotero find that it exists a MODS format, but it doesn't import it !!!
If I get a ref_zotero format from the same website with the same data, then if I use it in my own website, Zotero imports it.

There is certainly something wrong in the MODS but I don't know what.

By the way, how can I build easily my own MODS ?
At least where can I found a complete description compatible with Zotero ?

adamsmith

not all MODS you find in the wild is valid, no (e.g. INSPIRE, the large high energy physics website, serves invalid MODS). Which website?
One easy way to get valid MODS is to export from Zotero.

noksagt

@gduh:
See a working example of unAPI+MODS at:
http://arc.nucapt.northwestern.edu/refbase/show.php?record=10704

Your unapi resolver should include a format response similar to:
<format name="mods" type="application/xml" docs="http://www.loc.gov/standards/mods/"/>

zuphilip

Another example for unAPI implementation: https://www.gbv.de/wikis/cls/UnAPI (German website)

gduh

Thanks for all yours answers.

@adamsmith:
I export a valid MODS directly from Zotero. When I use it with my website, now it's ok. So, I compare it with the first one I used and the only difference consists in upper/lowercase in naming tags. For example "modscollection" instead of "modsCollection" !!!!!

So, now it's ok with unAPI/MODS.

Next step, I need to know how to build a well formed MODS structure with the correct list of fields recognized by Zotero. If at least, someone can tell me where I can find that list, it would be great !

zuphilip

The good thing about standardized metadata format is that they are standardized: "The standard [MODS] is maintained by the Network Development and MARC Standards Office of the Library of Congress with input from users." http://www.loc.gov/standards/mods/ . Try to stick to the standard and maybe there are some useful tools on these sites (MODS validator?).

The standardized MODS should then import neatly into zotero. There is the code for the mapping between MODS and Zotero: https://github.com/zotero/translators/blob/master/MODS.js

gduh

@zuphilip:
Thx for the links, I'll look at that !